Most collabs do their job quietly. One of them pays for the year.

Averages are the wrong tool here

If you judge ten collabs by their average performance, you will conclude the whole thing is marginal. Most of them were: a few hundred people saw a Story, one table filled, a photograph got used.

That is not how the returns are shaped. One of the ten behaves completely differently from the other nine, and averaging it with them hides the only result that mattered.

What a clip that travels actually does

A video that catches leaves the creator's own audience entirely. Tens of thousands of people see it, sometimes a hundred thousand or more, almost none of whom had heard of you. That is not reach you could have bought at a sensible price, and it is not reach that stops when you stop paying.

It also does not stop that week. Because it keeps being served to people searching for a place like yours, one clip carries on introducing you for months.

Work out what that was worth

Take the food cost of the collabs you ran in a season. Not the menu price, what it cost you. Then ask what you would have paid to put a video in front of that many people in your area, and how long that campaign would have run for.

For most venues those two numbers are not in the same league. The one that travelled paid for every quiet collab you ran alongside it, several times over.

You cannot pick it in advance

This is the part that frustrates people, and it is worth being straight about. Nobody can tell you which collab will be the one. Not the creator, not the platform, and not us.

What you can do is take cheap swings, often, at good odds. That means keeping the cost of each attempt low by using capacity that was not selling, and shortening the odds by matching properly: a creator whose audience is actually near you, making something about a place worth making something about.

And yes, you will get the odd miss

Somebody will post something flat. Somebody will be late. It happens, and pretending otherwise is what makes the rest of this sound like a pitch.

A miss costs you what the meal cost you. It does not cost you the season. That asymmetry is the whole argument: small, capped, repeated downside, against an upside with no ceiling on it.

What to actually track

Not the average. Watch for the outlier, and count the things that outlast the week.

• How many collabs you ran, and what they cost you in product rather than in menu price.
• Whether any single one went much further than the rest, and what it was about.
• How many pieces of content you can still use in six months.
• How many reviews and searchable videos are still on the internet doing work.