Specification document for a mobile app: the complete guide
A specification document that is too broad makes a mobile app budget drift. firstapp stopped asking for one and replaces it with a call, then a workshop on Figma.
Article summary
- The seven useful sections run from business context through to metrics, including the MVP, the tech stack, and monetization.
- Hard-paywall apps convert 5 times better than freemium at J35: the model is chosen before the design.
- 55% of trial cancellations happen on day 0, and onboarding is almost always missing from the document.
- An AI-generated text, 20 to 40 pages long, can price at €80,000 a project worth €20,000 without the invented features.
- firstapp replaces the specification document with a 30 to 45 minute call, then a Figma workshop of about 2 hours.
Summarize this article with
Everyone tells you to write a specification document before you contact an agency. It is the advice you read everywhere. It is also the advice that does the most damage on mobile app projects in 2026.
firstapp handles dozens of requests a month. For a long time we asked for specification documents. We stopped. This article explains what a good specification document should contain, and why, at firstapp, we do it differently.
The 7 sections of a specification document
1. Business context and goals
This is the section most founders rush through. It is also the most important.
Describe what problem your app solves, for whom, and why now. State your business goals at 6 and 12 months: number of active users, target MRR, target conversion rate. An agency that does not understand your business goals cannot make the right tradeoffs when it has to choose between two features.
2. Target user and use cases
Describe your target user in detail: age, mobile behavior, context of use. Then list the 3 to 5 main use cases in priority order. Not a feature list. Concrete scenarios. "A user opens the app for the first time and understands the value in under 2 minutes" is a use case. "Home page with a product list" is a feature.
3. Features by priority (MVP vs V2)
Use a simple two-column matrix:
MVP (launch) : the features without which the app has no value.V2 (post-launch) : the features that enrich the experience but do not block launch.
The rule: if a feature does not contribute directly to conversion or retention, it goes to V2. To understand what actually drives an app's revenue, read our article on mobile paywall strategy.
4. Technology choices and constraints
State your constraints: an imposed stack, an existing API, security requirements. If you have no view on the technology, say so. At firstapp, the default choice is React Native - a single codebase for iOS + Android, costs reduced by 20 to 40%, and a mature monetization ecosystem. To understand why this choice matters, read our article on React Native agencies.
5. Monetization strategy and paywall
This is the section most often missing. And the most critical.
According to RevenueCat's State of Subscription Apps 2026, apps with a hard paywall convert 5 times better than freemium apps at the median (10.7% vs 2.1% at J35). This choice has to be made before you design a single screen. Describe your model: monthly or annual subscription, freemium, trial. If you are still unsure, our article on freemium vs subscription vs one-time purchase will give you what you need to decide.
6. Third-party integrations
List the external services: RevenueCat or Adapty for subscriptions, analytics, push notifications, social authentication, APIs specific to your sector. Each integration means extra days of development. An integration forgotten in the specification document means a surprise invoice line mid-project.
7. Success criteria and target metrics
Define how you will measure success 3 months after launch: trial conversion rate, MRR, DAU, App Store rating, retention at J30. This section forces a healthy look at what actually matters and lets the agency align its choices with your real goals.
Newsletter
Once a week.
Last issue
We opened 10 apps to see how they sell a subscription.
+534 people read us each week
What a specification document often leaves out
Onboarding
Almost nobody describes their onboarding in the initial specification document. It is still the most important screen in your app. According to RevenueCat, 55% of trial cancellations happen on day 0, before the user has even used the app. Describe the journey you have in mind for a new user, even roughly. Our guide to mobile onboarding optimization gives you a basis to think it through.
Error states and edge cases
What happens if the user loses connection in the middle of a transaction? If their payment fails? These cases mean several days of development if they are not anticipated.
The ASO strategy
The app name, its store description, its keywords: these are decided before launch and they directly affect your organic visibility. Most founders think about them three days before submission. That is too late.
The post-launch plan
Who are your first 100 users? How do you acquire them? An agency cannot answer these questions for you, but it can align some development choices if it knows your launch plan.
Specification mistakes that inflate the budget
⚠️ These mistakes in a specification document are expensive:
Listing features with no priority. When everything is "essential," the agency cannot make tradeoffs. The result: a maximum quote that includes everything, or a project that drifts for lack of a clear frame.
