Given you have the story cards for your epics, how do you put them on a timeline methodically? How do you fix on how much you can deliver in a sprint with consensus?
Card estimation
- Fix on an estimation sequence (Fibonacci, GP, etc.)
- Have all the cards laid out in front of the team
- Cards on the same column would belong to the same feature
- Cards on any given row would deliver value to the user
- As a team, decide upon the smallest card that you can finish end-to-end (
grooming -> dev -> test -> review -> production) - Assign this card the smallest number in your estimate sequence (Ex: 1 in case of Fibonacci)
- Iterate over the card and estimate it as a team,
- Estimations should be based on the relative hardness of the cards estimated so far (clustering)
- Discuss when there’s difference of opinion, an estimate is done only when everyone agrees to it, in case of an inability to arrive at a consensus park the card
- Any assumptions should be called out and made note of (useful context when grooming them later)
- Note down the story number and respective points in a matrix, these are your estimates
You can do multiple rounds of estimation
Multiple rounds increase estimation accuracy as the common understanding of what a card is increases amongst the team
Roadmap
- Now that you have your estimates, keep them aside!
- Shuffle the cards
- Pick a random card and decide as a team if it can be done (in addition to the cards picked) within one sprint
- If yes, continue picking the next card
- If no, stop adding cards and decide on the final set of cards you can do within one sprint (feel free to remove some)
- Once you have the final set, note down the cumulative sum of the card estimates
- Repeat steps 2-6, note down all the estimates
- The average of all the estimates rounded off is the number of story points your team can do in one sprint. You can then come up with a burn up chart to get the release timeline