Software is mostly hours of work, so what an app costs follows from how much work goes into it. Two apps with the same description can be far apart on that.
Say a bike shop wants customers to book a repair online. In one version that is a web app in the browser: the customer picks a time and the shop sees the booking in a list. In the other version it is an app for iPhone and Android, with payment up front, a loyalty card, a notification when the bike is ready and a connection to the till. Both are called “an app for the bike shop”, and the second takes far more work.
How much it has to do
The number of features usually weighs the most, and a feature is more work than it looks on paper. Take that notification when the bike is ready. Someone in the shop has to be able to mark the bike as ready. The app has to ask permission to send notifications, and there has to be an answer to what happens if the customer refuses. Customers may also want to switch them off again. One line on your list soon turns into a few screens and a few exceptions. That is why it pays to choose sharply what belongs in a first version.
Design counts as well. A product built from standard parts takes less work than one where every screen and every movement is designed from scratch.
Which devices it runs on
A web app in the browser is one product that opens on any device. A real app has to work on iPhone and on Android, and those are two different systems. With tools like React Native you can share a large part of the code, but you still test on both, and every version goes through Apple’s and Google’s review.
So the question of whether you need an app is also a question about price.
Who works with it
One type of user is the simplest. Every extra type brings its own screens and its own permissions, meaning rules about who can see and change what. In the bike shop a customer only looks at their own repair, an employee at all the repairs of the day, and the owner may also want to change prices.
Which software it has to talk to
An integration lets two programs exchange data directly, so nobody has to retype anything. How much work that is depends partly on the other side: how well that software is documented and what it allows.
Only reading data is usually less work than also writing it back, because with writing everything in the other system has to stay correct. Felix, for example, is built to only read from a boutique’s till. There is more on integrations in Connect your software to your till and your accounting.
What is unusual about your product
Some requirements take a lot of work by themselves. An app that has to work with no signal needs extra work to store data and bring it back in line later. On WTN App drivers could sign for a load with no signal, and the chain reconciled when any party came back online.
Legal rules can weigh in too. If a load on WTN App contained hazardous material, the app generated the extra consignment note, with the right signatures on the right document. Online payments and AI are other examples of requirements like these. Mention them early, because they change the estimate more than an extra screen does.
What keeps running after launch
An app that is live keeps costing money. The software has to run somewhere, and that is called hosting. Apple and Google release new versions of their systems, and your app has to keep working with them. Services you build in, such as text messages, emails or AI, often charge per use, so that cost grows with your product.
Apple and Google also charge a fee for a developer account, and they take a commission on certain sales in the app. Those rules change, so check them before you do your sums.
Where you can save
You save the most by building less. A smaller first version is the simplest way, and what can be done by hand for now doesn’t need to be in software yet.
Another way is to find out earlier that something doesn’t work. Changing a screen in a sketch takes far less work than changing the same screen once it is built. I would rather throw away the wrong idea cheaply than ship it expensively, and you can test your idea before there is any software.
Sometimes what you are looking for already exists, and then an existing tool can be the right answer.
The right price
Founders sometimes think an app costs a lot of money. When that comes up, I first want to know what “a lot” means to you, because that differs widely from person to person.
I want to build value for the right price. If you have more to spend than the product needs, I don’t add things you don’t need just to use up that budget.
When we have talked it through, you get a written plan: what it will cost you and what you get for it.
Read next
What's your idea?
Tell me what you're trying to build. I try to reply within one working day.