What to estimate before you guess?

I am frequently asked how much it will cost to make things. I’ve needed to answer this question for probably 10 years now. Before I attempt to estimate features or consider the team doing the implementation, I often think about the following five things.

These have all been learnt the hard way.

How experienced is the project manager* (Geoff)?

This is a thorny one but the most important. You don’t want to say to “yeah… So Geoff, I’ve assessed your ability as a project manager and need to add $100k to the project ”. However, you do need to do that assessment, and ironically, if Geoff is an experienced project manager, he would already know that!

There are two parts to assessing Geoff, first does he have experience of running any project, and second has Geoff ever managed a software project? Geoff will need extra tools in his belt to manage the latter. If he doesn’t have them, he’ll need to learn and that takes time. If you are Geoff, be honest, add extra time if you’re inexperienced, it will work in your favour at the end of the project.

Rough calculation: add 10% extra for inexperience.

Has anyone thought about the real problems to solve? Do they involve fixing something real people do?

If the answer is no, chances are the project will swing and move (and take longer). It’s also an indication that this is a pet project so it becomes about making things the project manager wants rather than the users.

Rough calculation: Add a few weeks to help Geoff think the project through, if at the end, it’ s still not clear, try to avoid it or add 50% to the budget. It’ll give you that time to learn what’s important to do.

Is the project decision maker obvious or hidden?

Sometimes Geoff has a manager (Dorothy) and it can turn out that he doesn’t have enough authority to choose priorities or direction. In this case the best thing is to speak with Dorothy, if you can’t do that add 20% for project changes beyond those expected.

How many other ‘stakeholders’ are involved (aka how many logos need to be on the site)

Add 2% for every logo. It will take longer to manage, there will be more changes, more opinions and more negotiating. Stakeholders can be good, but they come at a cost to delivery.

What does the internal approval process look like?

The more steps in the approval process the longer things take and the more uncertainty around acceptance of new features you’ll face. I’d factor in a 5–10% increase in budget to counter this.

*Our team typically works with the rough structure of Product Manager, Designer, Lead Engineer + more Engineers/Designers/Geospatial folks when needed. The project manager in this case is the person in charge of a larger project where building technology is one piece. They could equally be a client.