1

As a research analyst, Debadeep works on finding new and exciting consumer insights from Google Search data for over 25 countries across Europe, the Middle East and Africa. He has published over 25 Think with Google articles and contributed to several market research papers about consumer behaviour.

AUTHORS

The Behavioural Architects

The Behavioural Architects are an award-winning global insight, research, and strategic consultancy . Their work is underpinned by the latest thinking from the behavioural sciences which they leverage to help organisations better understand and influence consumer behaviour. They apply behavioural inspired frameworks and methods to help address key strategic marketing and social challenges .

Sian Davies

Founder, The Behavioural Architects

Sian is global co-founder of The Behavioural Architects and has over 30 years’ experience in market research and strategic consulting . Over the past 12 years, Sian has pioneered the application of behavioural science insights in both qualitative and quantitative research methods .

Dipesh Mistry

Associate Director, The Behavioural Architects

Dipesh has close to a decade’s experience in research, which has centred around applying his strong academic background to the commercial research sector . Dipesh’s career has centred around creating innovative methods to address a range of behavioural questions .

‌Introduction‌

In 2020 we published a piece of research called Decoding Decisions: Making sense of the messy middle that helped explain how shoppers navigate the complexity of the web — or the “messy middle” as we called it. We started out by exploring “internet street”: online shopping reimagined as a digital high street, with every storefront, every product, and every point of view and opinion immediately at your fingertips.

Over the course of detailed literature reviews and simulated shopping experiments, we became more familiar with the behavioural science underlying our decision-making, eventually proposing a model formed around two closely related loops: exploration — an expansive mindset; and evaluation — a reductive mindset .

The reason for grounding our research in behavioural science was simple: people were foraging for information long before the internet existed, and long before there were streets. Many of the mental shortcuts our ancestors used when searching for their next reliable source of food turn out to be the same as those we now apply when figuring out where to buy our next pair of trainers . Abundant choice and limitless information have made both physical and digital environments “messier”, but the cognitive tools we use to navigate them have endured .

07 INTRODUCTION

want to know how these insights can be put to work at scale, so that every customer interaction is enriched and improved by an appreciation of how and why people make the purchase decisions they do .

To answer these questions — and more — we set about simulating three specific environments: Search, content, and online points of sale.

Content

Point of sale

We chose these three spaces because they represent places where marketers have a significant degree of control, and therefore the ability to experiment, iterate, and try new things .

If you’re already familiar with our original research, you can skip to the next chapter . But if you’re new to the messy middle,

you can find a quick recap below.

As in our original research, our insights are derived from a combination of literature reviews, observational studies, and rigorous survey-based experiments. And once again, we’ve partnered with behavioural science experts The Behavioural Architects to validate and inform our work.

08 INTRODUCTION

‌Revisiting the messy middle

The messy middle is a complicated place . An environment of abundant choice and limitless information. And it’s getting more complex every year as people are introduced to new modes of communication, new means of finding information, new forms of entertainment and expression, and — of course — new products and services to buy and new ways to buy them .

Shoppers are pretty complicated too. What they want can be a moving target, with intent to buy one product often ending with the purchase of something completely different. And all of this is governed by motivations that are hard to explore without the conscious mind inventing post-rationalisations .

In our model, the messy middle sits between the twin poles of trigger and purchase, and against a backdrop of exposure which represents all of the existing perceptions and feelings a shopper has about the brands, retailers, and products in a category .

INTRODUCTION

We codified the messy middle as two looping mindsets: exploration and evaluation . Shoppers can cycle between these mindsets as many times as they need or want, exploring available products and brands before evaluating the suitability of a smaller set of choices .

If none of those products or offers turns out to be quite right, they simply seek more information and options . And if a purchase is habitual or impulsive, they might skip the messy middle altogether . Finally, once a decision and purchase is made, the shopper’s subsequent experience with the product and brand becomes part of their background exposure, ready for the cycle to start again.

We identified a set of mental shortcuts and rules of thumb that people typically employ while moving through the messy middle . In the course of validating these findings we ran observational studies and simulated shopping experiments across multiple categories in over 20 countries across the globe. This process revealed a powerful set of common principles that illuminate how shoppers navigate the tangled web between trigger and purchase before ultimately deciding what to buy .

1

Ensure brand presence

2

Close the gap between trigger and purchase

3

Employ behavioural science creatively and responsibly

Inspired by our research, we ultimately recommended three key strategies for marketers operating in the messy middle:

10 INTRODUCTION

‌New developments in the messy middle

In times of economic uncertainty, consumers are forced to do more with less, to search harder for value, and be more vigilant with the household budget. And marketers now face a similar challenge, often working under budget pressures while being tasked with improving (and proving) ROI

to deliver profitable growth. In this context, understanding more about how people make decisions could offer a critical competitive advantage. Our insights are equally applicable to established and challenger brands, equipping marketers with ideas and inspiration so they can help shoppers navigate and exit the messy middle . And in the process make their brand and product more likely to be chosen .

A new generation of advertising tools and solutions, an increasing number of them powered by artificial intelligence (AI), promise to make implementation and scaling much easier . In the context of digital

marketing, AI will allow brands to anticipate individual consumer needs, enabling them to better connect new customers with relevant content, across channels, and in real time. But it’s important to remember that while AI will be an invaluable partner, it isn’t a replacement for marketers’ knowledge and experience . What will continue to set great marketing apart is the human ingenuity, insight, and intelligence behind it.

This new report focuses once more on insights into the messy middle of purchase decision-making: the space between trigger and purchase .

This may make it a richer source of insight for the performance marketer than the brand marketer . But that isn’t to suggest that brands aren’t crucially important to decision-making. Our research shows that leading brands are able to retain a meaningful share of consumer preference even when real-world or invented competitor brands offer vastly superior propositions . In other words, brand continues to be one of the strongest heuristics impacting consumer choice .

11 INTRODUCTION

There are many routes through the messy middle and each one is different. But while the combination of exposure, trigger, and path is unique to each shopper, our research shows that there are important commonalities in how people search for information, assess competing propositions, build confidence about a product, and ultimately choose where they are going to make a purchase .

The role of marketing in the messy middle then becomes a story about reaching people where they are and boosting their confidence. It’s about being present and having broad coverage to maximise customer connections across as many relevant opportunities as possible, using creativity to build brilliant assets that help close the gap between trigger and purchase, and crafting compelling propositions that influence the products consumers choose to buy and where they buy them .

The exploration and evaluation loops of the messy middle are fundamentally about how people discover and digest information.

In our original report we introduced the notion of “supercharging”, by which we meant boosting a proposition with insights derived from behavioural science. In this report, we illustrate that by also “supercharging” their Search results, their owned content, and their presence at points of sale (the places consumers make purchases), marketers can improve ROI and growth, and make a strong case for increased investment .

Observing the messy middle

1

But they constitute very different mindsets. In our original research we described exploration as an expansive activity, adding brands, products, and category information to our mental portfolios. Evaluation, on the other hand, is about narrowing down until we have a small enough set of options to make a decision.

The evolution of information foraging

CONTENTS

Anthropologists have applied the idea of “information foraging” to explain how people typically handle the challenge of finding information online.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

From optimal foraging theory …

Information-foraging theory1 is inspired by optimal foraging theory, developed in the 1970s, which describes how animals look for food by quickly assessing physical locations known as “patches” for their

potential to provide sustenance. Cues, such as scent, are used to gauge the quality of a given patch. And if that location doesn’t seem like it will provide a good return on their energy expenditure, the animal moves on to the next one.

… to information-foraging theory

People use a strikingly similar technique when foraging for information. The information patches we explore might not be physical spaces,

but we look for cues and signals in the same way. Search results,

digital ads, social content, and website copy are all types of information patches. We quickly assess them for the amount of information we’re likely to gain, and if we don’t like the “scent” of a given patch, we’ll carry on and look for another.

The information-foraging techniques we rely on have to account for a rapidly evolving information environment. Before the internet, information was limited and mostly local. We could ask for advice about what to buy from friends and family, from sales reps if we visited a physical store

or picked up the phone, and possibly from media such as television and newspapers (but only if we were lucky and happened to see an ad

relevant to our needs at that moment in time). The range of sources was so much smaller, and there was beauty in this simplicity.

Pirolli, P., & Card, S. (1995, May). Information foraging in information access environments. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 51-58).

CHAPTER 1

OBSERVING THE MESSY MIDDLE

Some of that beauty derived from how much easier it was to make a decision. With a limited set of options and sources, it was sometimes possible to look at everything. This sense of examining a problem from all angles instilled confidence, so that when a choice was made, it was with the reassurance that we’d done the best we could with the information at hand.

An explosion in product choice and information has made it harder to feel confident about making the right decision

Contrast that with today’s world of superabundant information. It’s now a genuine challenge to exhaust all sources of information for even the simplest products. For those of us hunting for the absolute “best”

choice, as opposed to something just “good enough”, there is always one more expert review, one more page of customer comments, one more combination of search keywords that could potentially confirm a bargain or save us from wasting time and money on an inferior item.

And then there’s the explosion in product choice, with many categories on popular online stores offering literally thousands of options. Where we once entered a physical retailer with a limited consideration set of products and a limited set of brand perceptions, in some categories we now have seemingly endless choice and information. For many kinds

of purchases, it has become much harder to feel confident that we’re making the right decision.

Confidence turns out to be a crucial factor in successfully navigating the messy middle, and we’ll return to it in depth in Chapter 4. But

for now we’re going to focus on the tools people use to navigate the information landscape.

CHAPTER 1

‌OBSERVING THE MESSY MIDDLE

The evolution of search behaviours

In addition to their increasing abundance, information patches in the messy middle also tend to change more rapidly than their physical counterparts. Our original research was inspired in part by a simple observation: the number of options and the amount of information people have when choosing what to buy has increased dramatically. In turn, this makes the business of buying something more complicated than it used to be. And that observation started with an insight from Google Search.

The internet is going from a place where people compare prices to one where they compare everything

In the U.K., search interest2 for “best” has increased over time, while search interest for “cheap” has declined, with the trend lines crossing in 2008. The previous leader, “cheap” describes a simple characteristic that is easily measured; “best”, on the other hand, is nebulous and requires a lot more information to define. We hypothesised that the transition from one word to the other was symptomatic of the internet going from a place where people compare prices to one where they compare everything.

To measure search interest over time for a topic, Google Trends looks at the proportion of searches for that term to all searches on Google at that time and location.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

In the U.K., search interest for queries containing “cheap” and “best” has been trending in opposite directions3

cheap

Search term

best

Search term

100

0 2004

2013

2023

FIGURE 01

The terms “best” and “cheap” are examples of modifiers — words that add nuance or precision to a query. In our original research we looked at long-term trends of modifiers such as “best”, “cheap”, “reviews”, “ideas”, and “deals” across categories. But the transformation of modifiers can also reveal how the messy middle is evolving within a category, showing us that over time searches in many categories have become more nuanced, precise, and personal.

Google Trends, United Kingdom, Search interest for queries containing “best” and “cheap”, January 2004 - July 2023.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

Let’s use two examples from the fashion category for illustration. First we have U.K. search queries related to the term “wedding dresses”. Over time, these queries reflect more specificity about the product the user

is seeking, for example the colour and style, but also more nuance about who the dress is actually for.

In the U.K., popular search queries related to the term

“wedding dresses” show increasing specificity over time4

2016

dresses

wedding dresses

2019

wedding guest dresses princess wedding dress

wedding dress london

2022

wedding guest outfits for over 50s floaty dress for wedding guest

long sleeve dress for wedding guest

teal dress for wedding

FIGURE 02

Secondly, Search can help us understand how a category is maturing and evolving in the face of a broader macro trend — in this case the intersection of the fashion category with rising interest and

expectations relating to sustainability. In the U.K., searches in the apparel category related to the term “sustainable” demonstrate a burgeoning sophistication, as users progress from exploring the concept to adding brands, subcategories, and value-based modifiers to their queries.

