CTR vs. Conversion Rate in Restaurant Ads: We Were Wrong About CTR
CTR measures clicks; conversion rate measures the next action. Neither proves a restaurant visit, paid check, incrementality, or profit.
CTR tells you how often an ad impression became a click; conversion rate tells you how often a defined group completed the next action. Neither one proves a restaurant visit or sale, and we are retracting our earlier rule that treated CTR as if it pointed away from purchases.
The corrected rule is less dramatic and more useful: compare ads doing the same job, pair platform-reported CTR with the next meaningful action, then keep following the guest toward the dining room and a paid check. That is the measurement job behind restaurant marketing ROI, not a reason to replace one magic number with another.
What exactly did we get wrong about CTR?
We made a broad claim from a comparison that mixed video and static ads.
“CTR is inversely correlated with purchase.”Our former internal rule, retracted after the August 28, 2026 evidence review
The claim implied that a high click-through rate was suspicious and that a low one might signal better buyers. That sounded plausible because curiosity clicks and accidental taps are real. But the analysis had confused format with performance: videos and statics behaved differently, so their combined CTR-to-purchase pattern did not establish a dependable relationship.
When the August 2026 review compared statics with statics in our own software-acquisition advertising, the high-CTR group cost $334 per purchase and the low-CTR group cost $385. Our second-best ad in that account carried a 3.4% CTR. Those figures do not prove high CTR causes better results either. They show why the earlier anti-CTR rule had to go.
This was not a restaurant campaign sample, and its purchase was a software purchase, not a seated guest or restaurant check. The transferable lesson is about method: do not compare unlike formats, then turn the pattern into a universal creative law.
CTR and conversion rate answer different questions
CTR diagnoses the ad-to-click handoff; conversion rate diagnoses whichever handoff you define next. A report becomes misleading when it says “conversion rate” without naming the event and denominator.
Use explicit formulas:
- Link CTR: eligible tracked link clicks divided by eligible impressions, using one consistent platform definition.
- Landing-page-arrival-to-action conversion: unique eligible landing-page arrivals with at least one identified completed action divided by those same unique eligible arrivals whose full fixed action window elapsed and whose action outcome was reconciled and known. The numerator is the successful subset of that exact denominator. Hold early actions out until the whole cohort matures; report open, immature, ineligible, unresolved, and automated records separately. If an arrival cannot be joined to the person who responded, publish arrival and completed-action counts separately and leave the person-level rate unknown.
- Action-to-visit conversion: verified completed visits divided by the exact mature group of eligible identified actions used in the denominator.
- Check-match coverage: the denominator is eligible verified completed-visit events whose fixed reconciliation window fully elapsed and whose matched or unmatched outcome was reconciled and known; the numerator is the subset with its own matched paid, non-voided settled check. Report immature, ineligible, open, and unresolved records separately.
A signup, claim, or reservation can be a conversion inside an ad report without being a restaurant visit. A visit can occur without a usable check match. A paid check can be directly attributed under a stated rule without proving the ad caused it.
That distinction is why a reservation needs its own arrival and check evidence. If that is the weak link in the report, see what a reservation is actually worth.
The evidence did not replace CTR with conversion rate
Conversion rate is closer to the business result only when its event is closer to the business result. A click-to-claim rate can expose a confusing landing page, but it still cannot tell you how many claimants arrived or what they bought.
Our August 2026 review found another reason not to stop at CTR. Reels with platform-reported CTRs of 14%–18% initially looked like they might be collecting accidental taps.
For those Reels placements, the platform-reported landing-page-view rate was 85%–87%. That is recorded platform activity, not proof that a person loaded or read the page.
Feed placements in the same software-acquisition analysis recorded a 30%–45% landing-page-view rate.
That observation does not make Reels universally better, and it says nothing about restaurant visits or profit. It shows that high CTR needed a second diagnostic. In that case, the platform-reported landing-page-view rate weakened the accidental-tap story.
The replacement is a pair of questions:
- Did the ad earn the intended kind of attention?
- Did those people take the next action the ad promised?
Then ask the restaurant questions: did eligible people arrive, did the visit reconcile to a paid check, and did the campaign make enough contribution to justify its full cost?
