What 77 Million Guests Told Us About Being Known
Surveys measure what brands say. We measured what guests actually did, 113 million transactions' worth, and what being known is worth to a restaurant's P&L.
Guests changed their expectations before most restaurants changed anything. In the AI age, people are known by the businesses they buy from. Remembered, recognized, reached personally. And they notice when they aren't. Boston Consulting Group (BCG) has built a whole research program on this shift (Personalized: Customer Strategy in the Age of AI) and scores brands on how well they deliver it (the Personalization Index).
McKinsey's What US Consumers Want from Restaurants in 2026 (Ben Mathews and Katharine Mattox, January 2026) reads the pressure from the diner's side. Value and pricing sit at the front of consumers' minds, and among diners planning to cut back, most said they would trade down within their restaurant of choice, using more promotions or ordering fewer and cheaper items, rather than switch to a cheaper restaurant. McKinsey's recommendation to operators names personalization as the next frontier, with AI tailoring offers to generational preferences and deepening brand connection.
Read that trade-down finding again, because it's the whole argument in one line: guests aren't trying to leave you. They're trying to spend less with you. Whether they stay is decided by whether you're still there when they decide.
We agree with the premise. But there's a limit to what surveys and mystery shopping can see. They measure what brands say and deploy. We measure what guests actually did. Our 2026 Restaurant Guest Engagement Report analyzed 113 million credit-card transactions from 77 million unique guests across 3,200-plus locations and $2.7 billion in sales. This piece is what those transactions say about being known, and what it's worth.
You Can't Be Known by Guests You Can't See
In most industries, identity comes with the transaction. A bank knows its customers by definition. A retailer gets a login and a ship-to address. In a restaurant, most guests pay and walk out. No account, no email, no name.
The loyalty program knows its members, and knows them well. That's exactly the problem: it's a well-lit room in a mostly dark house.
Loyalty data often overstates guest engagement. Your most enthusiastic guests opted in, and reading the whole base off that slice misleads the plans built on it.
This isn't an argument against loyalty. An opted-in guest is a real relationship at very reasonable cost. But you can't make up for what happens outside loyalty with loyalty alone.
So the whole personalization chain collapses at the first link: identification. You can't send the right message at the right time to a guest you can't see.
From the guest's seat, none of this is dramatic. They came in twice in March. They liked it. Nobody ever spoke to them again, and there are a dozen places within ten minutes of their house. They didn't leave you. They forgot you. Marketers feel this daily, pulling data from multiple sources, building the list, and still watching most transactions float by unattached to anyone they're allowed to reach. Total visibility isn't the goal; no one gets that. The realistic play is making the matchable portion of those unseen guests knowable and reachable, and the brands doing it are pulling away.
The Four Moves
The high-performing brands in our data aren't running secret plays. They run the same four moves, in order, and compound them.
Before any campaign question comes the identity question: what share of transactions can you connect to a person you're allowed to reach? Everything downstream runs through that number, because you can only talk to the guests you know. And the talking is where the gap shows up: guests reached with 1:1 marketing came back at 35% last year; everyone else came back at 14%.¹ A gap that wide is the argument, and a brand's identification rate decides how much of the base that lift can even reach.
| Guest group | Retention rate |
|---|---|
| Guests reached with 1:1 marketing | 35% |
| All other guests | 14% |
A guest you can talk to directly spends about a quarter more than one you can't: $72.18 in sales per guest reached one-to-one last year, against $58.91 for everyone else.¹ Direct doesn't mean loud, and it doesn't mean more email. Most brands already send plenty. The question isn't volume, it's who the message reaches and what it changes: the right offer, at the right time, to a person, not a list.
Guests rarely quit a restaurant. They drift, and then they've forgotten you. Left alone, almost no lapsed guests come back, about 1.4 in 100. Reached one-to-one, 6.4 times as many return. Better still is not needing the win-back at all: catch a regular the moment their rhythm slips, while they're still a guest you're keeping.
| Group | Returns per 100 |
|---|---|
| Left alone (organic) | 1.4 |
| Reached one-to-one | 9 |
Not opens. Not clicks. Dollars. Retained and re-engaged guests drove nearly half of trackable sales across our analyzed portfolio last year, more than $400 million, and 77% of total retained sales came from guests marketed one-to-one. The report's own conclusion is that retention has to move from a marketing KPI to a core business strategy. Put plainly: it belongs on the P&L, not in a campaign report.
The full 2026 Restaurant Guest Engagement Report goes deeper on all four moves.
Get the ReportWhat the Gap Is Worth
Start with how the industry budgets. The bulk of restaurant marketing spend still chases acquisition while the guest base leaks out the bottom of the barrel. That's not a growth strategy. That's paying premium prices to replace guests you already paid to win: replacement spending, dressed up as growth.
| Flow | Guests | Cost |
|---|---|---|
| New guest won | +1 | ~$30 to acquire |
| Active guests lost | −1.4 | ~$3 each to have kept with weekly email |
Now do the math underneath it.
For every new guest restaurant brands won in 2025, 1.4 active guests walked away. Winning that guest costs roughly ten times what keeping one with regular email does, and it barely pays for itself. A first visit breaks even at best. The money was always in the second visit, where the guests you can reach directly come back nearly three times as often.
| Group | Return rate |
|---|---|
| New guests reached directly | 32% |
| Everyone else | 12% |
Identification sets the ceiling. Guests reached one-to-one came back at 1 in 3; everyone else at 1 in 7. How many guests a brand can identify decides how much of the base that lift can even reach.
Direct reach changes every metric. A guest reached one-to-one spends about a quarter more than one who isn't, and 6.4 times as many lapsed guests come back when reached directly.
Retention is a P&L line, not a campaign stat. Retained and re-engaged guests drove nearly half of trackable sales, and 77% of those dollars came from guests marketed one-to-one.
The brands on the right side of that retention gap didn't out-spend the others. They knew more of their guests, and talked to them.
Three Actions for 2026
The report closes on three steps. Here they are in the terms of this argument:
You can't fix what you can't see. Connect transactions to real guest profiles, so you can tell who's active, who's churning, and which way retention is trending. Then hold that number to the same discipline and visibility you already hold sales to.
Loyalty reaches a fraction of the base. The report puts direct reach through loyalty at about 20% of guests, leaving the other 80% unknown and unreachable. Identifying more of them, beyond loyalty, is what raises the ceiling on everything else you do.
Once guests are identified, keep the brand in front of them. Don't wait until they disappear. Retention comes from communication that's ongoing, not occasional.
Go Deeper
The full 2026 Restaurant Guest Engagement Report, the behavior of 77 million guests in detail, is one click away.
Get the 2026 Restaurant Guest Engagement ReportAnd if you want to know what your own guest data would say, how many of your guests you can actually see and what the unseen ones are worth, that's a conversation we have with restaurant brands every week.
¹ Throughout: "one-to-one" means guests receiving attributable 1:1 marketing; "everyone else" includes guests whose outreach can't be attributed, reached by general advertising, broad social targeting, or nothing at all. The comparison is observational: same brands, same period, split by what each guest measurably received. Definitions follow the report's terminology section.