The 154 Problem.
Your on-time rate looks like a win. The guests it leaves out are the ones you cannot afford to lose.
Here is a real snapshot from a KitchenAI dashboard. One location, 910 orders. 78.7% landed within five minutes of the estimate the guest was given. On a report, that is a strong number. Most operators would take it and move on.
Then look at the box in red. 154 orders ran late by more than five minutes.
That 78.7% is an average. No guest experiences an average. Each one experiences their order. And for 154 of them, the restaurant made a promise and missed it.
The promise
An estimate is a promise
When you quote a guest fifteen minutes, you have set a clock in their head. They plan around it. They tell the table the food is coming. They decide whether they have time to grab a drink or take a call.
Blow past that quote by seven minutes and you did not just run seven minutes behind. You broke something you said. The guest does not grade you against your kitchen's capacity. They grade you against the number you gave them.
154 late orders is 154 broken promises. Some of those guests shrug it off. Some do not come back. A few leave the review. You rarely find out which is which, which is exactly the problem.
Direction matters
Late is not the same as early
Notice the dashboard splits the misses into two boxes on purpose.
40 orders came out early. Early is a minor issue. Food might sit a minute. The guest might not be back at the counter yet. Nobody writes a review about early.
154 came out late. Late is the miss that costs you. Late is the daypart that gets slammed and never recovers. Late is the third-party order that shows up cold and turns into a refund. Over and under are not the same failure, and rolling them into one number buries the one that actually hurts.On-time rate matters.
The blind spot
The metric most operators have never seen
Ask an operator their average ticket time and most can give you a number. Ask them how often they beat or missed the estimate the guest was actually given, and most cannot answer. Not because they do not care. Because nobody could measure it.
That gap between the promise and the plate has been invisible. KitchenAI puts it on the screen. Once you can see it, 154 stops being a mystery and starts being a list.
The fix
154 is a to-do list, not a verdict
The number is not there to make anyone feel bad. It is there to give you a lever.
Now you can ask the questions that lead somewhere. Which dayparts drive the late orders? Which stations? Is the estimate too optimistic, or is the kitchen missing an estimate it could hit? You either fix the quote so you can keep it, or fix the flow so you can hit it. Either way the gap closes, and next month's red box gets smaller.
Run the same look at your own volume. Take your order count, find your on-time rate, and count the guests on the wrong side of it. That is not a stat on a slide. That is the number of people who left less happy than they walked in.
78.7% is a fine grade until you meet the 154 people it does not count.
