Platform dashboards are built to show you growth, because growth is what keeps you on the platform. They are not built to show you profit. This chapter is about the handful of numbers that tell you whether delivery is actually working, and what to do when the answer is no.
📈 The five numbers that matter
Ignore the rest of the dashboard until these five are in front of you:
| Number | What it answers | Where it hides |
| Net per order | What actually reaches my account after every deduction | Nowhere. You calculate it: statement payout ÷ orders |
| Contribution per order | Net minus food cost minus packaging. What is left to pay rent | Your numbers, not theirs |
| Average ticket | Is this channel worth the kitchen disruption? | Dashboard, and it is usually honest |
| Repeat rate | Am I acquiring customers or renting them? | Dashboard, if the platform shows it. The number that decides whether promotions were an investment |
| Rating and cancellations | Am I about to lose visibility? | Dashboard. Watch the trend, not the number |
The first two are not on any dashboard, and they are the two that decide whether you should be doing this at all.
🗓️ A rhythm you can actually keep
The reason most restaurants do not track this is that the tracking plan was too ambitious to survive a busy month. Keep it to this:
- Weekly, five minutes: rating trend, cancellations, and any review that repeats a complaint. These are the things that get worse quietly.
- Monthly, thirty minutes: open the itemised statement, calculate net per order and contribution per order, and compare against last month. Check what promotions ran and what they cost.
- Quarterly, one hour: is each platform still earning its place? Which dishes sell on delivery and which never do? Is the repeat rate moving?
🔧 What to change when the numbers are bad
In rough order of how quickly they pay back:
- Cut the promotions that discount regulars. Immediate, costs nothing, and often fixes the whole problem.
- Remove the dishes that lose money or travel badly. Fewer refunds and better ratings, same week.
- Reprice for the channel, within your contract. Small and consistent beats one large jump.
- Push regulars to direct ordering. Slow, permanent, the highest return of anything here.
- Renegotiate, or drop a platform. Volume from one platform is a negotiating position with another. If a platform never reaches positive contribution after the four steps above, it is not a channel — it is a subscription you are paying for the privilege of cooking.
🚪 Deciding to leave
Leaving is a legitimate outcome, and it is easier if you prepare for it: know your notice period, and start moving regulars to direct ordering months before you need to. A restaurant that can leave a platform negotiates very differently from one that cannot.
Delivery done deliberately is a real second dining room with no rent. Delivery done by default is a kitchen working at capacity for someone else's margin. The difference is the numbers in this chapter, checked monthly.