97% of surveyed consumers read local-business reviews. Just 69% wrote one in the preceding year.
Those two figures, from BrightLocal’s 2026 Local Consumer Review Survey, describe different kinds of participation. They do not tell you how many restaurant guests left unhappy without saying anything. But they do expose a mistake you can make when assessing a service fix: treating the feedback people publish as a complete account of the experiences people have. citeturn912320view3turn997296view2
If you’re asking, “Are restaurant reviews representative of customers?”, this chart offers a useful starting point—not a verdict on your dining room.
About the visual: This original chart is an editorial test informed by Orbit’s findings about content and visuals—not evidence that this particular chart will improve engagement, comprehension, or traffic. citeturn912320view0
What the two bars actually measure
The 97% figure concerns reading reviews of local businesses. The 69% figure concerns writing a review during a specified period: the preceding year. BrightLocal’s survey sample comprised 1,002 U.S. consumers. citeturn912320view3turn997296view2
Three distinctions matter before you apply those numbers to restaurant guest satisfaction.
First, reading and writing are different behaviors. You might read reviews of six restaurants before booking a birthday dinner, then write nothing afterward. Your use of reviews does not require you to contribute one.
Second, the time frames are not interchangeable. The writing measure explicitly covers the preceding year. You should not silently attach that same window to the reading figure.
Third, the subject is local businesses, not restaurants alone. A respondent’s review activity could involve a dentist, a plumber, or a hair salon. These figures do not isolate restaurant visits, restaurant reviewers, or diners at your location.
The unit also matters. These are percentages of people, not percentages of visits. Someone who wrote one review after dozens of purchases can count as a review writer. That does not mean those dozens of experiences are documented.
The chart therefore shows a difference in customer review participation. It does not show how much of your restaurant’s service history is visible online.
Why 28 percentage points cannot become an unhappy-guest estimate
Subtracting 69 from 97 produces a 28-percentage-point difference. The arithmetic is straightforward. The interpretation is where things go wrong.
That difference is not the share of guests who were unhappy but stayed silent. Neither measure asks whether someone was unhappy. Neither measures whether they complained to a server. Neither identifies whether a restaurant corrected their problem.
Nor should you label the difference “readers who never write.” The two headline percentages do not provide the respondent-level overlap, and “did not write in the preceding year” is not the same as “never writes.”
Even the 31% who did not report writing a review in that period cannot be classified by satisfaction from this result. That group could include delighted customers, disappointed customers, indifferent customers, and people who prefer private feedback.
Imagine two tables after the same Saturday dinner service:
- One enjoys the meal, pays, and leaves without posting.
- Another finds the dining room uncomfortably loud, decides not to return, and also leaves without posting.
Both create the same public signal: nothing.
That is the problem with interpreting silent restaurant guests. Silence does not carry a satisfaction label.
The survey also does not establish that reviewers are systematically more negative—or more positive—than everyone else. It measures participation, not the direction or size of restaurant feedback bias. You cannot use it to calculate a “true” satisfaction score behind your star rating.
Public feedback is a selected view, not a guest census
Participation bias becomes relevant when the people who supply feedback differ in meaningful ways from the people who do not.
A guest may be more inclined to write because an experience was exceptional, because you requested a review, or because they routinely post about local businesses. Another may avoid public reviews even when something deserves attention. Those are possible selection mechanisms, not findings established by these two percentages.
The operational point is narrower: you should not assume that the people you hear from represent everyone you served.
Public reviews remain useful. A detailed account of a missed allergy handoff deserves attention regardless of whether the reviewer represents the average guest. Repeated descriptions of cold entrées can help you identify something worth investigating.
But usefulness is not the same as representativeness. A review can accurately describe one visit without telling you how frequently the problem occurs across all visits.
That distinction matters especially after you make a change. Suppose you adjust the host handoff to reduce waits after guests check in. Two weeks pass without another online complaint about seating delays.
You have observed no new published complaint. You have not yet established that the handoff works consistently. The restaurant may have served fewer parties, the affected guests may not have posted, or the problem may genuinely have improved. Review silence alone cannot distinguish those explanations.
Before you call the fix successful, look beyond the review feed
For that host-handoff change, check the evidence you already have: direct guest comments, manager notes, and observations from the host stand. Are guests still returning to ask whether they have been forgotten? Are servers still finding seated tables that nobody has greeted? Did anyone mention the wait privately even though no review appeared?
Those sources have limits, too. A floor observation covers only what someone saw; direct feedback still depends on someone speaking. They are additional views, not a hidden census.
But checking them prevents one narrow signal from becoming an oversized conclusion.
“No new reviews mention the wait” is a defensible statement. “The wait problem is fixed” requires evidence about the wait.
Before you declare a fix successful, check existing direct feedback and what your team is actually seeing on the floor. The absence of a complaint tells you what was reported—not necessarily what guests experienced.