Same brand,
different store experience.
A national study of how people experience convenience-store and quick-service food. Their antennae are up, they judge one location at a time, and quality is what registers.
Shoppers show plenty of goodwill toward chains overall. They have far less patience for individual locations. Most agree, 79%, that some locations of the same chain are noticeably better or worse than others.
A national study of 1,200 U.S. food-and-beverage shoppers · fielded July 3 to 10, 2026 (1,087 passed the attention screen and form the reporting base) · quotable with attribution to Seurat Analytics
Shoppers judge the location, not the chain.
The variance has structure. Our preliminary hypothesis is that it tracks how closely a chain holds its operators to one store. This is an association rather than a demonstrated effect of operating model, and the chains differ in many other ways.
Base: 810 shoppers across the six chains large enough to test, from the 892 who named a most-visited chain. 142 at the single-restaurant chain and 668 across the rest. p=0.0009.
Base: 761 shoppers who use a delivery or app ordering channel at least rarely.
The judgment is local. Shoppers are not weighing one brand against another; they are weighing your best store against your worst one, and they notice the difference.
WHAT OPERATORS CAN DO WITH THIS
- Rank sister stores against each other on repeat behavior; the customer already does
- Read reviews and complaints at the store level, where customers actually file them
- Ask of every chain-average metric: which stores is this number hiding?
- Assume the store-level judgment follows the customer into the app; the fulfilling location is a choice, not a detail
Defection is fast, quiet, and local.
Base: 587 regular c-store coffee buyers.
One pattern worth naming: frequency changes who fires a location, not how fast. Heavier shoppers switch more often, but the two-visit window holds across light, medium, and heavy alike.
said something the last time an order went wrong, and got it fixed
said nothing and ate it anyway
net change on the chain P&L when the dollars move to a sister store
These "quiet defections" are easily missed by most operators. In many instances, the customer who quits one branch shifts to a sister location run by the same company, so the holding company's P&L barely moves. There is no clear "net loss" anywhere on the books. When two locations are owned by different franchisees, the shift is also difficult to detect, because the effect is a slow leak, happening one customer at a time over the course of a year. Either way, no one is making a noisy exit.
WHAT OPERATORS CAN DO WITH THIS
- Track each store against its own predictive sales model1; branch defection shows up there before it shows up anywhere else
- Assume complaint logs undercount defection; silence is the common exit
- Win the visit after a miss; the second experience decides whether the first one mattered
After a remodel, operations drive the return decision.
as the reason shoppers returned
say a major remodel makes them more likely to visit a store (58%).
HOW THE STORE PERFORMS
- Food hotter or fresher than before
- Service noticeably faster
- Order accuracy improved
- Cleanliness held up during the rush
- Expanded selection of the products they want
WHAT THE STORE LOOKS LIKE
- New paint or interior refresh
- New signage or exterior look
- New layout
- New menu boards and kiosks
- Updated look around the pumps
Base: 685 shoppers who noticed recent changes at a store; each named the single change that most affected their decision to return.
Consumers care more about changes that affect their experience than changes to how the store looks. The remodel that wins is the one they can taste, time, and count on, and it beats the one they can only see by 2.5 to 1.
19% of shoppers who noticed changes said none of them affected their behavior at all. Improvement does not register automatically; that is the argument for measuring whether yours did.
WHAT OPERATORS CAN DO WITH THIS
- Sequence the operational fixes before the cosmetic spend, not alongside it
- If service will degrade during construction, close rather than run badly: two bad visits can cost a regular, and winning one back costs far more than keeping one
- Verify the operational lift after the remodel instead of assuming the investment ran it
The stickiness is earned, not automatic.
vs. 70% for shoppers of every other named chain
A good food experience converts a customer. Keeping that customer is a separate question, and the answer differs sharply by operator. Across all c-store prepared-food buyers, 27% still mostly go back to fast food after a good experience; the spread behind that average is what matters. Note that this measures depth of usage among shoppers who already buy c-store prepared food, not the category taking share: shoppers who report a swap are heavier food-away-from-home users on both sides, and their c-store share of total food occasions is no different from anyone else's.