Google Trends, United Kingdom, Popular search queries about the topic “dress”, January 2016 - July 2023.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

FIGURE 03

In the apparel category in the U.K., queries related to the term

“sustainable” can help us understand how the category is evolving5

Fashion mentioned

Brands start to appear

Diverse demands emerge

More specific modifiers emerge

2017

2018

2021

2022

sustainable clothing sustainable fashion

sustainable clothing sustainable shoes sustainable clothing UK sustainable clothing brands sustainable swimwear

[leading sportswear brand] sustainability

sustainable brands sustainable clothing brands sustainable swimwear sustainable jeans

sustainable dresses [leading clothing brand]

sustainable development goals [unisex clothing brand] affordable sustainable clothing best sustainable clothing brands

Note: The years featured were selected to show a noticeable change in consumer queries compared to previous years.

Google Trends, United Kingdom, Popular search queries containing “sustainability” in the Fashion category, January 2015 - July 2023.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

Shifting trends in Google Search show that consumer demand is a moving target — and marketers need to be able to respond in real time

While category demand (and therefore searches) will evolve over time in ways that reflect broader cultural and social trends, the COVID-19 pandemic created more radical shifts. To give an example, consider the search terms “dog friendly” and “hot tub” in the U.K. For years, these

How searches in a category can evolve to reflect broader cultural and societal trends6

dog friendly

Search term

hot tub

Search term

100

0 2013

2017

2022

FIGURE 04

two queries marched hand-in-hand, matching both short- and long-term trends almost exactly. Then, for the two years of the pandemic, they parted ways.

Google Trends, United Kingdom, Search interest for queries containing “hot tub” and “dog friendly”, January 2013 - July 2023.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

Before 2020, these two terms were often used by people planning holidays. Some wanted to know if their destination hotel had a hot tub and some wanted to know if it would accommodate canine guests. But when pandemic-related travel restrictions put holiday plans on hold, one term sank while the other soared. U.K. search interest for “hot tubs” rose to its highest level to date, as more people explored getting one for their own home. In doing so they caused the term to somewhat shift category from travel to home and garden. Once restrictions were lifted “hot tubs” shifted back towards the travel category and the two terms returned to their pre-pandemic correlational relationship.

In fact, the pandemic inspired short-term fluctuations in several categories. At the start of lockdown, U.K. search interest for “can you freeze” suddenly spiked as people began to respond to anxieties about food availability. While the spike was short-lived, it was followed by a larger than usual seasonal increase in search interest for “fridge freezer” over the summer months. For anyone playing close attention at the time, the spike may have represented an opportunity for brands to

mozzarella cheese

gnocchi

donuts

can you freeze...

salmon

mac and cheese

banana bread

brussel sprouts

Words commonly paired with the search term “can you freeze” in the U.K. during the COVID-19 pandemic7

FIGURE 05

yeast

meet consumer information needs, or hinted at the category interest that followed.

bagels

Google Trends, United Kingdom, Popular additions to queries containing “can you freeze”, January 2020 - December 2021.

CHAPTER 1

OBSERVING THE MESSY MIDDLE

During the COVID-19 pandemic, U.K. search interest for

“can you freeze” peaked in the early months of the first lockdown8

can you freeze fridge freezer

Search term Search term

100

Feb. 2020

March 2020

April 2020

May 2020

June 2020

July 2020

FIGURE 06

These examples demonstrate that consumer demand, as expressed through Search, is a moving target. Over time, people are getting better at searching for what they want, and are more confident that highly specific queries will return relevant results. Brands are improving too, appearing

in the right places with relevant messaging and content. But marketers still need to keep an eye on evolving trends and invest in tools that enable them to respond in real time whenever these unpredictable pockets

of demand appear. Again, this is an area where AI promises to help by enabling marketers to make confident decisions driven by predictive data.

Google Trends, United Kingdom, Search interest for queries containing “can you freeze” and “fridge freezer”, January 2020 - July 2020.

CHAPTER 1

‌OBSERVING THE MESSY MIDDLE

From foraging to searching

Humans are creatures of habit. Once an information patch demonstrates a high degree of reliability, we’ll return to it often. In fact, the way we engage with it may evolve and become more sophisticated as our familiarity increases.

People are complicated and the way we search for information on the internet is complicated. No two people are the same, and no two instances of decision-making are the same. And with the ever-accelerating rate of change as a multiplying factor, it’s reasonable to ask: how can marketers hope to keep up with so many different needs and such a high degree of uncertainty?

In the next chapter we’re going to take a closer look at how marketers can use behavioural science to stand out in a fast-moving information landscape.

CHAPTER 1

1Chapter summary

‌OBSERVING THE MESSY MIDDLE

The ways that people explore and evaluate information online mirrors the ways that animals forage for food. Both involve looking for cues and signals that quickly make clear whether a location is likely to contain what they’re looking for.

The techniques used in foraging for information have evolved in line with a rapidly shifting and evolving Search environment. People are looking for more nuanced and specific information about the products they’re thinking of purchasing, and this is reflected in the growing complexity of their search queries.

In the previous chapter we introduced a concept from academic literature: information-foraging theory. A key consideration in the theory is efficiency: people prefer ”patches” that provide the largest amount

Marketers need to keep an eye on evolving trends and invest in tools that enable them to respond in real time, whenever unpredictable pockets of demand appear.

Searching the messy middle

2

of relevant information for the least amount of time and effort.

Over the past 25 years, Google Search has become a familiar starting point for many information-dependent activities, including online shopping. The popularity of Search presents marketers with an opportunity to connect with billions of consumers at the precise moment when a want or need solidifies into a query. That in turn presents consumers with the opportunity to quickly review product propositions from multiple brands in one place.

CHAPTER 2 S UMMARY<�H�H�R=�?h�H �H �Hdent: 0pt;line-height: 121%;text-align: left;">In our original shopping experiments, we saw how the application of behavioural science helped participants explore and evaluate the information contained in different product propositions. Since the components of a well-crafted Search ad convey similar details about brand and offer, we decided to explore a related question: can behavioural science make a difference within the specific parameters of Search advertising?

25 CHAPTER 2

‌SEARCHING THE M ESSY M IDDLE

Evolving from theory

Tip: Every time we introduce a new behavioural principle we’ll provide a brief description of what it means. For a full definition of all the principles we reference in our research, visit the Behavioural science glossary at the end of the report.

The shopping experiments conducted for our original research measured how various behavioural science principles affected shopper preference. There are hundreds of potentially applicable principles that could be used to benefit consumer decision-making, but our literature review and shopper observations highlighted six that had significant potential to influence across both exploration and evaluation.

Six key behavioural principles often associated with consumer

decision-making in the messy middle

Hotel Brand

Certiﬁcate of excellence from [travel review company]

1,200 reviews

Free room upgrade

All-inclusive

Only two spaces remaining

Book now to get the best deal

FIGURE 07

CHAPTER 2

SEARCHING THE MESSY MIDDLE

Authority bias: the tendency to attribute greater accuracy and knowledge to the opinion of a perceived authority figure, unrelated to its content, and be more influenced by that opinion.

Power of free: the tendency to overvalue free products. It assumes that when a product becomes free, its intrinsic value for consumers increases.

Scarcity bias: based on the economic principle that time-, quantity-, or

access-limited resources are more desirable.

Social proof: the tendency to follow the opinion, advice, and behaviour of others in one’s own decision-making.

Category heuristics: shortcuts or rules of thumb that aid us in making a quick and

satisfactory decision within a given category.

Power of now: the tendency to focus on today rather than think about what tomorrow might bring. We discount the future in favour of today.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

Through those original survey-based experiments, we demonstrated across a range of categories that shoppers would be willing to shift away from their first-choice brand to a second-choice brand — if that second-choice brand made a compelling marketing case based on responsible use of the six behavioural principles. And when maximal expressions of those principles were put in place (think five-star ratings and third-party endorsements), a significant number of shoppers across

In the U.K., a significant number of shoppers will switch from their first-choice brand to an invented brand that’s been “supercharged” with the six key behavioural principles9

First-choice brand Supercharged invented brand

Note: The eagle-eyed among you may notice that some of the bars don’t add to exactly 100% — don’t worry, it’s not an error, but due to rounding to whole numbers.

FIGURE 08

48%

62%

63%

every category switched not just from a preferred household-name brand to their second-choice brand, but to an invented brand we created purely for the experiment.

28%	43%	44%	49%	50%	50%	50%	51%	53%	60%	60%					64%	65%	71%	73%	73%	75%	75%	76%	78%	79%	79%	82%	86%	87%
72%
	57%	56%
			51%	51%	50%	50%	49%
								47%
									40%	40%
											37%	37%	37%	37%	36%	35%
																	30%
																		27%	27%
																				25%	25%	25%
																							22%	21%	21%
																										18%
																											14%	13%

Cereal Flights (long haul)

Car Whisky Detergent Smartphone Face moisturiser

Mobile network provider

Cat food Clothing Children’s toy

Laptop Cinema Shampoo

Fitted kitchen Mortgage Power drill

Flights (short haul)

Make-up Energy provider Broadband provider Package holiday

Credit card Car hire Hotel Sofa ISA

Bathroom suite Mountain bike

Car insurance

52%

38%

Google/The Behavioural Architects, U.K., Decoding Decisions: Making sense of the messy middle, N=31,000 in market online shoppers, 2019.

CHAPTER 2

‌SEARCHING THE MESSY MIDDLE

Experimenting with Search Ads

Unlike the side-by-side comparison format we used for our original research, Search Ads have specific requirements and limitations.10

FIGURE 09

In our new research, we introduced ranking into our simulation that looked like a real Search results page

Original experiment design

Brand 1 Brand 2

Certiﬁcate of excellence from [travel review company]

400 reviews

Recommended

by [travel magazine] 400 reviews

Free room upgrade All-inclusive Free travel insurance Flexible booking

Only 10 spaces remaining

Book now to get the best deal

Limited spaces available Book this week to get the best deal

Experiment design for Google Search environment

hotels

[First-choice brand] - Hotel

FREE 50 page home style inspiration guide. The travel editor’s favourite. We know we aren’t the cheapest, but our rooms are proper plush. UK’s favourite for 18-24 yr olds.

Rating: 4.9 200 reviews

www.inventedbrand.com

Package holidays - [Invented brand]

A comfortable stay. The celebrity favourite. Sponsors of [U.K. travel show]. The UK’s favourite.

Rating: 5 1250 reviews

[Second-choice brand] - Great value, quality hotel

FREE 20 page top holiday destination guide. [Top travel guide]'s favourite. We know our rooms are cosy, but our rates are low. Honestly, how long will you spend in the room anyway! UK’s favourite for 18-24 yr old couples. Sponsor of of [U.K. travel show].

One of those limitations is how many characters advertisers can use. Search Ads have strict character limits, but our experiments sometimes exceeded these in headlines and descriptions. This was by design to measure the impact of multiple behavioural science principles.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

Supercharging ad copy can have a positive effect for both established

and challenger brands

1 2

hotels

2nd Favourite Brand - Great value quality hotel

FREE 50 page hotel destination guide. [Top travel guide]'s favourite. We Know our rooms are cosy, but our rates are low. *Honestly how long will you spend in the room anyway!*.

UK’s favourite for 18-24 yr old couples. Sponsor of [top U.K. travel TV show].

Rating: 5.0 1250 reviews

FIGURE 10

So for this experiment, while we kept three of the original six principles, we also added three new ones that could be easily articulated as text or using specific, opt-in features of Search Ads.

Order effect: the difference in response as a result of the order in which stimulus is presented, such as listing the brand versus benefit first

in the ad headline.

Costly signalling: signage that conveys information about the characteristics of the product, service or brand itself, but these characteristics must be costly for the company in order to signal wealth, such as

prime-time television slots or premium sponsorships.

