Market insights 21 studies

Market insights. What the research says.

Lecy’s market insights gather 21 public studies on customer loyalty, Google reviews and visibility in AI answers (AEO), each with what it changes for a physical shop or a franchise network.

Talk about your business

Featured AEO

45% asked an AI to recommend a local business, against 6% a year earlier.

BrightLocal United States, 1,002 adults

2026Your customers already ask an AI for adviceRead the study

01Loyalty 6 studies

Loyalty: why keeping your customers pays

Keeping a customer costs less than winning one, provided the programme asks nothing of them. 6 studies, from Harvard Business Review to the European Central Bank, on what brings a customer back.

Discover the Loyalty module

Harvard Business ReviewReview 2014

5 to 25×

more expensive to win a customer than to keep one

Keeping a customer costs far less than winning one

According to Harvard Business Review, acquiring a new customer costs five to twenty-five times more than keeping an existing one.

For your business

Every customer who comes back is a sale with no acquisition cost. For a local shop as for a network, loyalty is the first growth lever, before advertising.

Amy Gallo, “The Value of Keeping the Right Customers”, Harvard Business Review, 2014Read the study (new tab)

Harvard Business ReviewService companies 1990

+5%

more customers kept, and profits that can almost double

A few more loyal customers weigh heavily on the margin

Reichheld and Sasser’s classic study shows that a service company can raise its profits by almost 100% by keeping just 5% more of its customers.

For your business

Loyal customers buy more often, cost less to serve and recommend you. Counting visits one by one is how you see that lever move.

Frederick F. Reichheld and W. Earl Sasser, “Zero Defections: Quality Comes to Services”, Harvard Business Review, 1990Read the study (new tab)

BondUnited States 2023

17.9

loyalty programmes per consumer, only half of them used

One loyalty programme in two is asleep

In 2023, an American consumer belonged on average to 17.9 loyalty programmes, only about half of which were active.

For your business

The problem is not signing the customer up, it is getting them to use the programme at every visit. Every card to find, app to open or number to type makes them drop off.

Bond, The Loyalty Report 2023, via eMarketerRead the study (new tab)

European Central BankEuro area 2025

81%

of in-store card payments are contactless

Holding a phone to the counter has become the way to pay

In the second half of 2024, 81% of card payments at the point of sale were contactless in the euro area.

For your business

A loyalty programme that reuses this gesture asks nothing new of the customer: they hold up their phone, as they do to pay.

European Central Bank, “Payments statistics: second half of 2024”, 2025Read the study (new tab)

Journal of Consumer ResearchIn-store experiment 2006

34%

of cards completed when progress shows, against 19%

Seeing how far along they are brings customers back

A loyalty card whose progress is visible from the start is completed by 34% of customers, against 19% for a card with no head start.

For your business

Showing the customer, at every visit, their points and what is left before the reward is not a display detail: it is what brings them back.

Joseph C. Nunes and Xavier Drèze, “The Endowed Progress Effect”, Journal of Consumer Research, 2006Read the study (new tab)

Cisco12 countries, over 2,600 people 2024

75%

will not buy without trusting how their data is used

Trust comes from the data you do not ask for

75% of consumers say they will not buy from an organisation they do not trust with their data.

For your business

The less data a programme asks for, the less it makes people hesitate. A programme that is anonymous by default removes a reason to say no.

Cisco, Consumer Privacy Survey 2024Read the study (new tab)

02Satisfaction 6 studies

Satisfaction: why reviews decide before you do

Your customers read reviews before they walk in, and they want them recent. 6 studies on the weight of reviews, ratings and local ranking.

Discover the Satisfaction module

IFOPFrance, 1,003 people 2026

93%

of French people check reviews before choosing a shop

Reviews have become the shop window before the shop window

93% of French people check reviews before choosing a shop, up from 75% in 2021. Almost one in two distrusts a review that is two years old.

For your business

Your Google profile is seen before your shop. Recent, regular reviews there reassure the customer who is hesitating between you and next door.

