On this page
- Why our calculator shows a range
- One answer can move the price further than the whole range
- A wrong assumption we found late
- What discovery should deliver
- When a short call is enough
- What discovery costs and how long it takes
- What discovery does not do
- Where early decisions mattered in our projects
- Questions buyers ask us
- Next step
Short answer. Our calculator's ±15% covers the features you entered. It can't cover the decisions you haven't made yet, and one of those can move the price further than the whole range. In the example below, one yes-or-no answer (does your team need an admin panel?) moves the midpoint above the top of that range. Discovery turns open questions into decisions. For a small app with clear answers, a short call is enough. Otherwise we propose a discovery phase at a fixed price, and the estimate for the first release comes out of it.
Why our calculator shows a range
The app cost calculator sizes your app in engineering hours, then prices those hours at a blended planning rate of £45 an hour. That is a planning average, not one price: our rates are per engineer, about $30–90 an hour, depending on the engineer's level and experience. The hours come from three parts:
- a base for the type of app (450 hours for a simple app, 1,100 for a business app, 1,800 for e-commerce, 2,800 for enterprise);
- hours for each feature you tick (for example 25 for sign-in, 30 for push notifications, 120 for an admin panel);
- a backend, if the app needs one.
Then it multiplies. One platform is ×0.9. Two fully native apps are ×1.7. Custom design is ×1.1. A regulated sector (legal, finance, health) is ×1.2.
Finally, it shows ±15% around the result. We say it on the calculator page too: before discovery, an estimate is honest only as a range.
That ±15% covers one kind of doubt: how big each known feature turns out to be. It assumes the feature list is right. That assumption is the weak point.
One answer can move the price further than the whole range
Take a business app on Android only, with our standard custom design, sign-in, push notifications and analytics. The calculator gives 1,371 engineering hours, a midpoint of £61,695, shown as £52,500–£71,000.
Now answer one more question, one at a time.
| Open question | If the answer is "yes" | Hours | Calculator range | Midpoint | Against the first range (top £71,000) |
|---|---|---|---|---|---|
| Will users pay inside the app? | Payments, plus a custom backend | 1,579 | £60,500–£81,500 | £71,055 | just above the top |
| Will your team need an admin panel? | Admin panel, plus a custom backend | 1,638 | £62,500–£85,000 | £73,710 | above |
| Will users message you or each other? | Chat, plus a custom backend | 1,658 | £63,500–£86,000 | £74,610 | above |
| Is this a regulated service (legal, finance, health)? | Compliance work (×1.2) | 1,645 | £63,000–£85,000 | £74,025 | above |
| Must it work without a connection? | Offline mode | 1,460 | £56,000–£75,500 | £65,700 | inside |
| Do you also need an iPhone app, built natively? | Two native apps (×1.7) | 2,590 | £99,000–£134,000 | £116,550 | far above |
Only offline mode stays inside the first range. Payments clears its top by just £55.
Why do payments, chat and an admin panel jump so much? Each adds its own hours. Each also switches the backend from a simple REST API (200 hours) to a custom backend (350 hours). The calculator does this for payments, subscriptions, chat, an admin panel, AI features, live tracking and CRM integration, because all of them need real server-side work.
So the range is not the problem. The open questions are. Discovery exists to answer them.
Five of six single answers moved the midpoint past the top of the range. That is our own number, and you can check each row in the calculator by changing one answer.
A wrong assumption we found late
Some open questions never appear on a feature list at all. Our founder wrote about one of them on LinkedIn on 11 August 2026. Here it is as he posted it. The post does not name the client, so neither do we.
Eight years in, I can tell you exactly what a wrong assumption costs. Ten minutes in a document. Weeks once it is live. Here is one we found late. A law firm we build for has 36,000 clients inside its app. To book a consultation, a client opens a catalogue of practice areas and picks the service they need. That was the brief, and it was built exactly as asked. Their clients cannot tell those services apart. Pension, military, family, disability. From the outside it all reads the same. A person knows what happened to them. They do not know what it is called on a price list. So the first step of the journey is the one step they cannot take. Nobody finds that in a kickoff call about features. You find it by sitting with the firm and walking through what a client actually does in the hour before they call. That is the job of a business analyst, and it is the whole reason the role pays for itself. Caught in a document, that assumption costs one conversation. Caught in a live product, it costs the front door and the screens wired behind it. We are redesigning that first screen now. The real output of a good analyst is not a document. It is the list of things you no longer have to build. I cannot invoice anyone for the hours we did not spend. That is exactly why the role gets cut first.
Our calculator could not have caught this. "Can clients tell the services apart?" is not a feature, so it has no hours and no place in the ±15%. It shows up only when someone walks through what a client does before they call. That walk-through is the business-analyst work inside discovery. When the post went up, the team was redesigning that first screen.