Read CTR as a clue, not a verdict
The same CTR can describe four very different problems. Compare ads with the same audience, offer, placement, landing page, budget treatment, and measurement window before reading the pattern.
| What you see | What it may mean | What to check next |
|---|---|---|
| High CTR, strong next action | The ad and page may be working together | Mature visits and matched checks |
| High CTR, weak next action | Curiosity, a promise mismatch, or a page problem | Search terms, landing-page arrival, and action completion |
| Low CTR, strong next action | The people who click may fit, but the ad earns too few chances | Hook, first frame, and audience delivery |
| Low CTR, weak next action | The problem may sit in the offer, audience, ad, page, or tracking | Test each handoff before replacing everything |
These are diagnoses, not automatic pause rules. A low CTR does not rescue a bad campaign, and a high CTR does not make a campaign profitable.
One restaurant owner captured the real frustration:
“Although we do social media and digital marketing, we can't figure out why it's not working.”Restaurant owner
The answer is not another colored arrow in the ad account. It is a record that shows where an eligible person stopped.
Wait for the restaurant result before judging the ad
Visit and check rates need fixed windows and closed outcomes.
For action-to-visit conversion, start with eligible identified actions, wait until the full arrival window has elapsed, reconcile every visit outcome, and place only the successful subset with verified in-person completion in the numerator. Early visits wait with their cohort. Report immature, ineligible, open, and unresolved records separately.
For check-match coverage, start with eligible verified completed-visit events, let the fixed reconciliation window elapse, and determine a known matched or unmatched outcome. Count only visits with their own paid, non-voided settled check as successes. A claim, reservation, seated party, cover, unmatched check, canceled record, or voided check is not a substitute.
Test the reporting path before spending enough to care about the result. Create a real eligible action, verify the completed visit, and confirm its paid check appears under the correct campaign and amount. Then use a second test check: confirm it appears first, void or refund it, and make sure both the eligible check count and its sales leave the report.
CTR needs no arrival window, so it will stabilize first. That does not give it the right to decide the campaign before the slower outcomes mature.
A paid check is still not proof of incremental profit
Directly matched sales are attribution, not incrementality.
They tell you which paid, non-voided settled checks the chosen matching rule connected to the campaign, less recorded refunds. They do not tell you how many of those guests would have come anyway. The word “estimated” in restaurant attribution deserves the same scrutiny as “conversion.”
Contribution requires two cost ledgers. First subtract the ordinary variable costs of the matched checks: food, packaging, processing, delivery fees, fulfillment costs, and incremental paid service labor. Then subtract the full campaign cost: media, creative production, influencer compensation, commercial-use-rights payments, software fees, agency fees, separately billed reporting, separately billed cleanup, actual incentive-fulfillment cost, and incremental paid campaign-management labor when each applies.
Count every cost once. If a free item is in incentive fulfillment, do not repeat its food or labor under the check. Allocate shared software costs, agency costs, and production costs with one written rule, such as active campaign-days, and use it consistently across locations and campaigns. Value owner-management time separately at a chosen hourly amount, without also counting those hours as paid labor. If an ad will run from an influencer's account, keep that account permission separate from the commercial-use license and add a permission payment only when the written agreement charges one.
What remains is matched contribution under the stated attribution rule, not causal lift or total restaurant profit. Incrementality needs a prewritten control, such as a comparable holdout location, audience, or time window. Profit still includes fixed operating costs outside the campaign ledger. A sensible restaurant ROAS target depends on those economics, not CTR alone.
What changes after this correction?
CTR goes back to being a useful early warning. Compare like with like, name the click type, inspect the next action, and do not let an early platform metric overrule a mature restaurant result.
The new rule is deliberately not universal. A high CTR may be good, bad, or irrelevant to the campaign's final economics. A low CTR may expose weak packaging or a narrow but qualified audience. The evidence decides only after the event path and the restaurant outcome are both visible.
If your current report stops at clicks, Feast Revenue Analyzer helps you examine marketing activity alongside restaurant sales. Explore what your data says about the results, then use those findings to plan your next investment instead of letting CTR decide the budget.