Pratfall effect: faults

or flaws can make products more appealing, given that imperfection is seen as natural and more relatable.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

For example, seller ratings allow advertisers to leverage social proof in the form of both a star rating and a review count displayed in the ad. Similarly, leveraging your own first-party data in ad copy could help to personalise social proof in the form of recommendations specific to different groups of users. For a full definition of social proof and the other principles tested in this experiment, visit the Behavioural science glossary at the end of the report.

Other expressions were chosen not because they map to a specific feature of Search Ads, but because they lend themselves to a

text-only environment. Authority bias and costly signalling can both be expressed by quoting endorsements and sponsorships, assuming a brand can legitimately cite either of these factors. A free gift or service (think free shipping) is usually easy to articulate, and in the case of a simple destination guide also relatively easy to produce.

Beyond these textual expressions of our original principles, we wanted to know if shoppers respond differently depending on whether a brand or a benefit is mentioned first.

Expressions were chosen not because they map to a specific feature of Search Ads, but because they lend themselves to a

text-only environment

Another text-friendly principle we decided to test is the pratfall effect. This involves deliberately highlighting an imperfection in order to cast light on an adjacent benefit. In the example above, the hotel’s rooms are described as cosy — an admission that their square footage is small. But the ad goes on to emphasise value by reminding travellers that they’re probably not going to spend much time in the room anyway. In other words, an apparent imperfection, when framed correctly,

can humanise the brand.

CHAPTER 2

‌SEARCHING THE MESSY MIDDLE

So while the example above might look like a typical Search Ad, it has been carefully engineered to include a range of behavioural principles that we felt were most relevant to text-based ad formats. But did it earn a greater share of clicks?

Well, as learned in our original research, behavioural science can have a big impact.

Search Ads: Supercharged

Our Search experiment was mocked up to look and feel like a real Search results page (though it should be noted that the experiment was not conducted on the Google Search domain and did not use ads from existing advertiser campaigns).

Participants were asked to select a first-choice brand and second-choice brand in each category. They were then shown three Search Ads featuring their first-choice brand, second-choice brand, and an invented challenger brand. The ads contained different combinations and executions of the six behavioural science principles. We asked participants to choose which of the three they would click on and measured how often the second-choice brand or invented brand were

chosen. This allowed us to quantify the ability of the selected behavioural science principles to redirect preference away from the first-choice brand.

We set up our experiments to explore whether our Google Search simulations aligned with the findings from our original research:

There’s value in simply showing up when shoppers have a choice to make.

Supercharging with behavioural science can shift consumer preference.

Even unknown brands can win share in the messy middle with the right approach.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

Here’s what we found:

Showing up still works

One of the most revealing findings from our original research was that brands could gain preference share simply by showing up and giving shoppers a choice to make. Second-choice brands and invented brands didn’t even have to offer anything different — an identical proposition was enough to make some shoppers switch allegiance.

We can simulate what happened when the first- and second-choice brands were in

first and third position, with identical propositions

hotels

[First-choice brand] - Great value, quality hotel

FREE 20 page home style inspiration guide. The travel editor’s favourite. We know we aren’t the cheapest, but our rooms are

*proper plush*. UK’s favourite for 18-24 yr olds.

Rating: 4.9 200 reviews

www.stayinnhotels.com

Stay Inn Hotels - Hotel

A comfortable stay. The celebrity favourite. Sponsors of [U.K. travel show]. The UK’s favourite.

Presence

of second-choice brand in third position, with identical ad copy

[Second-choice brand] - Great value, quality hotel

FREE 20 page home style inspiration guide. The travel editor’s favourite. We know we aren’t the cheapest, but our rooms are

*proper plush*. UK’s favourite for 18-24 yr olds.

Rating: 4.9 200 reviews

FIGURE 11

To test this effect in Search, we used the findings of our experiment to simulate the effect of identical propositions in the first and third positions of the Search results.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

In the U.K., 15-34% of participants across all categories would have clicked on their second-choice brand when shown with an identical

proposition to their first-choice brand11

First-choice brand

Second-choice brand

66%

34%

69%

31%

69%

31%

70%

30%

70%

30%

27%

73%

27%

73%

27%

74%

26%

74%

26%

85%

15%

26%

74%

FIGURE 12

In this hotel example, setting aside the invented brand in second position, 74% of participants would have picked their first-choice brand in first position, but 26% would have clicked on their second-choice brand down in third position. The effect was visible across all the categories we tested. Even with no difference in copy between ads other than the brand, 15%-34% of participants would have clicked on their second-choice brand when it was shown in third position.

Smartphone

SUV

Hotel

Whisky

Mortgage

Broadband

Clothing

Shampoo

Moisturiser

Holiday

Mascara

Google/The Behavioural Architects, U.K., Decoding Decisions: Supercharging Search Ads in the messy middle, N=12,000 in market online shoppers, 2021.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

There could be several reasons for this behaviour — for example, people may be largely ambivalent between their top two brands. Regardless of the reason, all the second-choice brand needed to do to win share from the first-choice brand was to show up.

Applying behavioural science can boost across categories

In our original research we introduced the idea of “supercharging”, where second-choice brands and invented brands were given more powerful expressions of the tested behavioural science principles (think 1,250 reviews versus 200 reviews) to maximise their ability to win share of preference. Even with fairly unremarkable applications of behavioural science, in a simulated environment with little scope for creative expression, supercharged second-choice brands proved capable of winning over significant numbers of shoppers.

To test the impact of supercharging on Search Ads we kept the

We can simulate what happened when the second-choice brand was in third

position, with supercharged ad copy

hotels

[First-choice brand] - Hotel

FREE 20 page home style inspiration guide. The travel editor’s favourite. We know we aren’t the cheapest, but our rooms are

*proper plush*. UK’s favourite for 18-24 yr olds.

Rating: 4.9 200 reviews

www.stayinnhotels.com

Package holidays - Stay Inn Hotels

A comfortable stay. The celebrity favourite. Sponsors of [U.K. travel show]. The UK’s favourite.

Presence of second-choice brand in third position, with supercharged copy

FIGURE 13

[Second-choice brand] - Great value, quality hotel

FREE 50 page top holiday destination guide. [Top travel guide]'s favourite. We know our rooms are cosy, but our rates are low.

*Honestly, how long will you spend in the room anyway!* UK’s favourite for 18-24 yr old couples. Sponsor of [top U.K. travel TV show].

Rating: 5.0 1250 reviews

second-choice brand in third position, but amped up all the behavioural science expressions.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

When the behavioural science principles were charged to a higher degree, second-choice brands proved capable of seizing significantly more preference share away from first-choice brands. Even smartphones, a category with particularly strong brand loyalty, saw almost a quarter

of people choosing to click an ad for their second-choice brand.

In the U.K., 23-53% of participants would have clicked on their second-choice brand across all categories when behavioural science principles were charged to a

higher degree12

First-choice brand

Supercharged second-choice brand

50%

51%

49%

51%

49%

51%

49%

54%

46%

55%

45%

56%

44%

56%

44%

56%

44%

59%

41%

77%

23%

53%

47%

FIGURE 14

In the hotel category, supercharging lifted second-choice preference from 26% (share of clicks from just showing up) to 53% — more than doubling its share. In fact, hotels proved to be the category most responsive to supercharging with behavioural science.

Smartphone

Whisky

SUV

Broadband

Mortgage

Clothing

Shampoo

Mascara

Holiday

Moisturiser

Hotel

Google/The Behavioural Architects, U.K., Decoding Decisions: Supercharging Search Ads in the messy middle, N=12,000 in market online shoppers, 2021.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

Supercharging can level up challenger brands

The invented brands we created in our original experiments to simulate the impact of using behavioural science with

FIGURE 15

In our original research, we wanted to understand what happens when a completely new challenger brand enters the market. So we invented some brands for just that purpose. To measure this effect in Search, we introduced them again here.

The purpose of creating Search Ads for these invented brands was to simulate the impact of using behavioural science on a new market entrant: a brand that no one had heard of before, with no previous

associations to cloud their judgement. To give the challengers a slight leg-up, let’s view the results when the invented brand was in second position, sandwiched by the first-choice brand in first position and the second-choice brand in third.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

We can simulate what happened when the supercharged invented brand

was placed in second position

hotels

[First-choice brand] - Hotel

FREE 20 page home style inspiration guide. The travel editor’s favourite. We know we aren’t the cheapest, but our rooms are

*proper plush*. UK’s favourite for 18-24 yr olds.

Rating: 4.9 200 reviews

Invented brand

in second position, with supercharged copy

[Second-choice brand] - Package holidays

A comfortable stay. The celebrity favourite. Sponsors of [U.K. travel TV show]. The UK’s favourite.

FIGURE 16

www.stayinnhotels.com

Stay Inn Hotels - Great value, quality hotel

FREE 50 page top holiday destination guide. [Top travel guide]'s favourite. We know our rooms are cosy, but our rates are low.

*Honestly, how long will you spend in the room anyway!* UK’s favourite for 18-24 yr old couples. Sponsor of [top U.K. travel TV show].

Rating: 5.0 1250 reviews

It’s worth noting that in the Search experiment, invented brands had to rely on names alone to draw shoppers in. Real-world brands, on the other hand, may have benefited from shoppers’ existing awareness of their visual identity and any associated emotional response. In other words, the effects resulted solely from the strength of the invented brands’ textual propositions and the behavioural science being applied.

To return to the example of hotels, a supercharged invented brand was able to take a 34% share from a non-supercharged first-choice brand. Bearing in mind the first-choice brand will have enjoyed the benefit of name recognition, this is an impressive achievement. While they varied

in degree, supercharged invented brands across all categories were able to take share from first-choice brands.

CHAPTER 2

‌SEARCHING THE MESSY MIDDLE

In the U.K., 14-37% of participants would have clicked on an invented brand when

it was supercharged with powerful expressions of behavioural science13

First-choice brand

Supercharged invented brand

32%

63%

65%

66%

68%

69%

70%

73%

74%

75%

85%

86%

37%

35%

34%

32%

31%

30%

27%

26%

25%

15%

14%

FIGURE 17

Smartphone

Clothing

Shampoo

Holiday

Broadband

Mascara

Whisky

SUV

Moisturiser

Hotel

Mortgage

Implications for Search

There are implications here for both established and challenger brands.

For established brands with turf to defend, supercharging ad copy with behavioural science can protect hard-won consumer preference and expensive brand investment. For challenger brands, these results should be cause for optimism: even if your Ad Rank (including your bid) doesn’t currently reach first position, a strong proposition and ad copy

supercharged by behavioural science may allow you to compete and win a greater share of clicks in the messy middle.

Google/The Behavioural Architects, U.K., Decoding Decisions: Supercharging Search Ads in the messy middle, N=12,000 in market online shoppers, 2021.

CHAPTER 2

SEARCHING THE MESSY MIDDLE

Supercharging ad copy can have a positive effect for both established and challenger brands

Optimising ads with AI

Optimising for an information patch as broad as Search might seem daunting, but this is luckily one of the areas where generative AI —

AI systems trained to produce new copy, images, and other kinds of content — is already making a difference. Marketers know their brands best, and can engage with AI using natural language to write and refine ad copy infused with behavioural science at scale.

But attention earned by great ad creative can still be diminished by poor destination content. In the next chapter we’re going to explore how behavioural science can be used to build and support confidence once shoppers begin to engage with your owned media.

2Chapter summary

40 CHAPTER 2

‌SEARCHING THE M ESSY M IDDLE

Behavioural science helps us understand how people explore and evaluate in the messy middle. Marketers can influence consumer decision-making by responsibly applying these principles in their propositions and ad copy.

Our research found that by supercharging Search Ads with compelling expressions of behavioural science, shoppers across every category switched not just from a preferred brand to their second-choice brand, but to an invented brand we created purely for the experiment.

Once a shopper expresses interest, for many marketers the next logical question is: how do we boost their purchase consideration? But purchase consideration is something that matters to marketers, not consumers.

For established brands, supercharging ad copy with behavioural science can protect brand investment and maintain your position as a top choice for consumers. For challenger brands, supercharging your ad copy can help level the playing field and enable you to win a greater share of clicks.