Describing the "what" without the "why." "The app must have a gamification system" without explaining the goal does not let the agency challenge you or propose a more effective alternative.
Forgetting maintenance in the scope. Who handles iOS updates when Apple changes its guidelines? If it is not in the specification document, it will not be in the quote. Our article on the mobile app quote breaks down all of these cost items.
Confusing an MVP with "the full version minus a few features." A real MVP is the minimum version that validates your business hypothesis, not a complete app from which you removed three features.
The problem with AI-generated specification documents
⚠️ This pattern destroys budgets in 2026.
Since ChatGPT and its competitors became available to everyone, a new category of specification document has been showing up: documents of 20 to 40 pages, generated in a few minutes, that look ultra-complete and ultra-professional.
The problem? The AI does not know your budget.
It adds everything a specification document "should" contain according to its training data: an advanced notification system, an analytics dashboard, a gamification module, multi-language management, offline mode, social integrations... The founder receives the document, sees 35 well-structured pages, and thinks they did serious work. The agency receives that document and has to announce a budget of €80,000 for a project that is worth €20,000 without the 15 features the AI invented.
We have received AI-generated specification documents with entire sections on features the founder had never considered, did not want, and could not fund. But because the document looked serious, nobody dared to question the scope.
An AI-generated specification document is not a specification document. It is a wish list with no budget filter. And that is the first reason that, at firstapp, we stopped asking for one.
The call and the Figma workshop instead of the specification document
At firstapp, we stopped asking for specification documents. Out of conviction.
A founder who writes their specification document alone spends weeks formalizing hypotheses that have never been tested against technical and budget reality. And when the agency receives the document, it then spends hours taking the scope apart to propose something coherent.
Our approach: a first call, then a design workshop on Figma.
The first call lasts 30 to 45 minutes. We ask the questions that matter: what problem you solve, for whom, with what business model, what budget, what timeline. No form, no document to fill in. A direct conversation.
If the project lines up with our expertise, we move on to a design workshop on Figma. Together we lay down the visual and functional foundations of the app: the first screens, the onboarding journey, the paywall structure, the MVP features. In 2 hours, we produce more clarity than a 20-page specification document.
Why Figma rather than a document? Because a screen says more than a page of specs. When you see your onboarding drawn, you immediately understand what is missing, what is superfluous, and what actually matters. You make decisions on something concrete, not on hypotheses.
This workshop lets us propose a precise quote, line by line, with no budget surprise. So you can understand what that means in terms of overall cost.
If you want to compare how agencies approach this before you choose, our top 10 of the best mobile app development agencies gives you an overview of the French-speaking players, and the criteria to evaluate them.
A first call with no specification document
No specification document needed. Send us your project in a few lines, we set up a first 30-minute call, and if it lines up, we move on to a Figma design workshop.
FAQ on the mobile specification document
Do you absolutely need a specification document to contact a mobile app development agency?
No. A one-page brief is enough for a serious first contact: project context, target user, the monetization model you have in mind, and an indicative budget. At firstapp, we prefer a direct first call followed by a design workshop on Figma rather than a long document written in a vacuum. What matters is the quality of the thinking, not the length of the document.
Is a specification document generated by ChatGPT usable?
As a starting point to organize your ideas, yes. As a document sent to an agency, no. The AI generates features you did not ask for, and they will inflate your quote for no reason. Reread every line and ask yourself: "do I actually need this to launch?" Ruthlessly delete anything that does not answer that question directly.
How many pages should a mobile app specification document be?
Between 5 and 10 pages for the great majority of projects. Under 5 pages, the document is probably too vague. Beyond 15 pages, you have over-specified details that should be left to the agency's expertise - or you used AI without filtering. A short, precise document is always worth more than a long, approximate one.
Can you have the agency write your specification document?
Yes, and it is often the best approach. At firstapp, we replace the specification document with a design workshop on Figma: we start from the real needs, we draw the first screens together, we lay down the MVP foundations in a few hours. The result is more useful and more precise than a 20-page document written alone.
Should you include mockups in your mobile app specification document?
No, unless you worked with a professional designer. Non-professional mockups constrain the agency's thinking on onboarding and design. Describe the journeys in text and let the agency propose the visual approach. That is what you pay them for.