A late order costs a review, a repeat and the customer. Multiply across a year and it's the difference between a business and a hobby. Here's how to measure OTD properly — and how to move the number.
"We cut average delivery time by 11 minutes. Repeat rate went from 21% to 31% in a quarter. Nothing else changed."
Food quality is what you think customers judge you on. Delivery time is what they actually judge you on. A great meal 20 minutes late scores worse than a decent meal 10 minutes early. This isn't unfair — it's rational. A cold pizza isn't the pizza you paid for.
OTD — On Time Delivery — is the single strongest predictor of repeat ordering we see in the data. Higher than cuisine, higher than price, higher than review score. If you only have capacity to move one number this quarter, move this one.
Order placed → food ready. This is your kitchen's number. If it's over target, no driver in the world can save you — the order is already late before it leaves the pass.
Food ready → driver on the road. The silent killer. A bag sat on the heated rack for seven minutes is a bag that arrives lukewarm. Track it. Watch it fall.
Order placed → customer's hands. The one the customer cares about. Made of Make + Rack + Drive. If you want OTD to move, move the first two first.
1. Promise later, deliver sooner. If your current OTD is 42 minutes, stop quoting 30. An under-promise over-deliver pattern lifts review scores faster than anything else. The customer didn't want a fast delivery — they wanted the delivery they were promised.
2. Batch smarter, not harder. Doubling up orders is efficient only when both drops are on the same route. A "quick" second drop that adds seven minutes to the first is a disaster for the first customer. The platform should batch; the driver shouldn't.
3. Fire the courier sooner. Too many kitchens wait until the food is boxed before calling for a rider. A rider that's already three minutes away is a rider arriving at ready-time. Auto-dispatch, not manual-dispatch.
4. Coach to the outlier, not the average. A 34-minute average hides 12 orders that took over an hour. Those 12 are your bad reviews for the week. Find the pattern — same driver, same hour, same postcode — and fix it there.
Cut your average OTD by 10 minutes and you'll typically see repeat rate rise by 5–8 percentage points over a quarter. That's a growth number no ad spend can match — and it costs nothing beyond discipline.
It also works the other way round. OTD that quietly drifts from 32 minutes to 38 over a year will cost you repeat customers you never realised you were losing. The damage compounds silently. So does the benefit.
If you're on Andromeda, OTD lives on the pass screen, on the manager map and in the weekly summary. If you're not — the first step is measuring it properly. Averages lie; distributions tell the truth.
Book a 20-minute demo. We'll show Make, Rack and Drive Time broken out by order, by driver and by hour — on real kitchens.
Book a demo