How we estimate projects

Software estimation is like planning a trip. You make educated guesses on the cost and time required based on the destination, your experience and research. However, just as unexpected events such as flight delays or weather conditions can impact a trip, unexpected factors such as changes in project requirements or technical challenges can impact software development.

Like with planning a trip, software estimation requires experience, expertise, and a willingness to adapt to changing circumstances.

At AZKY we use T-Shirt sizing to estimate the duration and cost for building apps

We list all the features of your app and assign T-shirt sizes to them. Every line item is assigned a S, M, L, XL or XXL. This framework has time and again helped us successfully achieve estimates that capture the complexity of the build and are very close to the actual timeline of the build.

See below an example:

• S - Simple things like sign up/login
• M - ID verification with an API
• L - Stripe Connect marketplace

As a rule of thumb, features that take less than a day to complete come in the S or XS category. It may look like an overkill talking in "days" for basic features which could be implemented in a matter of a few minutes. But, every new feature has to be made fully responsive, has to go through a Q/A cycle and be tested for various scenarios to be marked as complete.

How can I use T-shirt sizing for my project?

Before we officially kick-off the project, we will ask you to fill up the T-shirt template with all the features and the flow for the app. Don't worry. We'll go through the entire process with you and polish it up together.

We use the formulas on the template to come up with very optimistic and very pessimistic time estimates. If the optimistic estimate is about four weeks then the pessimistic estimate will be about sixteen weeks.

That's a 4x difference! So how much time will it actually take?

The estimation is ballpark at the end of the day and depends on lots of factors, e.g. are we using our recommended design system or a custom design?

If you aim for as lean an MVP as possible then we are close to the optimistic estimate

We generally multiply the timeline estimate by 1.5x to leave the buffer for kickoff calls, wireframing, QA, testing and tuning with a few more S and M tickets, etc.

Overall, you should keep in mind a figure of six to eight weeks for a full build of your MVP. Business tools are generally far more complicated and can go on for months. But, we split the builds into phases. 8 weeks for phase 1, test and then move on to the next phase and add more features and use cases.

We hope this helps you to understand the process better. If you haven't already, please also watch the video above and if you still have questions about how this mechanism works, you can always contact us.