Every estimate argument starts the same way: "it's just a simple app". Then three weeks later the same app has a payment provider, an admin panel, two integrations and a data migration.
Estimation is not guessing. It is a process with a few known methods. Here are the four that actually get used, and what each is good for.
By analogy
"We built something similar and it came to this much." Fast, useful for a rough range, and wrong the moment the scope differs. Good for a first conversation, dangerous as a contract.
By stages
Each stage is estimated separately: prototype, core, integrations, release, support. More work upfront, but the client can see where the money goes and can stop at any stage. This is the format I use most, because it turns an argument about the total into a discussion about one stage at a time.
Expert estimate
An experienced contractor looks at the list of work and gives a range. Fast and usually decent for scoping, but it is a range, not a commitment. Treat it as a filter, not a quote.
Time and materials
Pay for hours. Flexible, good for exploratory work where nobody knows the shape yet. Requires control: without a written scope, T&M quietly becomes an open-ended retainer.
In practice the best result is a combination: a range by analogy, a breakdown by stages, and a fixed price per stage.
What raises the price
- Complex integrations (ERP, payments, CRM);
- Offline sync and conflict resolution — the part everybody forgets;
- Non-standard design and animation;
- Compressed deadlines;
- "One more button" loops with no acceptance criteria.
What lowers it
- Ready-made components and templates;
- A minimal, honest MVP scope;
- Content and copy supplied by the client;
- A realistic timeline with stages;
- A clear spec with acceptance criteria.
The cheapest project
Not the one with the lowest bid. It is the one where the scope was clear from day one: nobody paid for features that were cut, and nobody argued about what "done" means.
If you only take one thing from this: ask for the estimate broken down by stage, and ask what is not included. The second question is where the real number hides.
Cross-post. Full version: How to estimate a development project. Related: how to write a brief and MVP in two weeks.
Top comments (0)