March 15, 2026 · By Raksit Nongbua
Advanced Planning Poker: Handling Outliers and Deadlocks
The goal of planning poker is consensus, but what happens when the team is stuck? When one person sees a 2 and another sees a 13, or when three rounds of voting produce no movement, the session can feel like a waste of time. This guide explores advanced facilitation techniques to handle outliers, break deadlocks, and keep your planning sessions productive.
The Value of the Outlier
First, it is important to reframe how the team views disagreement. A wide gap in estimates is not a failure of the process — it is the process working exactly as intended. The gap is a signal that information is asymmetric. Someone knows something the others do not.
Instead of viewing an outlier as a "problem to be fixed," view it as a "source of hidden requirements." The goal of the discussion is not to pressure the outlier to change their mind, but to surface the assumptions that led to their number.
Techniques for Handling Disagreement
The "High-Low" Discussion The standard approach: ask the person with the highest estimate and the person with the lowest estimate to explain their reasoning. Pro-tip: let the low estimator speak first. They often have a simple solution in mind that others overlooked. Then let the high estimator explain the risks or complexities they spotted.
Identify "Hidden Work" Often, a high estimate comes from a team member including work that others assumed was out of scope (e.g., unit tests, documentation, refactoring). Ask: "What is included in your 8 that might not be in their 3?" This clarifies the Definition of Done in real-time.
The Three-Round Limit If the team cannot reach consensus after three rounds of voting and discussion, you have hit a deadlock. Continuing to vote usually produces frustration rather than accuracy. It is time to use a deadlock-breaking strategy.
How to Break a Deadlock
When you are stuck, choose one of these three paths:
1. Pick the higher estimate: If the gap is small (e.g., between 5 and 8) and time is running out, default to the higher number. It is safer to over-estimate complexity than to under-estimate it.
2. Split the story: A deadlock is often a sign that the story is too large or contains too many unknowns. Stop estimating, split the story into two smaller pieces, and estimate those instead.
3. Assign a "Spike": If the team genuinely does not have enough information to choose a number, do not guess. Create a "Spike" (a time-boxed research task) for the next sprint to gather the necessary data, and defer the estimation of the story until the spike is complete.
Conclusion
Disagreement in planning poker is where the real value lies. By handling outliers with curiosity rather than pressure, and having a clear strategy for breaking deadlocks, you ensure that your estimation sessions remain a tool for shared understanding, not just a source of numbers.
Run an advanced session now