Why Some Influencer Videos Work: One Restaurant, Two Opposite Internal Ad Rankings

Two influencer ads from the same restaurant received opposite internal performance labels in a July–August 2026 Feast observation, but the record does not identify the ranking metric or prove why. The defensible lesson is to test videos under comparable conditions and wait for mature guest and check outcomes before deciding that the video, restaurant, or influencer caused the result.

That contrast changes how you should run influencer marketing for your restaurant: commission several distinct videos, test them under comparable conditions, and judge each one by mature guest and check outcomes rather than by appearance alone.

Opposite internal rankings justify a controlled test, not a causal verdict

The paired labels make a controlled video test worth running; they do not show that the video caused either ranking or that the gap was material. The business was the same, yet the two influencer partnership ads were internally described at opposite ends of that restaurant's history.

“Two partnership ads added to his existing campaign; one is historically his worst, one is historically his best.”Feast internal operating note, July–August 2026

This was an operating observation, not a controlled experiment. The record does not establish the metric used to rank the ads, equal spend, equal platform-reported activity, identical audiences, identical offers, or identical run windows. It does not prove that influencer videos always vary this widely, nor that the video alone caused the difference.

The observation leaves a useful question open: when the restaurant is the same, how much of the result changes when you change the video while holding the other conditions steady?

A strong restaurant still has to test the ad

A restaurant supplies the raw material, but the paired record cannot tell us which part of either ad helped or hurt its internal ranking. The opening shot has to stop the right person, the food has to become specific and desirable, and the next step has to make sense. A beautiful dining-room tour can earn attention while doing little to move someone toward a trackable action.

That is why “good content” and “good direct-response creative” are different judgments. Production quality, organic engagement, and the influencer's personality can all be useful. None of them, by themselves, proves that an identified guest arrived or that a paid check settled.

The reverse is also possible. A rough-looking video can be the clearer ad if it shows the dish, gives a believable reason to visit, and sends the viewer to one measurable destination. The paired observation does not tell us which ingredient separated the two ads. It tells us not to collapse all of those ingredients into one verdict about the restaurant.

Change only the video if you want to learn about the video

A useful influencer-video test holds the commercial conditions steady. Write the campaign plan before spending:

If one ad runs from the restaurant's account and another runs from an influencer's account, you changed more than the video. A partnership ad needs separate account-level permission when it runs from the influencer's account; a commercial-use license for restaurant-account advertising is a different permission. Treat the identity shown on the ad as part of the test rather than claiming a video-only result.

Agree on files and permissions before the visit. Finished posts, raw footage, commercial-use rights, editing rights, and influencer-account advertising permission are different things. Moving an influencer video into a paid ad becomes much harder when those terms are negotiated after you find a winner.

Compare customer response, not just views

A video can get more clicks while another produces more reservations or purchases. Decide which result matters for the campaign, then compare both ads using that same measure.

Where your setup allows it, connect the reservation, offer claim, or order to a customer and a POS check. Compare the resulting sales with the costs of the ad, the content, and serving those guests. Sales are not profit, and directly matched sales do not prove every purchase happened because of the ad.

You don't need a separate measurement manual for each video. Use the same tracking method for both, give people time to act, and investigate a tracking gap before treating missing records as missing customers. The guide to tracking restaurant sales from Facebook ads explains the available methods.

If both videos struggle, look beyond the creative

If changing the video doesn't improve the result, the shared offer, audience, destination, or visit experience may be the problem. A better-looking edit won't fix a booking page that doesn't work or an offer people don't want.

That also changes what “underperformed” means. A video can miss the response goal yet still supply useful footage, explain the restaurant well, or work organically. Decide the job before testing it; otherwise an awareness video gets punished for failing a direct-response test it was never built to pass. Diagnose the job before calling an influencer video a failure.

“Ideally, we just test it and see how it works instead of going through the process of revisions when we don't even know this could actually work.”Feast founder, April 2026

The practical decisions are simple: keep a mature winner in rotation, repair a video when one identifiable part of the response path fails, stop a clearly losing paid test under your written rule, and retest when you materially change the video or offer. Do not keep spending because a post looks good, and do not kill a test whose visit or check window is still open.

You need another video ready after one test earns more spend

A repeatable testing program needs a pipeline, not one heroic shoot. The hard part is getting distinct local influencers approved, booked, reminded, through the visit, and back with complete files and the permissions required for the next test.

“Most people we work with don't know how to make content. I have to teach them. The biggest bottleneck to getting results with our product is making content.”Feast founder, March 2026

Remy is built for that operating gap. You choose whom to approve and which videos deserve a test; the software finds influencers, books visits, sends texts and reminders, and follows up for files so repetitive coordination does not set the pace. Use Remy to keep the video pipeline moving; use separate ad and check reporting to decide which video wins.

Remy is $300 a month per location: $150 for Remy and $150 for creator recruitment ad spend. Learn more about Remy.