Boosting confidence in the messy middle

3

Let’s take ourselves out of a marketing mindset and spend some time thinking like consumers.

For consumers, confidence is key in making a decision. We want to feel confident that the choice we’re about to make will address our needs. That it’s worth the resources we’ve expended getting to that point. That it will live up to expectations and not waste money. We might even want our purchase to reflect something positive about ourselves.

In other words, we want to feel confident that the choice we’re making is the best one, or at least not a poor one.

On the flip side, we might also want to feel confident that we’re avoiding any possible negative consequences, such as getting into an agreement for a service that isn’t trustworthy, that might waste our time and money, or that cannot easily be cancelled.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

For consumers, confidence is key in making a decision — we want to feel confident that the choice we’re about to make will address our needs

During our research we explored the degree of confidence shoppers said they felt about various products. And we discovered that the level of baseline confidence differs between categories. Across the categories included in our in-market studies, we found that fitted kitchens were the product people felt least confident about. This makes intuitive sense, as a kitchen is an expensive and infrequent purchase that shoppers generally have to live with for a long time. At the other end of the scale, shoppers felt most confident about breakfast cereal, an inexpensive, frequently

FIGURE 18

In the U.K., consumer confidence tends to be higher when purchase frequency is high or cost of product is low14

Low confidence

High confidence

Fitted kitchen

Cereal

Note: Grey circles represent the 31 different categories tested in our U.K. survey.

purchased product with relatively low stakes for making the wrong choice.

From this we can deduce that consumer confidence tends to be higher when the permanence of a purchase is low or when the cost of a product is low.

This confidence can reduce the information required to reach a decision by tapping into existing product understanding and prior experience.

Google/The Behavioural Architects, U.K., n= 600 prospective online purchasers per category, ages 18-75, January 2023.

CHAPTER 3

‌BOOSTING CONFIDENCE IN THE MESSY MIDDLE

We also found that confidence was strongly correlated with knowledge. Product categories where shoppers said they had a lot of knowledge tended to be categories where they reported a high degree of confidence. Bearing in mind what we’d already learned about how people forage for information, we realised that this was an area where we needed to dig deeper.

From information needs to product knowledge

Across the twin mindsets of exploration and evaluation, shoppers seek to address a range of information needs. Some of these are specific to each mindset, while a small number exist at the intersection of the two.

FIGURE 19

Consumers are looking for specific information needs in both the exploration and evaluation mindsets of the messy middle

Trigger

Range of options

Brands and retailers in category

Price range Inspiration

Contextual category information

Company ratings and reviews

Policies and warranties Expert guidance Product features

and benefits

Availability Delivery terms Price

Payment plan

Product rating and review

Purchase

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

When we’re in an exploration mindset

… we want to understand the brands and retailers in the category, what range of products might be available, and the range of prices on offer, and maybe find some inspiration. This helps us understand what is available and what we need to keep

in mind while making our purchase decision.

When we’re in an evaluation mindset

we are able to get our preferred product or service, and if what we’re considering is right

for our needs.

A brand’s owned content plays a particularly important role in providing information during both exploration and evaluation. Home pages and product pages engage both mindsets of the messy middle, as shoppers gain confidence in their understanding of what’s available before weighing up the suitability of specific products.

CHAPTER 3

‌BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Different building blocks for different categories

At the product level, we found that many information needs were common across categories. But our research also showed that shoppers prioritise different kinds of information for different products. For example, participants indicated that they were much less interested in guidance from industry experts when searching for information on high-frequency and low-risk categories, such as cereal, than for higher-stakes products that are infrequently purchased, like fitted kitchens.

At the same time, shoppers needed less information (they had a lower information threshold) when shopping for products that they enjoyed, such as make-up, than when shopping for less enjoyable products, like broadband. We hypothesised that this might be explained by shoppers having higher prior knowledge and existing exposure among categories they enjoy browsing.

Overall, we saw that products perceived as having higher stakes, by virtue of being more expensive or infrequently bought, seem to have higher information needs than the other categories. From this we can hypothesise that people might spend more time gathering information about these purchases, increasing the amount of time they spend in the messy middle.

We were able to validate these insights further using data from Google Search in the U.K. For example, when shoppers search within the Air Travel category they are much more likely to look for pricing information than any of the other types of need shown in the chart on the next page. Again, this makes intuitive sense, as for most flights the destination is fixed and the carriers are to some degree interchangeable — making price the most pressing variable. In contrast, the Laptop category sees more shoppers searching for advice in the form of expert opinions and reviews from

other users, which makes sense for a product with many more technical dimensions to consider.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

In the U.K., popular search modifiers suggest that the information needs of consumers in the messy middle vary by category15

Inspiration (includes search terms such as “for” + “ideas”) Offline interactions (includes search terms such as “near me” + “store” + “shop” + “shops” + “stores”)

Online interactions (includes search terms such as “delivery” + “pick up” + “customer service” + “return” + “online” + “app”) Price (includes search terms such as “cheap” + “price” + “deal” + “discount” + “offer” + “cheapest”)

Review (includes search terms such as “best” + “compare” + “review” + “vs”)

100

Air Travel

Grocery and Food Retailers

Flowers

Apparel

Laptops and Notebooks

FIGURE 20

Relative search interest

All of this information is valuable and shoppers invest a lot of time in finding, distilling, and digesting it. But you can still have too much of a good thing. Even with the information-foraging techniques at our disposal, overabundant information can create barriers that prevent

us from completing — or even starting — a purchase journey. Creating and maintaining engagement is critical to guiding potential buyers past these barriers. The beginning of a purchase journey can feel like facing a mountain of information. And as we’ll see in a moment, depending on the product, we don’t always have a lot of incentive to climb.

Google Trends, United Kingdom, Relative average search interest of terms about various information needs across 12 months. Inspiration consists of “the queries containing “for” OR “ideas”, Reviews consists of “best” OR “compare” OR “review” OR “vs”, Price consists of “cheap” OR “cheapest” OR “price” OR “deal” OR “discount” OR “offer”, Offline interactions consists of “near me” OR “shop” OR “shops” OR “stores” OR “stores” Online interactions consists of “delivery” OR “pick up” OR “customer service” OR “return” OR “online” OR “app”,
June 2022 – May 2023.

CHAPTER 3

‌BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Getting stuck in the messy middle

In addition to information needs, we identified a range of different psychological barriers that hamper progress through the messy middle. Many barriers are subconscious, but this isn’t always the case. Some are more applicable to the exploration mindset while some are more related to evaluation. A few are relevant to both.

Something as simple as a perceived lack of time can present a major barrier depending on how long a decision seems likely to take. Related to this idea is decision fatigue, which often surfaces in the evaluation

FIGURE 21

Psychological barriers can hamper progress in both the exploration and evaluation mindsets of the messy middle

Trigger

Lack of interest Lack of knowledge

Perceived lack of time

Choice overload Fear of missing out

Indecisiveness

Lack of urgency Seeking validation Risk aversion

Decision fatigue

Purchase

mindset, when shoppers begin to feel weighed down and averse to having to make yet another decision. And indecisiveness often manifests when the shopper has almost made up their mind, but cannot bring themselves to make the final call.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Just like information needs, barriers can differ between products. As mentioned earlier, user confidence tends to be higher for products where existing product knowledge is high, or where purchase permanence or the cost of the product is low. But the reverse is also true, with unfamiliar, long-term, or high-value purchases necessitating a higher information bar, creating a higher likelihood of barriers appearing.

Another factor in the appearance of barriers is how we feel about the product we’re thinking of buying. In categories we enjoy more, we seem less likely to be critically blocked by psychological barriers and may even stay in the messy middle for a prolonged period for the sheer pleasure of it. Conversely, in a category we don’t enjoy, we might drop out of the messy middle as soon as we encounter a barrier. And yet sometimes a decision has to be made, especially where essentials are concerned.

The degree to which U.K. consumers claim to enjoy shopping varies widely by category16

High enjoyment

Make-up

Low confidence

High confidence

Gas or electricity provider

Low enjoyment

FIGURE 22

After all, shopping for car insurance might not provide an endorphin buzz, but then neither does getting pulled over without a valid policy.

Google/The Behavioural Architects, U.K., n= 600 prospective online purchasers per category, ages 18-75, January 2023.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

The presence of other factors can raise barriers even when pleasure is involved. Buying an engagement ring or a new car should prompt positive emotions, but the cost and consequence of the purchase is still likely to provoke indecision, risk aversion, and choice overload, among others.

Finally, there is the perceived permanence of the purchase. How long we have to live with a decision also makes a difference to how likely we are to end up stuck in the messy middle. Our information needs are different if we’re buying something like a sofa, which we expect to hang on to for a

In the U.K., the perceived permanence of a purchase varies by category, further influencing the types of information need and barriers at play17

Enjoyment

Face moisturiser

Make-up

Package holidays

SUV car

Clothing Whisky Long-haul flight Smartphone Budget hotel

Children’s toy

Cinema tickets

Sofa

Mountain bike

Laptop

Confidence

Short-haul flight

Cereal

Power drill

Fitted kitchen

Bathroom suite

Mobile network provider

Shampoo

ISA account

Detergent

Cat food

Credit card

Mortgage

Broadband provider

Car insurance

Gas or electricity provider

Note: Size of circles represents the permanence of the product or service

— how long customers anticipate having it or needing it.

FIGURE 23

decade or more, compared to a bottle of shampoo that will be replaced in a matter of weeks.

Google/The Behavioural Architects, U.K., n= 600 prospective online purchasers per category, ages 18-75, January 2023.

CHAPTER 3

‌BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Building confidence

Confidence is a key factor in determining when shoppers feel ready to exit the messy middle and complete a purchase. As we acquire more information about a category, our confidence grows, though the process isn’t purely accumulative. There are multiple steps in most customer journeys, and if the quality of information degrades from one step to the next, confidence can be quickly eroded.

After exploring the relationship between information needs and confidence, we wanted to understand what additional role behavioural science can play when applied to a brand’s owned content. Our expectation was that the ways in which offers and propositions are articulated in a brand’s own assets are every bit as important as how they are presented in the ads and listings encountered elsewhere in the messy middle. But as shoppers move back and forth between exploration and evaluation, which behavioural science principles have the biggest impact on their level of confidence?

As we acquire more information about a category, our confidence grows, but if

the quality on information degrades then confidence can be quickly eroded

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Confidence boosts

To understand how brands can help build confidence by infusing their content with behavioural science, we took three of our invented brands and experimented with different ways of presenting information on websites we created for each of them.

Miller (sofa brand): a low confidence category — an expensive, long-term purchase.

Hyskeir (whisky brand): an average confidence category — and not considered a long-term buy.

4sure (car insurance brand): a higher than average confidence category — but definitely a necessity rather than a longed-for buy.

Most of our research into the messy middle has involved a mix of real and invented brands, but where websites are concerned, it’s hard to control for people’s prior experience. A good user interface (UI) can be undermined by bad customer service, while a strong brand can cover for any number of deficiencies in website experience. By using only invented brands, we were able to observe the impact of behavioural science on information and confidence isolated from existing brand perceptions.

Confidence to explore

We created home pages for our invented brands to measure the role of behavioural science in building confidence during the exploration

phase. The pages included different amounts of information, presented in different ways, to which varying expressions of behavioural science were applied. Participants were asked to choose which variation gave them most confidence about making a purchase in the category. Note that participants were not asked about their confidence in buying from the website owner’s brand, just about the category in general. Participants could also select “neither of the two” if neither of the options made them feel confident — which in practice might equate to a shopper diving back into the messy middle by returning to the Google Search results page.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Q. Which website makes you feel more confident in general about purchasing a sofa?

NEITHER

FIGURE 24

Let’s take the invented Miller sofa brand as an example. A sofa purchase tends to be relatively rare, relatively expensive, and something we plan to live with for a reasonable amount of time. So the amount of information needed to make us feel confident about making a purchase is likely to be quite high.

