Quickly grooming large backlogs

After a period of time every backlog builds up a number of Stories that will never get done. There are many reasons for why you wouldn’t want to do a particular story in a backlog, these include (but are not limited to):

  • Business changed its mind
  • Already fixed
  • Duplicate of another better written Story
  • Incredibly low or dubious business value
  • Badly written or just forgotten what it was about
  • Change in business strategy
  • Change in legal environment
  • Change in competitive environment
  • Not technically possible
  • It was a shit story to begin with

So how do we quickly identify dead stories and remove them from the backlog?

Whole Team Grooming

One way to quickly identify dead stories in a large backlog is to set up a session with the Product Owner, Team and relevant stakeholders and get them to review the backlog collaboratively.

Prior to the session print every story in the backlog and randomly place all of the story cards on a large table. When the participants arrive ask them to group the stories into related areas and while doing that to identify any stories that are no longer needed.

Have a laptop handy so any questions about the detail of a story can be quickly looked up.

Identify High Priority Stories

Add value to a backlog review by identifying any high priority Stories as you go. Sometimes in a large backlog some high priority stories can get lost or forgotten. This will help the product owner make sure that they are planning the right work into the upcoming iterations.

Top Tips

Print Story cards if you can. JIRA and Rally both have ways to do this. If you are using another tool it might be a good idea to export it into a csv and them use mail merge to create easy to print cards for large backlogs.

Take all of the seats out of the room to encourage people to keep moving and get the job done.

Many hands make light work. Book the session at a time when everyone is available and not towards the end of a sprint.

Make it fun. Bring doughnuts. Have a prize for the person who removes the oldest story or the worst written story.

Results

The first time we did this exercise in my current role we found that 11% of tickets were no longer needed.

The second time we did this was after a major release and we found that a massive 42% of tickers were DONE, not needed, duplicates, related to the old brand or functionality.