At Deeson, we’re always experimenting with how we can use Scrum in an agency environment.

The mechanism we’re seeing the most success with is planning poker.

Accurate estimations

As a solutions architect, I might have estimated tasks on behalf of the development team. But using this traditional method, we found our estimates were often higher or lower than actual time taken for the task. While these would often balance out, better decisions could have been made if estimates were more accurate.

I’ve found planning poker to be a rather neat idea and worth getting to know. The core concept brings the project team together to estimate on tasks. This team can include developers, project managers, user experience, designers and sometimes even the client.

The concept

Each team member uses a deck of cards representing numbers similar to Fibonacci’s sequence (½, 1, 2, 3, 5, 8, 13 ,etc). Once the numbers reach more than 13, the decks we use break into bigger increases. This is because laying a high numbered card is a sign that the task needs to be broken down anyway.

Agreement is the key

After a quick overview, each member simultaneously lays their estimate down on the table for each task. Team members with the highest and lowest estimations then state their cases and everyone estimates again until everyone agrees. 

(*whispers* Don’t tell the Scrum police, but we plan our tasks in hours and not days. This is because our clients buy hours, so we think in hours. But we always measure the accuracy of our estimates.) 

Bringing teams together

I find that when you start thinking about how long something is going to take you, it’s difficult to do without considering how it might be built, at least at a high level. 

This is where you can think about efficiency by reusing code or patterns. Planning poker brings everyone together which means you’re calling upon the much broader experience of your team to make the decision. This leads to better estimation. 

Deeson tips

The Deeson developers love the planning poker process because the team then owns a realistic estimate and have conversations that wouldn’t surface with the traditional process. 

Our top tip for making it work for you is, be flexible, but embrace the theory. 

I've found you can be too strict in using the card deck, although most purists would say to do this. 

For example, when trying to reach a consensus on an estimate, it’s best to stick to laying a single card to minimise the options available. So if you really think it’s four hours, lay a three and a one. This is important to us at Deeson because we plan in hours and we need to be as realistic as possible. 

The important thing is that you can justify why you did it, share with the team, listen to feedback and agree on the final estimate. 

Things to avoid

If you adopt strict estimation without embracing the theory, you go through the motions without the right amount of discussion or logical agreement. This, ultimately, provides a bad estimate. 

Planning poker is collaborative, accurate and fun. Give it a go….