Psychology of Planning Poker Voting - Anchoring and Groupthink
Noticed junior devs always wait to see what senior dev votes before showing their cards. Getting groupthink instead of independent estimates. How do you prevent anchoring bias and get genuine diverse opinions in planning poker?
Simultaneous reveal is critical - use physical cards everyone places face-down then flips together, or online tool with hidden votes until facilitator reveals. Never do verbal round-robin voting where first person anchors everyone else.
Also rotate who explains the story first. If senior dev always presents the story, their framing anchors the entire discussion before voting even starts. Let junior dev read story, ask clarifying questions, THEN estimate. This mixes up influence patterns and surfaces diverse perspectives.
Silent voting ritual: Everyone writes estimate on paper privately, facilitator counts to 3, simultaneous reveal. 30-second silent thinking time before reveal prevents reactive voting.
When estimates differ widely, ask quietest/newest person to speak FIRST about their reasoning. This combats HiPPO bias (Highest Paid Person's Opinion). Senior can't dominate if junior explains their 8pt estimate before senior explains their 3pt.
Blind voting online tool completely solves this. Everyone selects their card privately in the app, no one sees any votes until facilitator clicks "Reveal All" and all estimates appear simultaneously.
Eliminates anchoring entirely. We saw estimate diversity increase 40% after switching from physical cards (where people could see hesitation and card selection) to blind online voting. Juniors vote confidently now.
Rotate estimation facilitator. Different person leads each story discussion, senior devs sit out occasionally and just observe. When junior dev facilitates the estimation, they can't mentally anchor to senior opinion because they're running the show.
Builds confidence and reveals insights seniors miss. Our junior QA engineer caught a complexity the staff engineer overlooked because she was facilitating and had to think independently.
Controversial take: Some "groupthink" is actually good. Team converging on estimates after discussion means they've achieved shared understanding - that's the goal!
Forcing artificial diversity creates false precision. If team naturally aligns around 5pts after healthy discussion, that's calibration not bias. The conversation happened, perspectives were shared. Don't solve a problem that doesn't exist just because a book said anchoring is bad.