Menu Test Lab: A 2-Week Protocol for Validating New Items

Instead of guessing which 22 new dishes belonged on the permanent menu, a Cihangir restaurant ran a 14-day test lab. Six items were quietly cut; the rest survived with hard numbers behind them.

Three Clean Thresholds in 14 Days

The backbone of the protocol is three measurable thresholds: at least 80 sales per item, an average customer rating of 3.8/5 or higher, and a kitchen ticket time under 7 minutes. Miss two and the item is dropped automatically; miss all three and the decision needs no debate.

Week one is the warm-up. Week two is the real decision window. A low sales count in Week 1 may be a service problem — repeated in Week 2, it's a product problem.

Collecting the Data Honestly

A micro-survey ("Did you like it?") appears under each test item in the QR menu, visible only to the table that ordered. The POS tracks sales count; the kitchen display logs cook time in seconds.

• Servers log a 60-second verbal feedback note at the end of each shift.
• Waste and half-eaten plates are counted separately — the silent cancellation signal.
• If Friday night load breaks the kitchen-time threshold, fix prep flow first, not the dish.

The Story Behind the 6 Cuts

Of the six items dropped in Cihangir, four blew through the 7-minute kitchen threshold; two had decent ratings but never crossed 80 sales. A long-ticket dish slows the entire Friday service — and you'll see it in the tip line of nearby tables.

In place of the cuts, three surprise winners — each with over 120 sales in the test — earned permanent slots. Without the data, the call would have stayed with the chef's gut.

FAQ

How do I label test items on the menu? Add a "New — limited run" badge: it draws attention and pre-explains the absence if the item gets cut.

Are 14 days enough for every item? Yes for any dish hitting 6+ orders per day. For niche categories (vegan, gluten-free), extend to 21 days.

Where does food cost fit in? Only after the three thresholds pass. Then check if cost-of-goods exceeds 32 percent. If the item fails the three, you never need to argue about cost.

Found this helpful? Share it.