What discovery should deliver
Discovery turns an idea into a plan you can price: the users, the flows, the scope, the priorities and the first release. Whoever runs yours, ask for these six things:
- A scope sheet. Every feature, with a decision: in the first release, later, or out. Each line carries its hours, the question still open and the risk if it stays open.
- An out-of-scope list. What the first release will not include, in writing. It is also the reference for any change request later.
- A flow map. The screens and journeys of the first release, so you can see the app before engineering time gets expensive. Ask whether a clickable prototype is part of discovery or of the design phase.
- A phased plan. Design, first release (MVP), then expansion. In our calculator's detailed breakdown, design is 12% of the build, the MVP 55% and expansion 33%. Discovery replaces those model shares with your real numbers.
- Yearly running costs. Hosting, store fees and maintenance. Our model uses 17.5% of the build cost a year for maintenance, before hosting and store fees.
- A proposal for the first release. Ask whether it is a fixed price or an estimate. With us, it is an estimate, and development is billed hourly.
Here is the scope-sheet format, with hours from our calculator model. It is a template, not a client's data.
| Feature | Decision | Model hours | Open question to close | Risk if it stays open |
|---|---|---|---|---|
| Sign-in and profiles | First release | 25 | Which sign-in methods? | Adding social sign-in later is another 20 hours in the model |
| Push notifications | First release | 30 | Which events trigger a push? | Pushes depend on events the backend doesn't record yet |
| Payments | First release / later / out | 60 + custom backend | Which payment provider, and who handles refunds? | Moves the backend from REST (200 h) to custom (350 h) |
| Admin panel | First release / later / out | 120 + custom backend | Who edits what, and how often? | The same backend jump; roles agreed late mean screens rebuilt |
| Offline mode | Later | 90 | Which screens must work with no signal? | Two devices change the same record while offline |
Out of the first release (example): a loyalty module (275 hours in the model), a referral scheme (115), CRM integration (150) and AI features (250). Each goes on the list with the reason it waits.
When a short call is enough
The ±15% covers the features you entered. If your app is small and the answers are already clear, a short discovery call is enough, and a written proposal follows it. That first discovery session is free. When the scope is still fuzzy after that call, our next step is a meeting with a business analyst, before any proposal.
What discovery costs and how long it takes
Our rates are per engineer, about $30–90 an hour, depending on the engineer's level and experience, and we work time and material, fixed price or outstaff. There is no minimum project size. The first discovery session is free. A full discovery phase has a fixed price that depends on the product, and how long it takes depends on the scope. Design is billed hourly, and so is development after discovery.
What discovery does not do
- It does not remove every risk. Some questions only get answered in the build, such as how a third-party API behaves under real load.
- It does not lock you in. You keep the discovery documents whether or not you build with us, and the contract lets you reuse them with another team. Rights pass to you as each invoice is paid.
- It does not assume you will build. Discovery has three honest endings: you build with us, you take the plan to another team, or you stop. If the answers show that an off-the-shelf tool does the job, or that the first release won't pay back, the plan says so. Then you have spent the discovery fee, not the build budget. As Oleksiy wrote in the post above, the real output is "the list of things you no longer have to build."
Where early decisions mattered in our projects
- Rethought. Apple's Screen Time permission is sensitive. The onboarding explains the access before the system asks for it. That is a decision about order and wording, made before code.
- Gamble Count. The first session had to feel personal without feeling heavy or clinical. A few light questions set up the plan.
- Travel Security Kit. The Android app moved from product flow to a completed foundation within a 6-month delivery.
- Merezha Prava. A Ukrainian legal-services company. Thousands of its clients needed to see where their case stood without phoning the office. Automated case-stage notifications went into the app, and inbound calls to the office fell by 15%.
All four had a discovery phase before the build, notably Travel Security Kit and Rethought.
Questions buyers ask us
Yes. The discovery documents are yours whether or not you continue with us, and the contract lets you reuse them with another team. Rights pass to you as each invoice is paid, so they are yours once the discovery invoice is paid.
The scope agreed in discovery is the reference: anything outside it is a change. We size it the same way as the original scope, in engineering hours from the same model, at the rates in your contract, and send it to you in writing. Work on it starts only after you approve it.
Next Era Solutions is the trading name of AppVanta OÜ, registered in Estonia (reg. 17188835), so you contract with an EU company. Our engineers work remotely across Europe, from the UK to Portugal, the Nordics and DACH, and a few work in Ukraine. Rights in the work pass to you as each invoice is paid. How personal data is handled and transferred is agreed in writing for each project; ask us about yours.
US dollars or euro, and pounds are possible. Payment terms are 14 or 30 days, depending on the client. The calculator shows GBP, EUR and USD.
Usually within one working week, the same as a build.
Next step
Book a 30-min call and bring your calculator result. We'll tell you whether a call is enough or discovery will pay for itself.
Or run the app cost calculator first. See how we run new app and MVP projects, or browse all our services.