What does our estimation process look like?

We use such approaches depending on provided details from the client's side. The goal is to give the most precise estimate based on available information.

1. Project size

In this step, we define the size of the project. Here we provide the most abstract and uncertain calculation, but at the same time, it's the fastest way to give the client an understanding of the project scale, not any exact number.

For instance:

A small project is less than 50,000 USD

A medium project is 50,000 to 200,000 USD

A large project is more than 200,000 USD

2. Define people and technologies

This step defines all potential stakeholders and what technologies will fit the project.

3. Estimate by analogy

We compare the project to our experience in the past. The questions to ask at this step:

Have we finished similar projects before??

What were the costs??

What are the differences between these projects??

The rule of thumb is that the same amount of work done by the same people will take the same amount of time.

4. Work breakdown structure estimate (bottom-up)

This is the most accurate high-level estimate, breaking the project scope into categories such as infrastructure setup, front-end development, back-end development, testing, etc. Estimates are given for each type, and the overall cost is calculated based on these numbers. There are some rules to this step:

100% Rule: Estimate should cover 100% of work. There should be no hidden activities.

No Overlap Rule: Estimates from different areas should not describe the same activity.

80-Hour Rule: Every area of work taking more than 80 hours should be split into smaller ones.

The questions to ask in this step are:

What are the main areas of work?

What are the deliverables?

What are the activities needed to complete the job?

What are the risks, dependencies, and uncertainties?

The person giving the estimate should answer these questions. There should be no blockers for this step.

Less information would result in a broader estimate range, and more information would result in a narrower range. We can adjust the numbers based on further communication with the client.?

Every person should give two estimates for each area of work:

The low estimate, aka Aggressive But Possible.?

It means a 50% possibility of completing the work in a given time with reasonable effort. It is usually optimistic but not purely theoretical or based on luck.

The high estimate, aka Highly Probable.?

It means a 90% possibility of completing the work in a given time, considering any real problems that might arise.

We can calculate the buffer based on these numbers, which are added to the sum of all ABP estimates:

The buffer is calculated using the following formula:

The estimate can be given to the client as a single number with some added margin. In addition to that, the table will show the amount of uncertainty for every part of the work. This is opposed to the traditional approach of giving a broad range, which usually results in the client's expectations leaning towards the lower value.

If you like this information, please ask questions in the comment section.

I can also share the spreadsheet calculator for those who already want to use this method on projects. Just let me know!