At the most basic level, we might expect a home page to contain a few non-negotiable pieces of information. Let’s think of this as the Bronze version of the site:

Branding and details about the brand

Imagery and visual representation of the brand

Indication of the range of products on offer

Information about the product category

These elements are all connected to the various exploration phase information needs we identified earlier.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

The Bronze version of the home page we created for an invented sofa brand featured only the most basic information required

FIGURE 25

To begin with, we enhanced the page by ensuring more of these key information needs were addressed. We added an endorsement badge, an average customer review score, and information about delivery.

In addition, we increased cognitive ease by surfacing more high-level information likely to be already present on the site and making it more accessible. This involved tweaking hero text and imagery, and adding a price range scale so people could quickly assess the affordability of Miller sofas.

We can think of this as the Silver version of the site: basic information, polished and presented as clearly as possible, but without needing much additional effort or investment.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

For the Silver version of the invented sofa brand home page, we utilised several behavioural science principles

Social proof

User ratings surfaced, with good but not great avg. rating

Emotional priming

Hero imagery upgraded to encourage more of an emotional response

Authority bias

Endorsements from a reputable website

Delivery friction

Delivery info surfaced, with slow option only

Framing

Hero text upgraded to frame the information in a more appealing way

Anchoring

Price ranges surfaced to help people shop within their budget

Paradox of choice

A buying guide to help shoppers avoid information overload

Cognitive ease

Easy-to-digest options, but without imagery

FIGURE 26

Viewed through the lens of behavioural science, we can see that the way information has been arranged on the Silver version utilises several of the principles explored across our other research. And for this experiment we also utilised a few more.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Authority bias: the tendency to give weight to a source of information because of its perceived expertise. Represented here by endorsements from reputable websites.

Framing: our preferences are shaped by how information is presented, even if the choices before us ultimately lead to the same outcome. We explored this by varying the hero text to provoke different kinds of emotional response.

Emotional priming: the tendency

to be influenced by how we feel emotionally rather than purely rational considerations.

We explored this by varying the hero image on the home page to show a blank background, a domestic background, and imagery of a family.

Cognitive ease: we tend to favour decisions that require minimal cognitive effort.

We enhanced this by making the options more easily digestible.

Social proof: the tendency to follow and copy the behaviour and actions of other people in situations of ambiguity or uncertainty. Represented here by user

ratings and reviews.

Paradox of choice: having unlimited choice can actually lead to indecisiveness, dissatisfaction, and regret. We addressed this by providing a tool or a catalogue to help

shoppers avoid information overload.

Anchoring: we look for reference points (anchors) that we can rely on and adjust our judgements and decisions from. Anchors may be based on price, past experience, occasions, or social norms. We applied by indicating price ranges on the home page to help people shop within their budget.

Delivery friction: reducing the time and/ or effort (friction) in which the consumer would receive an item they are thinking

of purchasing. This can be thought of as analogous to power of now, which we

For a full definition of all the principles we reference in our research, visit the Behavioural science glossary at the end of the report.

explored in some of our other experiments. Examples would be reducing delivery times, or offering the availability for the consumer to pick up the item themselves.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

And how did the Silver version perform compared to Bronze? Pretty well, as it turns out. In fact, twice as many participants selected Silver over Bronze. And a lower proportion chose “neither” when shown the Silver option, meaning fewer people were likely to leave the site and look

Twice as many participants selected the Silver version of the home page over the Bronze,

showing that basic improvements can provide shoppers with a significant confidence boost18

NEITHER

61%

30%

FIGURE 27

elsewhere. This shows the value of simply optimising the information that shoppers are looking for, with basic improvements leading to a significant confidence boost among our shoppers.

Improving beyond Silver isn’t so easy. Some changes require additional investment while others must be earned, such as through excellence in product or service quality.

For the final enhancements, we upgraded the endorsement so that it came from a more reputable third-party authority, boosted review scores from 4.5 to 4.8, and improved delivery terms to be best in market. We

also upgraded the buying guide to provide shoppers with even more accessible information to aid their decision. A more advanced pricing tool allowed for even more nuanced pricing preferences, while enhanced hero imagery and text made category considerations even more intuitive. Let’s call this optimised site the Gold version.

Google/The Behavioural Architects, N=1000 in-market online shoppers of sofas, U.K., 2023.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

For the Gold version of the invented sofa brand home page, we optimised the behavioural science principle expressions even further

Social proof

User ratings surfaced, with a high 4.8 avg. rating

Emotional priming

Hero imagery enhanced to make decisions more intuitive

Authority bias

Endorsements from a well-known, reputable website

Delivery friction

Delivery info surfaced, with best-in-market terms

Framing

Hero text upgraded to frame the information in a more appealing way

Anchoring

Price ranges surfaced to help people shop within their budget, with more nuanced options

Paradox of choice

A tool to help shoppers

Cognitive ease

More easily digestible

FIGURE 28

And how did Gold perform against Silver? Directly comparing the two, a higher proportion of participants chose the Gold version over the Silver version. This is impressive considering the fact that information on the Silver version was already comprehensive and well-optimised.

CHAPTER 3

A significantly higher proportion of participants selected the Gold version of the home page over the Silver version19

NEITHER

58%

36%

FIGURE 29

In addition to sofas, we tested two other categories — car insurance and whisky — that differ in terms of confidence, enjoyment, and permanence. And while we had to make some product-specific tweaks to how information was presented on these pages, we saw strikingly similar results.

While the impact of the behavioural science expressions was very similar, there were certain ways of addressing information needs that worked better for some products than others. For car insurance, the confidence of our shoppers was particularly boosted by expert reviews, which had less impact in the other two categories. For sofas and whisky, the brand impressions created by home-page visuals were more effective. The sofa category also saw confidence boosted significantly by providing tools that allow shoppers to explore and understand the full product range. This feature was also beneficial for whisky shoppers, but had less impact on car insurance.

Google/The Behavioural Architects, N=1000 in-market online shoppers of sofas, U.K., 2023.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

FIGURE 30

We supercharged the product pages of our invented brands with behavioural science principles and information needs more closely tailored to evaluation

Authority bias Delivery friction

Social proof

Emotional priming

Framing

Cognitive ease

Anchoring

? Paradox of choice

Social proof

Emotional priming

Delivery friction

Authority bias

Anchoring

Framing

? Paradox of choice Cognitive ease

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

A note on the performance of Gold and Silver

While our Gold, Silver, and Bronze pages generally ranked in performance as their names suggest, we did encounter a few unexpected results. Sofa shoppers did not appear to have a preference for whether their product was delivered within a week or within a four- to six-week delivery window. If anything, they may have had a slight preference for the longer delivery timeframe. Our working hypothesis here is that quality assumptions related to lead times might explain the ambivalence over delivery windows, especially for a product like a sofa where shoppers want their product to be well made. While these results are statistically significant, every brand is different. So testing to find the right approach remains critical.

Another unexpected finding across all three products was that interactive tools and guides had little impact. Here we hypothesise that the effort of either paging through a guide or working out how to use an unfamiliar tool was perceived as being greater than the likely gain in information.

Again, we should offer a note of caution about real-world implications: as our page was a mock-up, shoppers couldn’t actually interact with the tools or guides. But as their presence had a low impact on confidence, it certainly suggests that research and testing would be wise before investing in this kind of functionality or content.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Confidence to evaluate

Our home pages were designed to meet information needs and reduce psychological barriers during the expansive exploration mindset. But we also wanted to understand what content would be most confidence-building for consumers in the reductive evaluation mindset. For this we created a near identical experiment using product detail pages rather than home pages.

We supercharged the product pages of our invented brands with behavioural science principles and information needs more closely tailored to evaluation

Anchoring

Cognitive ease

Social proof

Default bias

FIGURE 31

For the product page, we tested a different set of behavioural science principles and information needs more closely tailored to evaluation. But the process was otherwise identical.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

Participants were presented with variations of Bronze, Silver, and Gold, with features such as reviews, ratings, payment plans, and delivery terms all varying in presentation and strength of expression. The behavioural science principles explored in this phase of the experiment included:

Authority bias: the tendency to give weight to a source of information because of its perceived expertise. Represented in our experiment by endorsements from reputable websites.

Paradox of choice: having unlimited choice can actually lead to indecisiveness, dissatisfaction, and regret. We explored this by providing various trial periods.

Social proof: the tendency to follow and copy the behaviour and actions of other people in situations of ambiguity

or uncertainty. Represented here by user ratings and reviews.

Mental accounting: people tend to keep track of spending by attributing purchases to different mental budgets. We explored this by providing different payment plans, which are a desirable category norm.

Cognitive ease: we tend to favour decisions that require minimal cognitive effort. We enhanced this by making the options more easily digestible.

Default bias: a tendency to continue behaving as we always have done, manifesting in online shopping as a preference for default options. This bias was explored by providing a default delivery option.

Endowment effect: the tendency to place a higher value on objects we own. We

For a full definition of all the principles we reference in our research, visit the Behavioural science glossary at the end of the report.

explored this bias by using a tool that allowed shoppers to visualise a sofa in their own home (though participants could not interact with the tool).

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

FIGURE 32

We supercharged the product pages of our invented brands with behavioural science principles and information needs more closely tailored to evaluation

Anchoring

Mental accounting

THIS WHISKY WAS AWARDED BEST

HIGHLAND WHISKY 2023 BY YOUR CHOICE. CLICK HERE TO FIND OUT WHY.

Authority bias

Endowment

effect

Cognitive ease

Social proof Default bias

? Paradox of choice

Anchoring

Want to know why we have been endorsed by Your Choice? Click here to read the full review

Authority bias

Cognitive ease

Endowment

effect

Social proof Default bias

Mental accounting

Paradox of

choice

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

More than twice as many participants preferred the Silver product detail pages with basic optimisation, over the Bronze versions with no optimisation20

63%

11%

NEITHER

26%

FIGURE 33

As with our home page experiment, we found that basic optimisation towards shoppers’ most common information needs resulted in a Silver page that substantially outperformed the default Bronze page in all three categories we tested.

Google/The Behavioural Architects, N=1000 in-market online shoppers of sofas, U.K., 2023.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

More than twice as many participants preferred the Gold product detail pages with fuller optimisation, over the Silver versions with basic optimisation21

NEITHER

64%

28%

FIGURE 34

And when a Gold version of the page was fully optimised with expressions of our behavioural science principles, a similar step-change in confidence was visible between Gold and Silver.

As in our home page experiment, we saw some commonalities — such as the importance of expert reviews in all three categories — but also some salient differences. In the car insurance category, expressions relating to product features and visual layout were more important than in either of the other categories. Intuitively, this makes sense given that car insurance is a more information-dense category and therefore shoppers benefit from having clearer detail about what is being offered.

Google/The Behavioural Architects, N=1000 in-market online shoppers of sofas, U.K., 2023.

CHAPTER 3

‌BOOSTING CONFIDENCE IN THE MESSY MIDDLE

In the sofa category, a greater degree of relative importance was attached to delivery terms, which again makes sense as this is generally not a product category shoppers can take home themselves. For whisky, we noted that several information needs — price, consumer reviews, delivery terms, and range of options — showed similar levels of importance. This indicates that all of these information needs play a meaningful role in shopper confidence.

Boosting confidence and boosting sales

At the start of this chapter, we stepped out of a marketing mindset and began thinking like consumers. Now that we’ve explored how behavioural science can help answer information needs and boost confidence, we’re going to return to the marketer’s point of view.

We all want shoppers to feel confident, but marketers will understandably be wondering if all of this increased confidence is likely to lead to increased sales. To answer that question, we re-ran both the home page and product page experiments, but this time, instead of rating their confidence, participants were asked to choose the version that would make them more likely to consider buying a sofa from Miller, whisky from Hyskeir, or car insurance from 4sure.

