Fixed price vs hourly software development

The price is rarely the worry. What happens the week after you say yes is.

Fixed price or hourly is an argument about risk. Hourly puts the risk of the unknown on you; fixed puts it on us. Everything below is how that risk gets priced.

What frightens people is the weeks after they approve it, not the number itself. What gets decided without them. What turns up as an extra at the end.

Why scope for a week or two before quoting anything?

Because a price given before anyone has looked at the work is a guess. Guesses get corrected in one direction. So every service starts by looking. An audit week for automation. A week or two of discovery for custom software. Strategy and copy before a website is designed. An audit before any SEO work.

What that time buys is a list. Which systems have to talk to each other. Where the data actually lives. Who approves what. And the big one: which parts of the process nobody has ever written down. Unwritten process is where overruns come from.

It also buys you an exit. At the end of scoping you hold a document that describes the build. Take it elsewhere, or drop the idea. For custom software that discovery is free, so walking away costs you two weeks and nothing else. Our note on build vs buy for small teams argues for not building more often than you would expect.

What does "fixed scope" mean for you?

It means the thing being priced is a described build, not a bag of hours. The Services page says it in one line under Discovery: "Process mapping, data model, fixed scope. You know what you’re getting and what it costs."

Three things follow from that. They are the three worth checking with anyone who offers you a fixed price.

• The scope was written before the price was. If the number came first and the scope is still being described around it, that is not a fixed price. It is an opening position.
• The price covers a finished thing, not a stage of one. Migration, training and go-live are inside it. A system nobody has moved onto does not count as delivered.
• Overrun is the builder’s problem. If it takes longer than we thought, that is a risk we priced and got wrong. The number does not move.

A fixed price is only as fixed as the scope it points at. Read that document twice and the number once.

Fixed price vs hourly: when is hourly the right answer?

When nobody can yet describe what finished looks like. A fixed price on an undefined build is a bet one side has to lose. Usually it resolves as a thin build, defended line by line.

So we would say hourly, or a small time-boxed block, when:

• The work is exploratory. You want to know whether a model can read your documents accurately enough to be trusted. The answer might be no. A week that ends in a clear no is still a good week.
• The first job is reading somebody else’s system. Nobody can price a change to code they have not opened.
• The work has no end. Monitoring, maintenance, a monthly measure. There is no "done" to fix a price around.
• You want us on call for small changes. Fixing a price to work nobody has described is worse for you, not better.

The test works on any vendor, not just us. If the work can be written down, it should be fixed. If it cannot, anyone quoting it fixed is padding heavily or planning an argument with you later.

What changes the price?

Three things, in roughly this order. Integrations, data migration, and scope that arrives after the scoping. None of them are exotic. All three are routinely underestimated by the person buying.

Integrations are negotiations with other people’s software. Their API, their rate limits, their idea of what a customer record is. One integration is a task. Five is an architecture. The contractor CRM we are building plugs into email, calendar, video calls, WhatsApp Business and the bookkeeping package the business already runs on. That list is a real part of the build, not a footnote at the end of it.

Data migration is the least visible line in any quote and the likeliest to be wrong. Old data is never as clean as the person describing it believes. Duplicates, three spellings of the same customer, a field that quietly changed meaning and nobody wrote it down. Ask any vendor what happens to the records the import rejects. The answer tells you whether they have done this before.

The third is scope that arrives after the scoping. That one is less a question about the estimate than about what a supplier does on the day it turns up, so it has a section of its own further down.

What do you get to see during the build, and when?

Working software, every week, running on your own data. That is the Services page’s own promise for the MVP stage: weekly demos on real data. It is also what keeps a fixed price honest in your favour, not only in ours.

A demo on invented data proves the screens exist. A demo on your data proves the thing survives the mess you have. That hands you the only real leverage on a fixed-price build. You find out we built the wrong thing while there is still time to say so.

The rhythm carries past launch, and it differs by service because the work does. Custom software releases every two weeks. An automation is watched, with a monthly note on what it did. SEO and AEO get a monthly measure against the timeline the Services page states plainly: three to six months before results compound. A faster promise is usually a promise about ad spend.

What happens if the scope changes halfway through?

It gets written down, priced, and decided before anyone builds it. As its own small fixed thing, not as a quiet extension of the original. Absorbing it silently sounds generous and is not. It comes out of the finish of something else, or out of the relationship at the end.

The awkward version of that conversation is the honest one. "That is new, here is what it costs, now or in the next release?" A supplier who never once says that during a build is either not tracking scope or taking it out of somewhere you cannot see.

You can also change your mind about the whole thing. If discovery shows the build will not return more than it costs, we say so. We would rather do that than invoice you for finding out.

How do you get a real number for your own project?

Start from the floor. A website starts from €5,000, an SEO audit from €1,250, a first automation from €3,500 and custom software from €12,500, all excl. BTW. Each service shows its own on the homepage. The floor tells you whether a call is worth it. The fixed number for yours comes out of the scoping.

Describe the process, not the software. Thirty minutes is enough to say which services your problem touches, and roughly how many weeks it looks like. It is enough to say whether this is a fixed-price build or an hourly week of finding out. That is what the booking page is for.

Bring the boring things: the tools you pay for, the steps nobody has written down, and the part of the week everyone dreads.

If that is your week too, thirty minutes is enough to find out.

Book 30 minutes →