How much discussion time is too much in planning poker?
Our planning poker sessions are dragging on forever. Yesterday we spent 15 minutes debating a single story about adding a filter to a report page.
How long should discussions last before we just pick an estimate and move on? What's the right balance between thorough analysis and efficiency?
Rule of Thumb: 5 Minutes Max Per Story
Here's the protocol that works for most teams:
- Read story + clarifying questions: 2 minutes
- First vote + reveal: 30 seconds
- Discussion (if votes differ): 2-3 minutes
- Second vote: 30 seconds
- Decision: immediate
Total: ~5 minutes per story
When to Cut Discussion Short:
- Team is debating implementation details (not estimation)
- Same two people keep arguing back and forth
- Discussion reveals missing requirements (defer story, get PO clarification)
The 2-Vote Rule:
If votes don't converge after 2 rounds, take the higher estimate and move on. You're spending more time estimating than it would take to just build the thing.
Signs You Need to Stop:
- "We've been discussing this for 10 minutes"
- "We're designing the solution instead of estimating complexity"
- "Nobody has new information to add"
Remember: Planning poker is about quick, good-enough estimates, not perfect ones. The goal is predictability, not precision.
I use a timer. Literally set a 5-minute timer on my phone for each story. When it beeps, we vote one last time and move on. Teams adapt quickly when they know time is limited.
The 5-minute timer idea is great! I'll propose this to the team. We definitely fall into the trap of designing solutions instead of just estimating complexity.
Also: if a story consistently generates long debates, it's probably too big or too vague. Either split it into smaller stories or send it back to the product owner for more detail.