Marketers who invest in anticipating and answering the information needs of

consumers may be rewarded with increased purchase consideration

And we found that exactly the same factors that inspire shopper confidence also drive consideration. This suggests that there is reciprocity at play — that marketers who invest in anticipating and answering the information needs of consumers may be rewarded with increased purchase consideration. Helping to boost shopper confidence turns out to be a win-win.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

FIGURE 35

During exploration, the same factors that inspire consumer confidence also inspire consumers to consider purchasing22

Exploration Evaluation

Sofas

Car insurance

Whisky

Emotional priming Authority bias

Importance in driving consideration

Cognitive ease Emotional priming

Cognitive ease

Anchoring Authority bias

Text inspiration Social proofing

Cognitive ease Social proof

Social proof Authority bias

Delivery friction

Anchoring

Paradox of choice

Delivery friction

Framing

Framing Framing

Visual/text inspiration

5 Service friction

5 10 15 20 25

Importance in driving confidence

Google/The Behavioural Architects, N=6000 in-market online shoppers of sofas, whisky and car insurance, U.K., 2023.

CHAPTER 3

BOOSTING CONFIDENCE IN THE MESSY MIDDLE

FIGURE 35 CONTINUED

During evaluation, the same factors that inspire consumer confidence also inspire consumers to consider purchasing23

Exploration Evaluation

Sofas

Car insurance

Whisky

Authority bias

Authority bias Anchoring Anchoring Authority bias Mental accounting Social proof Defaults Social proof Defaults Mental accounting Endowment Range of options Mental accounting Cognitive ease Paradox of choice Endowment Paradox of choice Defaults Social proof Endowment Cognitive ease Anchoring

Importance in driving consideration

5 10 15 20 25

Importance in driving confidence

Google/The Behavioural Architects, N=6000 in-market online shoppers of sofas, whisky and car insurance, U.K., 2023.

BOOSTI
3
Chapter summary
69 CHAPTER 3
‌NG CONFIDENCE IN THE MESSY MIDDLE

Confidence is key to exiting the messy middle and making a purchase. But underlying levels of confidence vary widely by category, as do enjoyment and perceived permanence of a purchase.

Information needs and psychological barriers vary by product category, but common needs and barriers can be mapped to the exploration and evaluation mindsets in the messy middle.

Optimising website assets for the varying information needs and adding a touch of behavioural science principles can boost the confidence of the shopper in the shopper journey.

C ONTENTS
The same factors that inspire shopper confidence also drive consideration. This suggests that marketers who invest in anticipating and answering the information needs of consumers may be rewarded with increased purchase consideration.

4
‌Shopping in the messy middle

Exploring and evaluating in the messy middle isn’t only about deciding what product you want to buy — you also have to figure out where you’re going to buy it from. And just as the internet age has seen a rapid increase in the number of available products, the number of sellers has risen in tandem. This entanglement between products and points of sale invites several interesting questions:

Is it easier to encourage consumers to change what they buy or where they buy it from?
Can retailers24 also use behavioural science to drive consumer choice?
What can brands do to encourage consumers to buy directly from them?
How can marketers build propositions that inspire confidence without relying exclusively on brand or pricing?

To answer these questions, we ran a new experiment focused on nine product categories: trainers, dog food, whisky, laptops, face moisturiser, second-hand cars, flights, hotels, and mobile SIM cards.25 This means that in addition to our nine invented brands, we also invented nine retailers to simulate the entry of new sellers into the marketplace.

C ONTENTS
CHAPTER 4 SUMMARY
Retailers” is used here as a catch-all term for any point of sale from which consumers purchase items or services.
Each category was tested with a minimum of 800 in-market shoppers.”

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 36

We tested nine invented retailers to simulate the impact of behavioural science principles on new market entrants

Category Brand Retailer

Trainers

Laptop

Products

Moisturiser

Dog food

Whisky

Second-hand car

Flights

Services

Hotel

SIM

Once again, we applied some of the most powerful behavioural science principles revealed by our original research: social proof, authority bias, category heuristics, and power of free. Full definitions of all these principles and more can be found in the Behavioural science glossary at the end of the report.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

But while those principles work well for exploring and evaluating products, there are some additional factors that come into play at the moment of purchase decision-making. So for this experiment we added delivery friction26 — how quickly consumers receive an item — which can be thought of as analogous to power of now, which we explored in some of our other experiments. And to measure the importance of value to purchase decision-making, we added price and promotion as a variable, testing products at full price as well as with discounts of 10% and 15%. It’s important to note here that no behavioural science was applied to how this pricing information

FIGURE 37

We tested the impact of different factors in the exploration and

evaluation stages, and in the purchase decision-making moment

Product (what)

Retailer (where)

Social proof

Power of free

Authority bias

Delivery friction

Category heuristics

Price and promotion

Product brand

Retailer brand

was presented to shoppers. And as with our previous experiments, we have to acknowledge the value of existing brand equity, with household-name retailers enjoying a meaningful head start.

In our experiment, shoppers were asked to choose first-choice and second-choice brands for both product and retailer brands. They were then shown eight simulated shopping scenarios with different expressions of our chosen behavioural science principles. For example, the st�H�H�R=�?h�H �H �Hroduct. The purpose of this design was to see which elements were most important in the joint proposition of product + retailer.

For service-based categories we used risk aversion to explore how retailers can enhance a proposition with better terms for cancellation after purchase. This was an alternative to delivery friction, which we use for physical products. For example, instead of “free next day delivery” (minimise friction in delivery), we had “no cancellation fee” (mitigate the risk around any decision with a service). For services where there is no physical product to deliver, mitigating the perceived risk around cancelling a service offers an equivalent but more realistic proposition brands could use for those categories.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 38

We simulated shopping experiences for first-choice brands and second-choice brands, varying the ways we expressed different behavioural science principles

OPTION 1

Your favourite trainers retailer

Trainer brand

4.9

22 reviews

Minimum shipping fees apply

Top industry award

Free high-visibility laces

Durable and lightweight

£90.00

OPTION 2

Your second-favourite trainers retailer

4.0

Free standard delivery and returns

Top industry award

Free 30-day access to training app

Proprietary cushioning technology

Now £76.50 — £13.50 off!

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 39

How we tested price and promotion in most categories

Participants were asked to state how much they would be willing to pay for a product in the category they are considering making a purchase in.

“For the trainers you are looking to buy, how much would you be looking to pay? Please be as accurate as possible using the slider below.”

£25

£90

£200

We either piped that price directly into the simulation, without a promotion, or added a 10% or 15% discount.

£90

Buy now

£90

Now £76.50

Buy now

How we tested price and framing in the second-hand car category

“For the second-hand car you are looking to buy, how much would you be looking to pay? Please be as accurate as possible using the slider below.”

£1,000

£27K

£50K

We framed that price in relation to the average market price for second-hand cars.

£22,950

£4,050

below market average

Enquire now

£24,300

£2,700

below market average

Enquire now

CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

Retailers, products, and pricing

Analysing the thousands of decisions modelled across our product and retailer experiment allows us to draw some category-wide conclusions about how these elements contribute to decision-making.

One of our first goals was to understand the relative “stickiness” of product and retailer preference. We began with a relatively level playing field, comparing the effect of simply asking shoppers to choose between combinations of their first-choice brands and second-choice brands and retailers with identical propositions.

When shoppers were shown their first-choice retailer offering both their first-choice brand and second-choice brand,, an average of 27% decided to switch brand preference to their second choice. However, when shoppers were asked to choose between buying their preferred brand from either their first-choice or second-choice retailer, 48% were willing to switch.

In other words, all things being equal, product preference appears to be much more durable than retailer preference

And this effect was amplified once we began to apply behavioural science. We found that supercharging product propositions with behavioural science was enough to shift preference away from first-choice brands to second-choice brands, or — as shown in the upcoming charts — to completely invented brands. While there was some variance by category, the pattern across product brands was consistent across many categories (though travel and second-hand cars behaved a little differently, as we’ll explore later).

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

In the U.K., supercharging invented product brands

with behavioural science had a significant impact on brand

preference across all categories27

Products

First-choice product brand

Invented product brand

Laptop

SIM

Whisky Moisturiser Dog food Trainers

Flights

Hotel

Second-hand car

32%

68%

48%

52%

48%

69%

31%

71%

29%

73%

27%

75%

25%

75%

25%

75%

25%

FIGURE 40

However, on the retailer side of the equation, the effect of behavioural science was significantly more pronounced. Even completely invented retailers with supercharged propositions were able to attract large numbers of shoppers away from their first choice and second choice. Around a third of shoppers were prepared to switch from their first choice to an invented product brand, but more than eight out of 10 switched to an invented retailer.

Google/The Behavioural Architects, U.K., n=844-1001 prospective online purchasers per category, ages 18-75, 2022.

CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

In the U.K., supercharging invented retailers with

behavioural science had an even more pronounced impact on brand preference28

Retailers

First-choice retailer brand

Invented retailer brand

91%

94%

Whisky

Dog food

Flights

Trainers Moisturiser

SIM

Hotel

Laptop

Second-hand car

15%

85%

15%

85%

16%

84%

21%

79%

27%

73%

32%

68%

FIGURE 41

This presents a risk and an opportunity for established retailers and newcomers alike. While this data shows that shoppers are very willing to switch to a different point of sale, that doesn’t change the absolute number of sales waiting to be won or lost. And with behavioural science as a powerful persuasive factor, the key for retailers is simply to win more than they lose.

Retailer vs. product

After finding such a substantial difference between product and retailer “stickiness”, we wanted to dig deeper. So the next step was to calculate importance scores for each of the elements in our experiment, to see if any patterns emerged in their contribution to both propositions.

Google/The Behavioural Architects, U.K., n=844-1001 prospective online purchasers per category, ages 18-75, 2022.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

We found that influence was divided almost equally between those principles relating to product (the what) and those that have more to do with the retailer (the where). This outcome was neither predetermined nor expected. And while we saw less stickiness for retailers, it tells us that both aspects — what and where — are important. Consumers

In the U.K., we found that influence was divided almost equally

between behavioural science principles relating to product (the what) and those that have more to do with the retailer (the where)29

Product brand 17%

Prices and discount 30%

Social proof 15%

Power of free 9%

Category heuristics 8%

Delivery friction or risk aversion 8%

Authority bias 7%

Retailer brand 8%

Note: The eagle-eyed among you may notice the sum total is 102% — don’t worry, it’s not an error, but due to rounding to whole numbers

FIGURE 42

want the right product and they also want a convenient shopping experience from retailers.

For marketers this means that whether you’re focused on retail or product marketing — or both — there are behavioural science levers on both sides that influence consumer decision-making. Ideally these would be delivered in harmony, making it easier for consumers to bundle the right product with the right experience.

Google/The Behavioural Architects, U.K., n=8,645 prospective online purchasers (of categories: trainers, laptop, face moisturiser, dog food, whisky, second-hand car, mobile SIM, flights, hotel), ages 18-75, 2022.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

Secondly, we see the continuing importance of brands in decision-making. This was a recurring question from readers of our original research, so we’re pleased to shed more light on it here. As we’ve covered elsewhere, brand remains important. But, between products and retailers, not all brands are created equal.

The influence of the product brand was twice as large as the influence

of the retailer brand when it came to influencing shopper preference30

Product brand 17%

Retailer brand 8%

FIGURE 43

In the chart below, the light blue segment representing the influence of the product brand is much larger than the dark blue one representing the retailer brand — in fact, it’s twice as large:

And this makes intuitive sense. If a shopper already has an item in mind (a new pair of name-brand trainers, for example), it’s much easier to influence where they buy those trainers than to change their mind about which brand they want to buy. For retailers, this means that having the right range of products curated and available in stock is vital.

Google/The Behavioural Architects, U.K., n=8,645 prospective online purchasers (of categories: trainers, laptop, face moisturiser, dog food, whisky, second-hand car, mobile SIM, flights, hotel), ages 18-75, 2022.

CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

Thirdly, it will be no surprise that pricing and promotion were important factors, accounting for 30% of decision-making in our experiment.