IFOP for Guest Suite, 2026Read the study (new tab)

BrightLocalUnited States, 1,002 adults 2026

74%

want reviews less than three months old

A good old review is no longer enough

Three consumers in four look for reviews written in the last three months, and 97% read reviews of local businesses.

For your business

What counts is not the stock of reviews but the flow. Offering the review at the moment of the visit, every day, keeps that flow going without asking anything of the team.

BrightLocal, Local Consumer Review Survey 2026Read the study (new tab)

BrightLocalUnited States, 1,002 adults 2026

47%

avoid a business with fewer than 20 reviews

Under 20 reviews, one customer in two walks on by

Almost half of consumers would not visit a business with fewer than 20 reviews.

For your business

For a shop that is opening or a new site in a network, the first weeks matter: every visit to the till is a chance for a review.

BrightLocal, Local Consumer Review Survey 2026Read the study (new tab)

Harvard Business SchoolIndependent restaurants 2016

+5 to 9%

in revenue for one more star

One more star shows in the till

A one-star increase on Yelp leads to a 5 to 9% increase in revenue for an independent restaurant; the effect is not seen in chain restaurants.

For your business

For an independent, the rating is a commercial asset. Listening to feedback in private to fix things fast, and offering the Google review at the right moment, works on both sides.

Michael Luca, “Reviews, Reputation, and Revenue: The Case of Yelp.com”, Harvard Business School, 2011, revised 2016Read the study (new tab)

Northwestern UniversityE-commerce 2017

270%

higher purchase likelihood with five reviews than with none

The first five reviews change everything

A product showing five reviews is 270% more likely to be bought than a product with none, and purchase likelihood peaks between 4.0 and 4.7 stars before falling towards 5.

For your business

The first reviews do most of the work, and a perfect rating is not the goal. What convinces is real reviews, left by real customers.

Spiegel Research Center, Northwestern University, “How Online Reviews Influence Sales”, 2017Read the study (new tab)

GoogleOfficial documentation 2026

local ranking factors: relevance, distance, prominence

Google says it: reviews count in local ranking

Google ranks local businesses on relevance, distance and prominence, and states that more reviews and positive ratings can help a business’s local ranking.

For your business

You choose neither the search nor the customer’s distance. Prominence, though, can be built: a complete profile, regular reviews, replies to reviews.

Google, Business Profile Help, “Tips to improve your local ranking on Google”Read the study (new tab)

03AEO 9 studies

AEO: why AIs are becoming your new shop window

Your customers ask ChatGPT, Gemini or Perplexity for advice, and these AIs only cite a handful of places. 9 studies, six of them from Profound’s Research Hub, on this new shop window.

Discover the AI Visibility module

BrightLocalUnited States, 1,002 adults 2026

45%

asked an AI to recommend a local business, against 6% a year earlier

Your customers already ask an AI for advice

In one year, the share of consumers who used an AI tool to find a local business went from 6% to 45%. ChatGPT leads, at 31%.

For your business

The question “a good café near here?” is now also put to ChatGPT, Gemini or Perplexity. Knowing what they answer about your business becomes as useful as knowing your place on Google Maps.

BrightLocal, “Half of consumers are asking AI for business recommendations”, Local Consumer Review Survey 2026Read the study (new tab)

SOCi2,751 brands, about 350,000 locations 2026

1.2%

of locations recommended by ChatGPT, 11% by Gemini

Being recommended by an AI is far rarer than ranking well on Google

The locations studied appear in Google’s local top 3 35.9% of the time, but only 11% are recommended by Gemini and 1.2% by ChatGPT: 3 to 30 times harder.

For your business

A good place on Google Maps does not guarantee a place in an AI answer. The two are measured separately.

SOCi, Local Visibility Index 2026Read the study (new tab)

SOCi2,751 brands, about 350,000 locations 2026

4.3 ★

average rating of the locations ChatGPT cites

AIs recommend the best rated

The locations ChatGPT cites have an average rating of 4.3 stars.

