This restaurant online ordering case study is about a mid-size F&B group in the Tangerang area, running six outlets of a casual dining brand. I have anonymized everything, but the story is real, and the lesson is one I keep watching businesses learn the hard way.
The group came to the same realization many restaurant owners reach around their third profitable year: the delivery marketplaces that brought them growth were also quietly eating their margins. On every order that came through a major food-delivery app, roughly 20 percent of the ticket disappeared in commission. On a business already running thin F&B margins, that cut was the difference between a healthy month and a nervous one.
So they decided to build their own online ordering channel. What follows is what actually happened, including the two moments the project nearly died, because the tech was the easy part. The war was everywhere else.
The Margin Math That Started It
Let me make the numbers concrete, because they are what justified the whole project.
The group was doing about Rp 900 million a month in revenue, with roughly 35 percent of that, around Rp 315 million, flowing through delivery apps. At a 20 percent commission, they were paying about Rp 63 million every month to the marketplaces.
Building their own ordering site and app was quoted at a one-time build plus modest monthly hosting and payment fees. Even assuming they could only pull a third of their marketplace orders onto their own channel in the first year, the commission they would save paid back the build in well under a year. After that, every order kept in-house was margin they got to keep.
The math was obviously in favor of building. Which is exactly why owners underestimate these projects. A clear ROI on paper hides how hard the non-technical parts are.
What Almost Killed It the First Time: Customers Would Not Switch
The system launched clean. Ordering worked, payments worked, the kitchen received tickets. And for the first six weeks, almost nobody used it.
This was the first near-death moment. The owners had assumed that if they built a good channel, customers would naturally move to it to be nice, or because it was theirs. Customers do not care whose channel it is. They cared about habit and incentive, and both pointed at the app they already had on their phone.
What turned it around was not a better website. It was reshaping the incentive:
- A price and promo advantage on the direct channel. Items were priced a little lower on their own app than on the marketplace, funded directly by the commission they were saving. Ordering direct became the smart choice, not the loyal one.
- A reason printed on every bag. Every marketplace order that went out carried a small card: order direct next time, here is the discount code. They used the marketplace's own customers to migrate people off the marketplace.
- Loyalty that only lived on the direct channel. A simple points system that the app could offer and the marketplace could not.
Slowly, the direct channel share climbed. The lesson: building the channel is 30 percent of the work. Giving customers a real reason to change a comfortable habit is the other 70 percent. This is the same dynamic I described in Traditional Retailers Are Losing Quietly, Not Suddenly, where convenience is a habit, not a preference.
What Almost Killed It the Second Time: Staff Would Not Trust the Screen
The second near-death moment came from inside the kitchen.
The new system fed orders to a screen at the pass. But the staff did not trust it. During busy service, they kept one eye on the new screen and, out of nerves, also had someone manually re-checking orders on paper, which slowed everything down and created exactly the chaos the system was supposed to remove. When a single order was missed early on, the whole team's confidence dropped, and some quietly went back to taking direct orders by phone and WhatsApp, defeating the point.
Fixing this was not a software fix. It was an operations fix:
- Run both systems in parallel, briefly and deliberately. For two weeks the new screen ran alongside the old method, so staff could see it never missed. Confidence is earned by evidence, not by a memo.
- Make one person the owner per shift. A single staff member per shift was responsible for the order screen, so accountability was clear and nobody assumed someone else was watching.
- Design the screen for a stressful kitchen, not a calm office. Large text, loud new-order sound, an obvious unmissable state for a fresh ticket. The first version was too subtle for the noise and speed of real service.
Once the staff trusted the screen, the manual double-checking stopped, and the channel finally ran the way the math had promised.
What the Numbers Looked Like After a Year
By the end of the first year, the direct channel carried a meaningful share of what used to be marketplace volume. The commission saved on those migrated orders more than covered the build and the ongoing costs, and every future order on the direct channel was margin retained rather than shared.
Just as valuable, and harder to price, they now owned the customer relationship. They knew who ordered what and how often, could message customers directly, and were no longer fully dependent on a marketplace that could change its commission terms whenever it liked.
The Takeaway
This restaurant online ordering case study is really about where the difficulty lives. The margin math for leaving a 20 percent marketplace commission is compelling, and the technology to build your own channel is routine. The project nearly failed twice, and neither time was technical. Customers would not switch until the incentive made switching the obvious choice, and staff would not trust the new screen until evidence, ownership, and better design earned it. If you are weighing your own direct-ordering channel, budget most of your energy for changing customer habits and staff behavior, because that, not the software, is what decides whether the ROI on paper ever reaches your bank account.