Assessing results across all the categories we tested, price and discounting made the biggest overall contribution to decision-making. Next came product brand with 17%, though retailer brand only ranked joint fifth with an influence score of 8%.

Social proof was the strongest behavioural science principle, with the rest largely making equal contributions. This reinforces the results from our previous experiments, showing just how important reviews are in boosting confidence. Comments and star ratings have to be earned by good products and good service, but they shouldn’t be out of reach for any determined brand or retailer.

Social proof was the strongest behavioural science principle, showing just how important reviews are in boosting confidence

Behavioural science vs. price

For both brands and retailers, economic realities mean it isn’t always easy to pull the big lever marked “discounts”. And if that’s the case, then our results offer plenty of cause for optimism. After all, 70% of decision-making in our experiments was still driven by factors other than price.

In fact, in many instances, the cumulative effect of applied behavioural science matched that of discounting.

As we discussed at the end of the previous chapter, there could be a beneficial reciprocity at work here, with shoppers rewarding those brands and retailers who take the time to provide them with the information they need, in a format they find easy to understand, and that shows awareness of the underlying human behaviours that guide our decision-making.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 44

In the U.K., 70% of decision-making in our experiments

was driven by factors other than price31

Power of free

Prices and discount

Delivery friction

70%

Other factors

30%

Price and discount

Authority bias

Brand

Social proof

Category heuristics

The most effective behavioural science principles varied by category, but to go back to our trainers example, appropriate expressions of social proof, category heuristics, and power of free had as much cumulative

impact in our test as price and discounting. In fact when we supercharged propositions with the strongest expressions of the relevant behavioural science principles, they had more impact on shopper choice than a 10% discount in the laptop, trainers, and moisturiser categories.

Some of the principles we’ve described are easier to implement than others, especially if you already have a strong review base or expert endorsements. As we saw in the previous chapter, some improvements might even require a significant investment of time or money, such

as logistical adjustments to shorten delivery windows or provide free shipping. But before you discount, it’s worth looking at all the other levers available to you — especially the ones that aren’t as costly for your brand to activate.

Google/The Behavioural Architects, U.K., n=8,645 prospective online purchasers (of categories: trainers, lap-top, face moisturiser, dog food, whisky, second-hand car, mobile SIM, flights, hotel), ages 18-75, 2022.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 45

We found that combinations of behavioural science principles have as much cumulative impact as price and discounting32

31%

Prices and discount

30%

Social proof

Category heuristics

Power of free

Beyond brands and pricing, behavioural science gives us a powerful lens through which to analyse other key factors that affect our decision-

making. The impact of these other factors did vary category by category, with some clearly more price-sensitive and brand-sensitive than

others. However, there were no categories where the relative impact of the behavioural science expressions was so small as to suggest that it wouldn’t make a difference in situations where margin or other considerations prevent discounting.

Throughout the various phases of our research, we’ve been surprised again and again by the power of these expressions to help shoppers clarify their needs, gain confidence in their choices, and ultimately make the decision to buy. And perhaps no finding has surprised us as much as realising that a little bit of applied behavioural science can have the same impact as offering a meaningful discount on price.

Google/The Behavioural Architects, U.K., n=8,645 prospective online purchasers (of categories: trainers, laptop, face moisturiser, dog food, whisky, second-hand car, mobile SIM, flights, hotel), ages 18-75, 2022.

CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

A quick look at: Retail realities

While our research into retail decision-making focuses on the online environment, the physical reality of how people shop is admittedly more complicated. E-commerce — or “internet street” as we characterised it in our original research — is still growing and there are now countless online retailers that make up the backbone of the digital economy. But over the past decade, the line between internet street and high street has begun to blur.

FIGURE 46

Across Europe, the Middle East and Africa, as many as

two-thirds of consumers now shop in a channel agnostic way33

Online Offline Agnostic

61%

67%

65%

69%

33%

58%

10%

29%

14% 19%

11%

24%

22%

2019

2020

2021

2022

2023

As many as two-thirds of consumers now shop in a channel agnostic way, moving from digital to physical retail and back again as their needs

change. And as the chart on the next page demonstrates, both online and offline media drive online and offline retail traffic.

Google / Savanta / Cint / Kantar, Ireland, U.K., Austria, Germany, Switzerland, France, Italy, Portugal, Spain, Belgium, Denmark, Finland, Netherlands, Norway, Sweden, Croatia, Czech Republic, Estonia, Greece, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Slovenia, Egypt, Ghana, Israel, Kenya, Nigeria, Saudi Arabia, South Africa, Turkey, U.A.E., Smart Shopper Research, Online - product buyers of Consumer Electronics, Home and Garden, Fashion, Toys, Beauty, Food. n = min 1000 per category, per year.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 47

Both online and offline media drive online and offline sales, but in this study digital

media drove a substantial 65% of e-commerce’s media-generated revenue

Digital

Traditional

43%

29%

71%

35%

65%

46%

54%

E-commerce

Bricks and mortar

The left-hand side of the flowchart34 shows the proportion of revenue driven by digital and traditional media that flowed to e-commerce or physical channels (both channels have non-media-driven revenue not included here). Both digital and traditional media were successful at driving omnichannel sales. However, while revenue from digital media was split almost 50/50 between the two sales channels, revenue from traditional media showed a 70/30 split in favour of physical retail. Overall, digital media drove 65% of e-commerce’s media-generated revenue, showing how dependent the channel is on digital media activity. Find out more about omni-channel retail here on Think with Google.

In addition to the separation between online and offline becoming narrower, so too has the divide between manufacturers and retailers. Logistics, digital storefronts, warehousing, and fulfilment have all been significantly commoditised, leaving brands to focus on what they do best. Technological solutions now exist for relationship management, customer service, and other aspects of the shopping journey. This has all contributed to a trend of manufacturers opening direct-to-consumer stores on top of their traditional retail sales channels.

Google / Analytic Partners ROI Genome, Global, Get Retail Ready: A meta study in driving business success, n=152 omnichannel retailers, 2019-2020.

CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

Full-spectrum retail

In addition to exploring the relationship between choice of product and choice of retailer, we also wanted to take a closer look at different flavours of retail. The retail environment varies considerably depending on product category and whether manufacturers also

sell directly to consumers. So we included all of these scenarios in our experiment, creating invented aggregators and direct-to-consumer brands as necessary.

Physical products and services

To measure the difference between these two broad categories, we deliberately chose to test a mix of physical products (dog food, whisky, laptops, face moisturiser, and second-hand cars) and services (flights, hotels, and mobile SIM contracts). And, for the most part, our results show only minimal variance between the two groups. However, a few unique insights do stand out. Firstly, promotions appear to have a larger impact on travel services than on physical products, with flights and hotel contracts exhibiting greater price sensitivity.

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

FIGURE 48

Promotions appear to have a larger impact on travel services,

such as flights and hotels, than on physical products35

39%

Flights

35%

Services

Hotel

31%

SIM

30%

Trainers

Prices and discount 30%

29%

(average across categories tested)

Whisky

28%

Products

Dog food

27%

Laptop

25%

Moisturiser

21%

Second-hand car

Google/The Behavioural Architects, U.K., n=844-1001 prospective online purchasers per category, ages 18-75, 2022.

CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

Conversely, in the case of second-hand cars, pricing (as expressed through price framing)36 had much less impact. Instead, social proof was more important. This perhaps speaks to a combination of the cost of purchase and a lack of mechanical knowledge — especially as the financial and safety consequences of buying the wrong car could be severe. Social proof is a powerful principle — in fact it was the most powerful behavioural science principle in all nine categories — but its power on second-hand cars was 1.7X greater than its average effect in other categories.

D2C or not D2C?

Alongside digital-native manufacturers selling directly through online channels, several other categories allow for direct-to-consumer (D2C) sales (think hotel chains that offer direct booking or fashion stores that sell their own brands alongside other labels). But to keep things simple, our investigation into this area focused on goods rather than services. We also limited ourselves to those categories where it was feasible for the same product to be available from both resellers and direct from the manufacturer: trainers, dog food, whisky, laptops, and moisturisers.

The economic pros and cons of launching a D2C channel are beyond the remit of this research. But there was one important question we wanted to answer: do consumers respond differently to D2C compared to any other retailer?

Based on our experiments, the short answer is no. Our results suggest that manufacturers going D2C are likely to achieve preference shares similar to those of average category retailers, providing they are present during the decision-making process to give consumers a choice. In a variant of our product/retailer experiment that included a D2C option, the results were roughly in line with those of second-choice or challenger retailers.

When price and proposition were kept the same, on average 42% of shoppers across categories chose the D2C option over their preferred retailer, compared to an average of 47% for their second-choice retailer. When supercharged, the average was 73% compared to 77%,

respectively. Looking at the range in results across categories, there was considerable overlap between the intervals.

It’s worth noting here that the same sorts of expressions were used across all nine categories with only two exceptions. For the second-hand car category we used price framing rather than an explicit percentage discount. This is better aligned to the market reality for the cat-egory, both as a convention widely used on second-hand car sales websites and, especially, given the context of the car market in recent years, which has seen price inflation as a result of supply chain shortages and manufacturing delays. And for service based categories we used risk aversion to explore how retailers can enhance a proposition with better terms for cancellation after purchase as an alternative to delivery friction. For example, instead of “free next day delivery” (minimise friction in delivery), we had “no cancellation fee” (mitigate the risk around any decision with a service).

CHAPTER 4

SHOPPING IN THE MESSY MIDDLE

36%

57%

35%

51%

68%

91%

56%

88%

Our results suggest that manufacturers going D2C are likely to achieve preference shares similar to those of average category retailers37

Range across categories where participants chose a second-choice retailer or D2C

Range in response

Average across categories

When retailer “just shows up”

Second-choice

retailer

100%

47%

Direct to

consumer

100%

42%

When retailer is supercharged with two behavioural science principles

Second-choice 0% retailer

100%

77%

Direct to

consumer 0%

100%

73%

FIGURE 49

So while each manufacturer will have to weigh up the costs and benefits of embracing D2C, our research shows that if the numbers make sense, the channel could be terrain worth exploring within the messy middle for brands who can build propositions with insight and intelligence.

Google/The Behavioural Architects, U.K., n=1,000 prospective online purchasers per category, ages 18-75, June 2022.

Cha4pter summary

89 CHAPTER 4

‌SHOPPING IN THE MESSY MIDDLE

Exploring and evaluating in the messy middle isn’t only about deciding what you want to buy — consumers also have to figure out where they’re going to buy from.

When it comes to the moment of purchase decision-making many of the behavioural science principles we’ve tested

for explore and evaluate also play a role. But there are additional principles and factors crucial here — for instance delivery friction (how soon you can receive an item), and the role of pricing and promotion.

People’s preference for specific products and services is stronger than it is for retailers. Completely invented retailers with supercharged propositions were able to attract large numbers of shoppers away from their first-choice and second-choice retailers — on average more than eight out of 10 in our tests switched to an invented retailer.

While promotional activity does play a key role in decision-making, supercharging propositions with the best expressions of relevant behavioural science principles can have as much, if not more, impact on what people choose.

‌ Conclusion

We began this research project with a hunch that age-old behavioural traits are the key to understanding online decision-making. And we end our latest bulletin with arguably the most striking insight to date: that “charging” a proposition with modest expressions of behavioural science can be as persuasive as a 10% price cut.

Ensure brand presence

One of the most surprising findings from our original research was that simply showing up could be enough to cause a meaningful proportion of shoppers to switch from their first-choice brand to a challenger. But what does showing up mean in an ever-expanding digital landscape?

Broad coverage has moved from being an aspiration to a necessity. Marketers — and retailers — need to show up and connect with consumers, wherever they are. And while there’s no doubting the challenge of being present in all the viable information patches of the messy middle, marketers do at least have a powerful ally in making this happen. As we enter a new era of marketing, a new generation of tools, including Google AI, will enable businesses to connect new customers with relevant ads across channels in real time.

