Pull up the last payout statement from your delivery app. Find the line marked commission, or service fee, or whatever they renamed it this quarter. That number is rent, and you pay it on every single order.
Worse, the customer on the other end has no clue you exist as anything but a logo in a list. To them the app is the restaurant, the booking, the shop. You cooked the food and packed the bag, and the app keeps the relationship.
Good news is buried in that same statement. A solid chunk of those buyers would order straight from you if you gave them one good reason and made it stupid easy. Moving even a handful over each week adds up, and the tools that do it can run quietly in the background.
Run the numbers before you change anything
Start with what the apps actually cost you. Commissions usually land somewhere between 15% and 30% of the order, depending on your plan and how much promotion you buy on top. Call it a quarter of the ticket, give or take.
Take a $40 order. At 25%, the app keeps around $10 before you count any extra marketing or delivery add-ons. Run that same order through your own online ordering page and you're mostly paying card processing, figure roughly 3%, so a little over a dollar.
That's about $9 back in your pocket on one $40 order. Move 20 orders a week to direct and you're looking at something like $180 a week, or close to $9,000 a year. Those are rough numbers, and your real figure depends on your ticket size and your plan, so spend ten honest minutes with an actual statement before you trust any of it.
- Pull one week of payouts and add up the commission and fees.
- Divide by the number of orders to get your cost per order.
- Subtract what a direct order would cost you (mostly card processing).
- That gap is what each moved customer is worth. Keep the figure handy.
Get the contact info the app won't hand over
The apps guard the customer list like a vault, because that list is the whole business. You will not pry it loose, so stop trying. What you can do is capture the relationship yourself the moment an order leaves your counter.
Drop a small card in every bag. Keep it plain: a QR code to your direct ordering page and a line telling people to text a word to a number to join for a perk. Something like "Text PARK to this number for $5 off your next direct order."
- A QR code straight to your own ordering page, not a homepage they have to dig through.
- One clear perk for ordering direct, stated in dollars.
- A text-to-join word so signing up takes five seconds on the subway home.
One honest note: most people toss the insert. That's fine. You're not converting everyone, you're converting the regulars who already like you, and those are the ones worth keeping off the app.
Loyalty texts people actually keep
Text messages get read, nearly all of them, usually within a few minutes. That reach is exactly why you can't abuse it.
Rules come first. You need real permission before you text anyone marketing, meaning they opted in on purpose (that text-to-join signup counts), and every message needs an easy way to stop. Break that and you're not just annoying, you're exposed to real fines.
Now the useful part. An automated system can send the right message at the right moment without you ever touching your phone:
- A welcome text with the promised perk, seconds after someone joins.
- A nudge a couple weeks after a first direct order, with one small reason to come back.
- A quiet win-back after 30 or 45 days of silence.
- The occasional heads-up: a new menu item, holiday hours, a slow-Tuesday deal.
Keep it rare and keep it worth reading. One or two texts a month that save someone money will always beat weekly noise that gets you muted or blocked.
Nudge the direct order without being a pest
Once someone has ordered direct a single time, the whole job becomes making the second order the obvious choice. Automated follow-up does the remembering so you don't have to.
Set up a few simple triggers and let them run:
- A reorder reminder timed to how often that customer usually buys.
- A points or punch-card style reward that only tracks on your direct channel.
- A birthday or anniversary note with a small freebie attached.
The trick is that the good stuff lives only on your own page, never on the app. If the loyalty perk, the birthday freebie, and the fastest reorder all live in one place, people quietly learn the route and take it.
Where this falls short
Automation moves the customers you already have. It does not put you in front of strangers, and that's the one thing the apps genuinely do well. A hungry person three blocks away scrolling at 8 PM finds you there, so keep a foot in that world for discovery and treat direct ordering as where you hold on to the people you already earned.
Two more limits worth saying out loud. If your direct ordering page is slow or clunky, none of this holds; people bounce back to the app out of pure habit. And if you discount so hard to win the direct order that you lose money on it, you've just moved the problem instead of fixing it, so run the perk through the same math you ran on the commission.
Worth doing this week
You don't need all of it at once. Start with the pieces that pay for themselves fastest.
- Pull one payout statement and work out your real commission per order.
- Order a stack of bag inserts with a QR code and a text-to-join word.
- Set up the keyword and a welcome text that delivers the perk automatically.
- Write three follow-up texts (welcome, two-week nudge, win-back) and put them on autopilot.
- Pick one reward that exists only on your direct channel, and price it so you still make money.
If wiring up the texts and the follow-ups sounds like one more job you don't have time for, that's the kind of setup we build for NYC shops so it runs on its own. Either way, the first move is free: read the statement and see what the apps are really charging you.