What custom software actually costs, and what moves the number

Nobody publishes real prices for custom software, so most non-technical founders shop blind. You describe the same idea to five shops and get quotes that vary by five times. It feels like some of them are lying. Usually none of them are. They are quoting five different projects, because your description left room for five different readings.

Here is what actually moves the number, roughly in order of impact.

1. How much of the scope is genuinely decided

The single biggest driver is not features. It is ambiguity.

A shop pricing a vague brief has two options: pad the estimate to cover the unknowns, or quote the smallest honest reading and handle the rest with change orders. Both are rational. Neither is what you wanted.

The fix is not a longer requirements document. It is deciding the handful of things that change the architecture: who logs in and what they can see, whether data is shared across customers or isolated, what has to keep working when an integration is down, and what has to be auditable later. Five decisions like that will move a quote more than fifty feature bullets.

What to do: before you collect quotes, write one page answering those five questions. Send the same page to everyone. Now you are comparing prices instead of comparing interpretations.

2. Whether it has to survive other people

An internal tool used by six people who all know each other is a fundamentally different build from a product strangers pay for. Multi-tenancy, roles, billing state that survives a failed card, password resets, audit logs, and the ability for someone else to deploy it are not polish. They are structure. They have to be decided in the first sprint, because retrofitting tenancy into a single-tenant codebase is close to a rewrite.

This is the line that most often separates a $15,000 quote from a $60,000 quote for what sounds like the same app.

3. Integrations, counted honestly

Every system you already run that the new thing has to talk to is a line item. Not because the API call is hard, but because someone has to handle the case where the other system is down, rate limits you, changes its schema, or returns data that does not match what its documentation claims.

A useful rule when you are budgeting: each real integration is a meaningful fraction of a simple app's total cost, and the second and third are not cheaper than the first unless they share a pattern.

What to do: list the systems by name. CRM, billing, phone, scheduling, accounting. Ask each shop to price them individually. The ones that quote a single blended number for "integrations" have not thought about it yet.

4. Who is actually doing the work

Rates vary enormously, and the headline rate tells you less than you would think. What matters is the ratio of senior time to junior time, and whether anyone senior is accountable for the outcome rather than just available for questions.

A cheap hourly rate with no senior oversight produces code that works in the demo and becomes expensive in month six. A high rate with a senior operator who never touches your project is worse. The thing to price is accountability, not headcount.

We wrote a separate piece on how to find out who is actually writing your code, because the answer is rarely on the website: what you are actually buying when a dev shop quotes you.

5. What happens after launch

Most quotes cover building the thing and stop there. Then the invoice arrives for the parts nobody scoped: hosting, monitoring, the first round of real-user bugs, the security patch, the dependency that went end of life.

Ask for the twelve-month number, not the build number. A build quote that looks 20 percent cheaper and carries an unpriced maintenance relationship is not cheaper.

The ranges we publish

We think withholding prices is a tactic, so here are ours as reference points rather than as a claim about the market:

• Stabilization Review, ten business days, written 30/60/90 plan: $2,500 flat. This is the diagnostic step for a project someone else started.
• Local business website, one page, live in about a week: $999. Multi-page builds with AI included start at $2,500.
• Industry AI builds such as voice agents and operating dashboards: from $2,500.
• Custom software builds: scoped per engagement. Our estimator will give you a range from a handful of questions without a sales call.

Use those as calibration. If another shop's number is far off in either direction, that is worth a conversation rather than an assumption.

The question that gets you a real quote

Ask every shop the same thing: what would you cut first if my budget were half this, and what would break?

A shop that has thought about your project answers immediately and specifically. A shop that has not will talk about flexibility and partnership. The answer tells you whether they understand the build well enough to have opinions about its structure, which is the same thing as whether their number means anything.

---

If you want a second read on a quote you already have, that is a reasonable thing to ask for and it does not need to be a sales call. Send it over and we will tell you what we would want clarified before we would sign it.