· 6 min read
Why we talked a client out of a Telegram Mini App
The client wanted a polished catalogue inside Telegram. We ran the numbers and proposed plain buttons. A year later we revisited the conversation.
A café chain came to us asking for a Telegram Mini App: a catalogue with food photography, a cart, custom design — essentially a delivery app, but inside Telegram. The budget under discussion was around $11,000.
We built a button-based bot for $4,200. Here is how we got there, because the reasoning applies well beyond food delivery.
What we calculated
A 60-item menu across seven categories. Average order around $10. The core scenario: open the bot, pick three or four items, pay. That is it.
A Mini App is effectively a web application inside Telegram. That means markup, your own navigation, loading states, network error handling, responsive behaviour and separate testing across three Telegram clients. It is full front-end development, not bot configuration.
The estimates came out like this:
- Button-based bot: 6 weeks, $4,200
- Mini App with the same functionality: 14 weeks, $10,900
- Difference: $6,700 and two months
What the premium actually buys
You are paying for visual freedom. A Mini App lets you show large photography, build a custom grid, add animation and brand colours. For a clothing or furniture catalogue, where the decision is made by eye, that pays for itself.
For a wrap, it does not. Someone ordering lunch for the third time this week is not browsing photographs. They are looking for what they had last time.
The number that decided it
We pulled a year of order history: 64 % of orders repeated the previous one exactly or with a single substitution. That reframes the problem entirely. The core requirement is not "display a catalogue" but "let people repeat an order as fast as possible".
In the button-based bot we made "Repeat last order" the first button on the first screen. Three taps to payment. In a Mini App the same path takes at least five, because the interface has to load first.
A year after launch, the repeat button accounted for 38 % of all orders — more than any menu category.
When a Mini App is the right call
We are not against Mini Apps. We are against reaching for one by default. It earns its cost when at least one of these holds:
- 01The choice is visual: clothing, furniture, jewellery, travel
- 02More than 150–200 items, so browsing without search and filters is hopeless
- 03You need complex inputs: a map, seat selection, an availability calendar
- 04Brand presentation matters because the brand is part of the product
If none of those apply, buttons are cheaper, faster, and work on any connection.
How it ended
A year later we raised the Mini App question ourselves: the bot had stuck and orders had passed a thousand a month, so we offered to build one. The client declined — said the buttons were fine and the money was better spent on a second location.
That is arguably the best outcome a technical decision can have: people stopped thinking about it, because it works.