How to estimate a web development project

Think of a number, double it and then run 16 months late

Software estimates.

Everyone's favourite subject.

The rule of thumb is "think of a number, then double it. Double it again and again, then tell the client. Realise you were wrong, multiply it by 6 and then overshoot by six months".

But why is that?

It's because you're starting from scratch every time.

Building from first principles.

Trying to estimate unknowns.

It doesn't have to be like that.

There are patterns to the work you do.

Stuff you can reuse.

Things you know how to do.

You can choose to build your business around that and suddenly you've reduced the number of unknowns.

And with fewer unknowns come more accurate estimates.

The key is to stop saying yes to every bit of work and focus in on what you're already good at.