WHAT OPERATORS CAN DO WITH THIS
- Treat trial and repeat as two separate problems; the first good visit converts, the second one decides whether it lasts
- Protect quality consistency once trial starts; the first good verdict converts, and the first bad one reverses it
- Watch trial-to-repeat store by store2, the same lens as Finding 02: some stores convert the new habit, and others leak it back
The base is prepared-food buyers, but the question itself asked about buying food, not prepared food specifically. Chain grouping follows industry reputation, not a published definition; bases of 103 and 416, p=0.004.
What quiet defection is worth at your store
In this study, 29% of shoppers have fired a location of their most-visited chain, half of them within two bad visits, and most without complaining. Put your own numbers in to size what traffic drifting to a sister store means in dollars.
Illustrative arithmetic on your inputs and the study's national rates. It is a sizing, not a measurement; a measured read of your own network is a different exercise.
What's behind this page
The study fielded 1,200 respondents across six modules and 266 variables. This page publishes four findings. The rest is inventory: here is the shelf, without the values.
Screener and habitsn=1,087 · 30+ items
- Not covered on this page: the full frequency and channel distributions, daypart mix, and every named-chain frequency table.
The location-experience modulen=1,087
- Not covered on this page: the full agreement battery beyond the two items shown, and the attention-triggers ranking.
The most-visited chain modulen=892 with a named chain
- Not covered on this page: the taste-consistency figures, the repeat-switcher rate, the full went-wrong response set, and the switching-to-defection link.
Changes noticed and the remodel batteryn=685 who noticed changes
- Not covered on this page: the full 18-change ranking in order, and the complete 12-item win-back ranking.
The occasion shiftn=519 c-store food buyers
- Not covered on this page: the reasons-for-shift ranking and the ordering-option adoption figures.
Trust signals, coffee, and GLP-1n=1,087
- Not covered on this page: the trust-signal ranking and the full GLP-1 cut.
Every finding cuts by: age, gender, household income, urbanicity, census region, ethnicity, visit frequency (three ways), prepared-food heaviness, and primary chain type. Named-chain tables exist and are available by request. No individual chain is named on this page.
Go deeper
The topline above is free and quotable with attribution. Underneath it, the ladder:
The full cross-tab set: every finding by every dimension, with bases, wording, and the data tables behind each chart.
Buy the deep diveYour segment, region, or question, run against the file. Any fees credit in full against a diagnostic signed within 90 days.
Request a cutInstrument and respondent-level file under license, for research teams that want to run their own analysis. Write to research@seuratanalytics.com.
Ask about licensing¹ A predictive sales model is a store-level forecast of what a location's sales would have been, built from that store's own history and from the movement of comparable stores around it. When a store's actual sales run consistently below what its model predicts, that gap is the earliest visible evidence of quiet defection, and it appears even while chain totals hold steady. ↩ back
² Trial-to-repeat is read from a store's transaction pattern over time. New-customer volume shows the trial, and the share of that volume that returns in later weeks shows the repeat. Comparing those two curves, store by store, shows which locations convert a first visit into a habit and which locations lose it. ↩ back
A fuller description of how Seurat measures store-level effects is on the main site: seuratanalytics.com
HOW THE STUDY WAS RUN
- Sample
- 1,200 U.S. adults fielded; 1,087 analytic base after a reading-attention screen
- Population
- Adults who visit a quick-service restaurant or convenience store for food or beverages at least monthly
- Field window
- July 3 to 10, 2026, via the Prodege online panel (study IO-127548)
- Coverage
- National, with a Northeast core sample across CT, NY, NJ, PA, MD, DE, and OH
All figures are reported on the analytic base of 1,087 respondents who passed the reading-attention check, except where a question was asked of a defined subgroup: location-switching incidence is among the 892 who named a most-visited chain, and switching speed among the 255 who reported switching; remodel figures are among the 685 who noticed recent changes; occasion-shift figures are among the 519 who buy prepared food at convenience stores. Frequency bands in the toggle are visit frequency of the respondent's most-visited chain: light is monthly or less (n=295), medium is 2 to 3 times a month (n=298), heavy is weekly or more (n=299). Percentages for check-all questions do not sum to 100. Full data tables and question wording available on request. Findings are quotable with attribution to Seurat Analytics.
Shoppers reward what stores do well. Many of the best changes have already been made, and measured.
The operational changes that win customers back have been run at multi-site retailers already, and their results are on the record. Seurat keeps that evidence: the proven levers for stores like yours, ranked by impact, measured across the stores that ran them and matched to stores like yours.