Every restaurant operator has gut instincts. Most of the time, those instincts are pretty good — you don’t survive in this business without learning to read a room. But gut instinct alone can also lead you down some expensive roads.

There’s a better way. Before you commit to a big change, treat it like an experiment. Make a guess about what will happen, test it, and look at the results.

How Seven-Eleven Does It

Seven-Eleven Japan gives their store managers access to the previous week’s sales data and asks them to make a guess: based on what sold, what should we stock more of next week? What should we cut?

They’re not just restocking shelves. They’re making a specific prediction — "I think this item will sell better if I give it more shelf space" — and then checking whether they were right. Over time, every store gets smarter about what their customers want because they’re always testing and learning.

The same idea works in a restaurant.

What This Looks Like on the Line

Say you’ve got a hunch that a spicy chicken sandwich would sell well as a lunch special. Instead of adding it to the permanent menu, printing new menus, and buying a case of chicken, try this:

  1. State your guess. "I think a spicy chicken sandwich priced at $14 will sell at least 15 per day at lunch."
  2. Run it for two weeks as a special. Train your servers to mention it. Put it on the specials board.
  3. Track the numbers. How many sold? What was the food cost? Did it slow down the line? Did lunch traffic go up?
  4. Decide based on what happened, not what you hoped. If it sold 20 a day and the margins were good, put it on the menu. If it sold 5, it’s not a failure — you learned something useful without committing to a menu print run.

This Works for Everything

Not just menu items. You can test a lot of things this way:

  • Staffing changes. "I think adding a runner on Friday and Saturday will cut table turn time by 10 minutes." Try it for three weekends and check.
  • Hours. "I think opening an hour earlier on Sundays would bring in a breakfast crowd." Test it for a month before making it permanent.
  • Online ordering. "I think offering a 10% discount for first-time online orders will bring in new customers." Run it for 30 days and see if those customers come back.
  • Vendor switches. "I think this new produce supplier will save us 8% on weekly orders." Run both side by side for a few weeks and compare quality and cost.

The Point

A wrong guess isn’t a failure. It’s only a failure if you don’t learn from it. When you write down what you think will happen and then check whether it did, you learn something every time — whether the idea works or not.

The best operators aren’t the ones who always guess right. They’re the ones who test small, learn fast, and stop pouring money into things that aren’t working.