A direct channel does not announce when it starts paying for itself. There is no statement telling you the fee is now cheaper than the commission it replaced. You have to look, and three numbers are enough.
📊 The three numbers
| Number | How to get it | What it tells you |
| Cost per direct order | Monthly system cost ÷ direct orders that month | Compare it against your commission per order. This is the whole business case |
| Conversion rate | Orders ÷ visits to the ordering page | Whether the page works. Low means friction, not lack of interest |
| Repeat rate | Customers with more than one direct order ÷ all direct customers | Whether you are building a habit or paying for one-offs |
Track cost per order monthly. It starts embarrassing — the first month it might be $40 an order — and falls fast. Watching it fall is what keeps you going through the months where the channel looks like a waste of money.
🔍 Reading a low conversion rate
If people reach the ordering page and do not order, the food is not the problem — they already chose you. Something in the path is. In order of how often it is the culprit:
- Forced account creation. The single biggest killer. Guest checkout, always.
- The wait time revealed at checkout. Show it early or lose them late.
- Too many taps, or a required modifier on a simple item.
- Delivery fee or minimum appearing at the end. Same problem as the wait time.
- The page is slow on a phone, especially with big menu photos.
Diagnose it the direct way: order from your own site once a week, on your phone, and notice where you sigh.
🗓️ A rhythm that survives a busy month
- Weekly, five minutes: direct orders this week versus last. Any failed payments or errors reported.
- Monthly, twenty minutes: cost per direct order, average ticket direct versus marketplace, and the share of total orders that are now direct. That share is your real scoreboard.
- Quarterly: repeat rate, and whether the offer from topic 5 is still pulling.
The share-of-orders number is the one to put somewhere visible. Moving from 5% direct to 20% direct changes what your restaurant earns without selling a single extra plate.
🔧 What to do with what you find
- Conversion low? Fix the path before spending anything on promotion. Sending more people to a page that does not convert just costs more.
- Direct ticket smaller than marketplace? Your add-ons are not working. Back to topic 3.
- Repeat rate low? People tried it once. Something about the experience did not beat the app — usually the wait, or having to re-enter a card.
- All three fine but volume flat? This is a topic 5 problem, not a topic 6 one. Nobody knows the channel exists.
The direct channel is slow in a way that marketplaces are not: no burst of orders on day one, just a line that bends upward over months. Judge it on the trend across a quarter, not on any single week. And keep the marketplace — it is doing a different job, and now you know exactly what that job costs.