Why your ad spend isn’t working

You just saw a number that probably didn’t sit right. Money spent, very little to show for it. Before going any further, it’s worth asking the obvious question: is that number real, or is it just a calculator built to make you want to buy something? It’s a fair thing to wonder. So here’s the honest answer: every figure on this page comes from Meta’s own published advertising mechanics. Not from us, and not from guesswork.

Here’s the short version: Meta’s ad system needs to see roughly 50 purchases a week, for every single ad you run, before it has enough data to start spending your money wisely. Most independent restaurants never come close to that number. A typical $1,000-a-month budget produces only a fraction of the orders the algorithm needs, so it never finishes learning. It just keeps guessing, week after week, and you keep paying for the guesses.

And the guessing isn’t the only cost. Every person who clicks your ad and doesn’t order is gone the moment they leave the screen, with no way to bring them back unless you pay Meta again to try to find them. That’s the real loss: not just the ad spend, but the customers you already paid to reach who simply disappear.

When you search for what a “normal” restaurant ad campaign should look like, you’ll find numbers like these: a median click‑through rate of 1.85%, a median cost per sale of $38.15, and a median return of $1.56 for every dollar spent, before costs — pulled from an analysis of nearly 35,000 food and beverage brands running ads on Meta throughout 2025.

But that pool includes national chains and well‑funded operators spending tens of thousands of dollars a month, with marketing teams whose full‑time job is optimizing every dollar. If you’re spending $1,000 a month and managing it yourself between shifts, you’re not running the same race. The comparison was never fair to begin with.

Meta’s ad system needs to see roughly 50 purchases a week, for every single ad you’re running, before it has enough information to figure out who’s actually going to buy from you. Until it crosses that number, the system isn’t optimized. It’s still guessing, indefinitely.

So far, this has been about why the ad spend itself underperforms. But there’s a bigger number sitting underneath all of it, and it has nothing to do with Meta’s algorithm. Across all online shopping, only about 1.4% of website visitors make a purchase. Paid social ads, the exact kind running on Facebook and Instagram, convert at just 0.9%, the lowest of any traffic source there is.

That means even a perfectly optimized campaign, run by a brand with an unlimited budget, would still lose more than 99 out of every 100 people who click.

Technically, yes. Getting one ad out of the learning phase takes roughly $6,400 a month, out of reach for most independent restaurants. But even with that budget, the deeper problem doesn’t go away: that 0.9% conversion rate isn’t a learning‑phase symptom, it’s the rate even fully optimized paid social campaigns get. A bigger budget fixes the algorithm’s confusion. It doesn’t fix what happens to the customer after they leave.

The fix isn’t spending your way to a better‑optimized ad. It’s building a way to keep the people the ad already brought you, so a non‑converting click stops being a dead end and starts being the beginning of a relationship you control.