Most delivery launches die one of two ways. Some drag on for months while the owner keeps researching software and repainting the logo. Others go live over a weekend with no zones, no delivery-fee logic, and no idea whether a $25 order 4 miles away makes money or loses it. Both are failures of sequencing. What follows is the order that actually works, the one behind the roughly two-week launches we run with restaurants on our platform.
Days 1–3: decide the economics before touching software
Start by pricing a single delivery on paper. Take your average ticket, subtract food cost, packaging, and what it truly costs to get the order 2 miles away: courier wage or per-drop fee, plus a share of the idle time between runs. If that order loses money at your average ticket, delivery will only scale the loss. Set a minimum order value, a delivery fee, or both before you go any further. There's a worksheet for all of this in the unit economics article.
Then draw the zone you can actually serve, which is usually smaller than the one you'd like to. Begin with whatever you can reach in 15–20 minutes of real driving. A tight zone that delivers hot food beats a wide one that delivers cold food and collects one-star reviews, and you can always widen it later once you have data to widen it on. The zone-design details are in our delivery zones guide.
Last, cut the menu, because not everything travels. Fries that show up soggy and delicate plating that shows up shaken will cost you the repeat order. Launch with the 60–70% of the menu that survives 20 minutes in a bag, and keep a list of what you dropped; some of it earns its way back once you fix the packaging.
Days 4–8: put the ordering channel on your own domain
Customers need one place to order, pay, and track their food, and you want that place to be yours from day one. Restaurants that launch on marketplaces alone spend their whole delivery life handing over an effective 30–40% per order, and they never actually meet their own customers. Lead with your own channel and the arithmetic flips: the marketplaces become an optional way to get discovered rather than the ground you stand on.
In practice this window covers the menu upload with photos and modifiers, payment setup, delivery fees and minimums per zone, and opening hours. On our platform all of that runs alongside everything else, since an ordering website and a branded app are built from the same menu and go live together. What makes the app worth having even on a brand-new service is the conversion gap: apps convert up to 35% of visitors, sites up to 15%, and marketplace listings around 3%.
This is also when you settle how orders reach the kitchen. Printing a ticket off an email doesn't count as a system. Orders should land on a single screen with statuses (new, cooking, ready, out for delivery) so nobody is retyping anything when the rush hits. That's order management, and setting it up takes an afternoon rather than a week.
Site, app, zones, courier tracking, and kitchen screens — one platform, one launch plan. We'll show you the timeline on your menu.
Days 9–11: couriers — start smaller than you think
The classic mistake is hiring three couriers for a launch that generates eight orders a day. Start with one courier per shift, or go hybrid: your own driver on the dense core zone, a third-party fleet for the edges and the overflow. The break-even math between owning and renting couriers deserves its own article, but at launch volumes flexibility beats any theoretical cost per drop.
What matters more than headcount is visibility. From the first order you want to see where every courier is, how long each drop took, and which orders sat waiting on the counter. With logistics management running, assignment happens on its own, the customer gets live tracking, and you collect the delivery-time numbers you'll lean on when you widen the zone. Bolt this on after launch and you end up relearning your own operation from scratch.
Days 12–14: run test orders like a hostile customer
Before you open to the public, place real orders yourself: pay, wait, receive. Order the awkward stuff, the soup, the dessert, the combo with three modifiers. Order out to the far edge of the zone during the rush. You're checking four things: whether the food arrives in a state you'd happily pay for, whether the promised time matches the real one, whether the kitchen screen holds up with three orders going at once, and whether the chain from payment to payout works all the way through.
Fix whatever breaks, then start small: regulars first, a note to your existing customers, inserts at the counter. Hold the marketing push until week three, once the operation has survived a real Friday.
The first month: watch three numbers
Watch delivery time, repeat rate, and margin per order. Delivery time predicts retention: operators consistently find that customers who get late orders just stop reordering, usually without a single complaint. Repeat rate shows whether the product survives the bag, and margin per order shows whether you set the zone and fees right. All three should sit in your reports without any spreadsheet work. If one of them drifts for two straight weeks, do something about it: shrink the zone, adjust the fee, or repackage the item.
Do this next
Block out three days this week for the economics work: the paper math, the zone, the menu cut. It needs no software, no budget, and nothing you can't undo, and it's the part that decides whether the other eleven days build a business or an expensive hobby.
3M+ orders have run through launches like this one. Bring your menu, leave with a launch date.