Estimation Poker

Introduction

Estimation poker, also often referred to as planning poker, is a good technique for a team to estimate points. In order to do this, a workshop is scheduled with the team. The raw material is the work items to be estimated. Each individual work item is now estimated using the following interesting procedure:

Explain

One of the participants will take the lead and explain each work item, what needs to be realized, and, if possible, what techniques, platforms, and services are to be used.

Estimates

Each participant has a set of cards that depict the values of the applied scale. This can be a set of playing cards, but predefined sets of cards are often available for known scales. After a work item is explained, all participants will lay a card from their own set down on the table that specifies which value they attach to this work item. This value represents the total complexity or amount of work that must be done to complete that work item, not just the amount of work for a certain role.

Evaluate

Once the cards are on the table, they’re evaluated. If the range of values is close to each other, the evaluation is simple. For example, if only fours and fives are showing, there’s nearly consensus. The team now simply chooses one of these values.

Explain

When the values are further apart, something more needs to be done. The two participants furthest apart now briefly explain why they chose their values. Keep in mind that neither of the selected values is wrong, even though their opinions differ. Maybe the customer opted for a two, and a developer for an eight. Perhaps the customer underestimated the complexity, or the developer misinterpreted the work item.

Re-estimate

If the values differ, the participants take their cards back from the table and again put a card on the table which represents the value for complexity based on the new insights. It’s likely that the values are now closer together. An estimate is then achieved.

If the values still differ strongly, the evaluating, explaining, and re-estimating can be repeated. If, after several rounds, no consensus is reached, this is usually because the work item requires further investigation. In Agile projects, this is indicated by choosing the maximum value from the scale.