The order lands 25 minutes late. The customer doesn't complain, doesn't ask for a refund, doesn't leave a review. They just never order again, and your dashboard files it as a successful delivery. That's the expensive part of slow delivery. The cost shows up weeks later, on a different line of your P&L, where nobody thinks to connect it back to a Tuesday-night courier shortage.
The reorder curve drops with every 10 minutes
Track cohorts of customers by how long their last delivery took, then count who orders again within 30 days. The same staircase shows up across the operators we work with and in published retention studies. Customers whose food arrived in under 30 minutes reorder at roughly 40–50%. In the 30–45 minute band it's about 30–40%. At 45–60 minutes, 20–30%. Past the hour you're down to 10–20%, and most of those survivors are regulars spending loyalty you built earlier.
The rule of thumb that falls out of the staircase: every 10 extra minutes shaves about a quarter off the odds of a next order. A regular is worth hundreds of dollars a year (we ran that math in why a $20 order is worth $400), so a chronically slow zone becomes one of the most expensive problems you have.
Customers punish broken promises, not long waits
The absolute number matters less than the gap between what you promised and what you delivered. Tell a customer 50 minutes and get there in 48, and they're happy. Tell them 30 and make them wait 45, and they're gone, even though the second wait was shorter. Two things follow. First, don't let marketing write the delivery promise; let the data write it. Second, quote by zone and by hour. A Friday 7pm promise should not read the same as a Tuesday 3pm one.
We'll show you the monitoring setup that catches slow orders while they're still fixable.
Where the minutes actually hide
Owners assume slow delivery means slow drivers. Pull the timestamps and the picture is usually different. The order sits unconfirmed for four minutes because the tablet was buried under receipts. The kitchen finishes at minute 22, but nobody assigned a courier until minute 20, so the bag sits on the pass while the food cools. The courier takes two other orders first because dispatch batched by convenience instead of geography. The address sits at the far edge of a delivery zone that someone drew as a circle on day one and never revisited.
None of that shows up if you only track door-to-door time. Break every order into four segments (accept, cook, wait-for-courier, ride) and the fix usually costs nothing: auto-assignment instead of a dispatcher's gut, zones redrawn as polygons around real streets, kitchen start triggered by courier ETA. That's the job of logistics management, and it routinely finds 10–15 minutes without anyone driving faster.
One slow week damages three channels at once
The reorder curve is only the first bill. Slow orders are also where bad reviews come from. Rating complaints cluster around "took forever" and "arrived cold," which is really one complaint wearing two hats. A drop from 4.6 to 4.3 stars suppresses conversion for every future visitor, including the ones you'd have delivered to on time. Then there's the operational tax. Late orders generate "where is my food?" calls, and every call ties up someone who could be taking orders. Operators tell us that on a bad Friday, one late-running zone can eat more staff attention than the rest of the shift put together. Speed problems never stay in logistics. They bill marketing and support too.
The effect compounds in the other direction as well. Fast, predictable delivery is one of the few edges a local brand has over the marketplaces: your couriers serve one kitchen, theirs serve forty. A brand that consistently beats its own promise earns the kind of repeat behavior no promotion can buy, and it shows up in the repeat-rate line within two to three months.
Watch the 90th percentile, not the average
Averages lie about delivery. A 35-minute average can hide the 15% of orders that ran over an hour, and those are the orders doing the retention damage. The numbers worth checking weekly: promised-vs-actual gap, p90 delivery time, and the share of orders past your promise, each split by zone and daypart. A live monitoring panel catches today's late orders while an apology and a bounce-back coupon can still change the outcome; weekly reports catch the structural offenders, the one zone or shift behind most of your late deliveries. In our experience the misses are rarely spread evenly. A handful of zone-hour combinations usually account for the bulk of them.
Do this next
Export last month's orders with timestamps. Compute the reorder rate for customers whose delivery beat the promise against those who waited 15+ minutes past it. That single comparison, built on your own data, will tell you whether your growth problem is really a logistics problem. For most operators past 30 orders a day, it is.
Auto-dispatch, smart zones, and live monitoring — one platform, launched in about two weeks.