91 CONCLUSION

‌CONCLUSION

Close the gap between trigger and purchase

Showing up is crucial, but how you show up can be decisive. At every touchpoint, marketers need to lean on their creativity to provide content and messaging that draw shoppers to their patch and satisfy their information needs. With new channels and surfaces emerging all the time, it won’t be possible to hand-craft content for every possible surface, environment, and scenario. Instead marketers will need to integrate behavioural science into flexible campaigns, with AI taking on the burden of customising specific formats. But with brilliant baseline assets, those campaigns may earn reciprocity by helping shoppers explore, evaluate, and exit the messy middle.

Employ behavioural science creatively and responsibly

Being present is one thing, but being present and helpful requires something more. After all, it’s highly unlikely that yours will be the only brand showing up at any given moment in the messy middle. The

Employ behavioural science creatively and responsibly

behavioural science principles explored in this report provide a powerful toolkit to help marketers enhance the effectiveness of their ads and owned content. What will ultimately set brands apart is how they harness creativity and human insight to craft compelling propositions that address the underlying behavioural drivers governing what people choose to buy and where they choose to buy from.

92 CONCLUSION

‌CONCLUSION

Ensure brand presence

Close the gap between trigger and purchase

At the conclusion of our original research we made three recommendations that are worth revisiting here

These broad conclusions were the result of our initial exploration of the messy middle — a territory that only reveals its topography as a result of patient, large-scale research. Happily, after several years of further study, we still believe they hold, albeit with a couple of refinements.

‌ Appendix

Behavioural science glossary

These definitions were developed in partnership with The Behavioural Architects. It’s important to note that hundreds of different behavioural science principles exist, but these are the ones we used in our research. Definitions can also vary depending on context — these definitions are drawn from our own interpretations.

Affect bias: the tendency to make decisions based on how we feel rather than any rational or thought-out reason. Imagery is often powerful due to the potential for high affect.

Ambiguity aversion: when choices have to be made between alternatives that involve uncertainty and unclearness about the future, it has been found that people prefer choices involving clear probabilities to those involving vague probabilities.

Anchoring: we look for reference points (anchors) that we can rely on and adjust our judgements and decisions from. Anchors may be based

on price, past experience, occasions, or social norms. Price anchoring can be supported by indicating price ranges on a home page to help people shop within their budget.

94 APPENDIX

Authority bias: the tendency to attribute greater accuracy and knowledge to the opinion of a perceived authority figure and be more influenced

by that opinion. That authority can be achieved through fame, power, or position. In advertising and marketing, this cognitive bias is quite

common, being found in product recommendations by professionals and testimonials by celebrities or public figures.

Category heuristics: a category heuristic is a mental shortcut that helps us make a quick decision within a given category without having to explore in detail all the characteristics of each option in depth. In other words, category heuristics are shortcuts or rules of thumb that aid us in making a quick and satisfactory decision within a given category. Heuristics that link to category norms and trends are particularly effective.

Cognitive ease: the natural desire a person has to make decisions and judgements using “System 1” thinking, a type of thinking that happens automatically and with little effort, driven by instinct and experiences. When information is presented in a way that requires minimal cognitive capacity, it enables someone to make decisions using “System 1” thinking.

Costly signalling: signals communicate information to others through a message and the way the signal is communicated can boost the credibility of the message. Costly signalling is a type of messaging aimed to convey a brand has considerable wealth, e.g., a sponsorship of a prime time TV show. This attributed wealth can be a positive sign of the brand’s health, quality, or performance.

Default bias: we tend to stick with what has already been selected for us

— the “default” — or continue behaving as we always have done. This can manifest in online shopping in choosing preferences that have been

pre-selected rather than making our own decisions.

95 APPENDIX

Delivery friction: reducing the time and/or effort (friction) in which the consumer would receive an item they are thinking of purchasing. This can be thought of as analogous to power of now, which we explored in some of our other experiments. Examples would be reducing delivery times, or offering the availability for the consumer to pick up the item themselves.

Diversification bias: a choice heuristic that says we tend to seek more variety when making several choices at once than when making choices sequentially, one at a time, over a period of time.

Emotional priming: primes are subconscious influences on our thinking and behaviour caused by different cues or stimuli, for example images. Emotional priming is the tendency for cues or stimuli to influence how we feel emotionally and the impact this can have on decision-making.

Endowment effect (linked to Loss aversion): the tendency to place a higher value on objects we own and that are in our possession.

Framing: decisions and preferences are influenced by how information is presented to us, even when the choices before us ultimately lead to the same outcome — for example, if a food item is presented as “10% fat” or “90% fat-free”. The product is the same, but consumer responses can be very different.

Halo effect: positive perceptions of an object can significantly influence our perception of adjacent objects. This can be useful for introducing new products by embedding them in existing ecosystems, or in partnership with beloved brands.

IKEA effect: consumers often place a disproportionately high value on products they created or had a role in creating.

96 APPENDIX

Loss aversion (linked to endowment effect): we overvalue and hate losing what we already have. Typically a loss hurts twice as much as

a gain feels good. This bias explains the power of a free trial. People

may also take huge, sometimes irrational risks, in order to keep possession of something.

Mental accounting: people carry around different running money tabs in their heads. We compartmentalise income and spend from different

mental accounts. We are also more or less reckless with money depending on where it has come from.

Mere exposure effect: repeated exposure to something leads to a more positive feeling about it. In particular, increasing familiarity can help to build trust in a logo, brand, or institution.

Paradox of choice: we think more choice is better, but there can be problems with excessive choice. Decisions become less likely (choice paralysis) and choice satisfaction may drop. This can be addressed by providing a tool or a catalogue to help shoppers avoid choice and information overload.

Power of free: the tendency to overvalue free products. The demand for a product is significantly greater at a price of exactly zero compared to a price even slightly greater than zero. Power of free is a source of irrational excitement that can be highly persuasive in the moment.

Power of now (or Present bias): the tendency to focus on today rather than think about what tomorrow might bring. We discount the future in favour of today. More specifically, it is the human tendency to put more value on rewards that are available now than rewards in the future.

Power of now is also known as present bias.

97 APPENDIX

Pratfall effect: the tendency for interpersonal appeal to change after an individual makes a mistake, depending on the individual’s perceived competence. The pratfall effect explains why people we hold in high esteem seem even more sympathetic if/when they make a mistake.

Likewise, when we see someone we don’t hold in high esteem make a mistake, we appreciate them even less. The pratfall effect extends to brands that may admit to mistakes or flaws from time to time.

Prospect theory: the tendency to choose options with a higher probability of a positive outcome, even if the potential rewards are lower. Conversely, we tend to avoid options with a higher probability of a negative outcome, even if the potential loss is smaller.

Scarcity bias: people place a higher value on an object that is scarce and a lower value on those that are abundant. Scarcity is a mental shortcut that assigns a value to an item based on how rare or limited it is. The

more difficult an item is to acquire, the more value that item has. These perceptions of scarcity can lead to irrational consumer behaviour.

Social proof: the tendency to follow the opinion, advice, and behaviour of others in one’s own decision-making. The recommendations and comments of others can be very persuasive.

98 APPENDIX

Thanks for making it this far!

As a reward, here’s some bonus content

Let’s get ready to bundle (or unbundle)

Here, we explore the benefits that bundling (or unbundling) can bring to consumers and businesses alike.

‌Let’s get ready to bundle (or unbundle)

Bundling and unbundling of products and services continues to be a hot topic in marketing. In fact, we’re not immune ourselves, initially

bundling this section into Chapter 5 before deciding to unbundle it as an appendix, because, unlike the other studies represented here, we haven’t conducted any new primary research on the topic. Still, we thought there were some interesting observations we could share...

Bundling involves combining multiple products or services into a single packaged offering, often including a discount. Think fast-food meal deals where you can add a drink and a side for less than the cost of buying all three items separately. That’s a bundle. Unbundling consists of breaking up packages of products and services into their separate constituents and selling them individually. In recent memory, unbundling has transformed industries as varied as music, telecoms, and TV.

This phase of our research primarily used literature review, supplemented by an audit of leading retailers to understand how they apply bundling

in the real world. According to published work on the subject, there are four main types of bundles, each conferring different levels of benefit to consumers and businesses:

99 APPENDIX

‌FIGURE 50

We grouped different types of bundling into four categories

Premium Streaming

Premium Delivery

Pure bundling:

you can’t buy individual goods at individual prices. Often invisible, we don’t realise there was a bundle until the product gets unbundled. For example, this can happen with airline travel.

Mixed bundling:

the individual components are also available

for purchase individually. For example, this can happen with meal deals at fast-food restaurants.

Partial mixed bundling: the bundle is offered, along with one of the bundled items separately, but not the other.

Tying:

the purchase of item A is tied to buying item B, although item B might be sold separately (partial

bundling) or not (pure bundle).

Bundles of behaviour

Our literature review revealed nine behavioural science principles that are particularly relevant when it comes to developing bundling and unbundling strategies.

IKEA effect: consumers often place a disproportionately high value on products they created or had a role in creating. Customised or personalised bundles are likely to be valued more and returned less.

Loss aversion: people are more motivated by avoiding a loss than by acquiring a gain of equal value.

100 APPENDIX

Prospect theory: people tend to choose options with a higher probability of a positive outcome, even if the potential rewards are lower. Conversely, they avoid options with a higher probability of a negative outcome, even if the potential loss is smaller. This is often applied by showing the cost of a bundle as a single figure, but itemising individual savings.

Cognitive ease: people prefer to receive information in a way that reduces mental effort and allows them to make intuitive decisions. Default bundles and a single contract for multiple services can reduce hassle and increase cognitive ease.

Framing: decisions and preferences are influenced by how information is presented to us, even when the choices before us ultimately lead to

the same outcome — for example, if a food item is presented as “10% fat” or “90% fat-free”. The product is the same, but consumer responses can be very different.

Power of free: people are particularly drawn to things that are free or offer a substantial saving. We often overvalue deals containing a free item.

Mental accounting: people tend to think about spending in terms of separate categories or budgets. Bundling items from complementary categories can be a way to reduce cognitive strain.

Diversification bias: this bias describes our tendency to want more variety when bundling multiple decisions into one. Infrequent grocery shoppers who visit the supermarket with long lists tend to buy a wider variety of items.

1 01 APPENDIX

Bundling in the messy middle

For shoppers exploring an unfamiliar category, bundles can answer several cognitive needs. A well-constructed bundle can inspire confidence, evoke feelings of control, and fend off anxieties associated with loss aversion and mental accounting. Unbundling can also appeal during exploration if it offers an innovative way to access the category,

e.g. only paying for what you need.

Bundles can also be an effective bridge between exploration and evaluation, by leveraging diversification bias and reducing the perceived risk of a purchase. Cross-selling and upselling are often successfully employed once a shopper has entered evaluation mode, close to the point of purchase. But if used inappropriately these tactics can confuse consumers, causing them to loop back into exploration to assess offers from other brands and retailers.

Finally, both bundling and unbundling contribute to post-purchase experience, as they can encourage consumers to shortcut the messy middle with repeat purchases.

Acknowledgments

1 02 APPENDIX

‌ ACKNOWLEDGEMENTS

This work was influenced by many, and we thank them for their advice, support and contributions.

From Google: Celena Baggio, Gerald Breatnach, Laura Smith-Roberts, Zarina de Ruiter, Selina Rennie, Liz Cracknell, Justine L’Estrange, Nisha Mathews, Carla Puttini, Rachael Powell, Tommy Howell, Ruth Ballett, Sarah Bradley, Sophie Florian, Raffaella Stratta, Alanna Petroff, Lucile Le Goallec, Bénédicte Ibert, Florence Corbasson, Asia Borowska, and Phil Wilson.

From outside Google: Andy Goll, Sam Kelly, Anais Le Joncour, and Veronica Beecham from The Behavioural Architects. Alicia Birr, Pilar McCrory Huertas, Alistair Rennie, Lucy Sinclair, and Philip Walford. Our creative agency partner, Redwood BBDO.