For your business

Ratings and reviews are the material AIs read. Recent reviews, left at every visit, are the first lever for getting into their answers.

SOCi, Local Visibility Index 2026Read the study (new tab)

ProfoundUnited States, 7.5 million conversations 2026

×2

buying conversations on ChatGPT in one year

People increasingly choose by talking to an AI

Across 7.5 million ChatGPT conversations analysed, the share with buying intent rose from 13.9% to 19.2% in a year, and their volume more than doubled, to around 533 million a week by the study’s estimate.

For your business

A growing share of decisions, including “where shall we eat tonight”, goes through a conversation with an AI. Being cited there means being on the list.

Profound, “Commercial conversations in ChatGPT more than doubled in a year”, August 2026Read the study (new tab)

ProfoundMillions of prompts 2025

3 to 4

brands cited per ChatGPT answer, against 6 to 7 before

There are fewer and fewer places in an AI answer

After an October 2025 update, ChatGPT cites 3 to 4 brands per answer on average instead of 6 to 7; average brand visibility fell by 31%, and more than 85% of brands lost ground.

For your business

An AI answer is a short list. If you are not on it, the customer does not see you at all: it is not a results page you can scroll.

Profound, “ChatGPT’s entity update: Fewer mentions, tougher competition”, November 2025Read the study (new tab)

Profound56 participants, 221 tasks 2026

92.8%

of observed shopping journeys end with no click to a website

The shortlist is the new shelf

Participants shopping through ChatGPT ended 92.8% of their tasks with no meaningful click to a website; the brands finally chosen had a 24% share of voice in the answers, against 11% for the brands passed over.

For your business

Customers often decide inside the answer, without visiting your site. What the AI says about you weighs directly on their choice.

Kevin Indig, Eric Van Buskirk and Jasman Singh for Profound, “The shortlist is the new shelf”, July 2026Read the study (new tab)

ProfoundOver 32 million citations 2026

×8.4

citations of google.com in Google AI Mode in two and a half months

Google’s AI cites its own profiles more and more

Between April and June 2026, google.com’s citation share in Google AI Mode grew 8.4-fold, to second place among cited domains. Restaurants and local services are among the most affected, and the business profile becomes the first touchpoint.

For your business

For a shop, the Google profile is no longer just a link to the website: it is what Google’s AI shows. Complete, up to date, fed with recent reviews.

Profound, “Google AI Mode’s shift to citing itself”, July 2026Read the study (new tab)

ProfoundUnited States, over 2 million conversations 2026

×2.5

site visits for a brand after a mention by Gemini

Being cited by an AI brings people in

Across more than 2 million tracked conversations, a mention by Gemini makes a visit to the brand’s site within seven days 2.5 times more likely (1.5 times for ChatGPT); 42% of those visits happen within 24 hours.

For your business

A mention is not only about image: it triggers visits, often the same day. And most of them are not attributed, which is why measuring the answers themselves matters.

Profound, “The AI mention effect”, July 2026Read the study (new tab)

ProfoundAbout 80,000 prompts per platform 2025

40 to 60%

of the sources AIs cite change from one month to the next

AI answers move every month

For identical questions, 40 to 60% of the domains AIs cite are different from one month to the next (54.1% for ChatGPT), and 70 to 90% over six months.

For your business

A place in an AI answer is never secured. It is measured regularly, engine by engine, and worked on over time.

Profound, “AI Search Volatility: Why AI search results keep changing”, July 2025Read the study (new tab)

Before you quote

Read these numbers with this in mind

None of these studies is about Lecy. They describe what makes customers come back, choose and recommend a business; we built Lecy on them, but results depend on each business.
Several studies cover the United States or e-commerce: each card says so. The “For your business” line is our reading, not the authors’.
AI answers change from one session to the next: nobody can promise a place in ChatGPT. It can be measured, and what moves it can be worked on.

19 sources

Sources

Figures taken from the original publications, checked in September 2026.

Place the tag, the rest follows

NFC tag delivered within 3 days. No commitment, cancel at any time.

Get my tag Talk about your network