Planning Poker | Agile Scrum Master
Planning Poker is a consensus-based estimation practice where team members independently select a size from a shared scale, then discuss differences until they converge. It improves shared understanding and surfaces hidden work, risks, and assumptions before committing to a plan. It is commonly used during backlog refinement and Sprint Planning to size user stories that are small enough to deliver and to decide when an item should be split or clarified. Key elements: a relative scale, a facilitator, independent selection, time-boxed discussion, convergence rules, and captured assumptions and dependencies.
How Planning Poker works
Planning Poker is a team-based estimation technique that builds shared understanding through structured, time-boxed discussion. Each participant selects an estimate independently, then the team explores differences and converges on a relative size. It reduces anchoring bias and helps ensure the estimate reflects the team’s collective knowledge, including hidden work, risks, and dependencies.
Planning Poker works best as a short learning loop, not as a negotiation. The goal is transparency (make assumptions and scope drivers visible), inspection (use disagreement and outliers to spot uncertainty), and adaptation (decide the next action: split, clarify, prototype, spike, or defer). The estimate is a useful byproduct, but the primary outcome is better near-term decisions and a healthier flow of delivery.
Purpose and Value
Planning Poker addresses several core needs in Agile estimation:
- Equal participation - everyone estimates independently, reducing seniority and role bias.
- Reduced anchoring - simultaneous reveal prevents early opinions from steering the group.
- Focused clarification - differences surface unclear scope, missing acceptance criteria, and hidden dependencies.
- Decision support - outliers trigger actions that improve outcomes, such as splitting work or running discovery.
Planning Poker scale choices
Planning Poker uses a shared scale so estimates are comparable. The scale is often non-linear to reflect increasing uncertainty as items get larger.
- Modified Fibonacci - values such as 1, 2, 3, 5, 8, 13, 20, 40, 100 to discourage false precision.
- T-Shirt sizes - XS to XL for early sizing when detail is limited.
- Simple sequence - 1, 2, 3, 4, 5 when work is well understood and tighter bands help.
- Uncertainty marker - a “?” or “needs discovery” option to avoid forcing a confident size.
Whichever scale you choose, keep it stable long enough to learn from outcomes. Recalibrate using reference items and what you observe during delivery, rather than changing the scale whenever estimates feel uncomfortable.
Steps to run Planning Poker effectively
A Planning Poker session is a repeatable routine. The steps below keep it fast and focused on understanding and next decisions.
- Present the item - restate intent, acceptance criteria, and any known constraints or dependencies.
- Clarify scope - align on what is included, what is excluded, and what “done” means.
- Estimate independently - each person selects a value without discussion to avoid anchoring.
- Reveal together - show estimates simultaneously.
- Discuss differences - the highest and lowest estimators explain scope drivers, risks, and unknowns.
- Decide and move - converge when it is decision-ready, or stop and trigger splitting or discovery if it is not.
- Capture assumptions - record risks, dependencies, and uncertainty drivers that could change the size.
Benefits of Planning Poker
Planning Poker improves more than the estimate. Its main output is shared understanding that reduces rework and improves flow during delivery.
- Improved clarity - hidden work and missing acceptance criteria surface before implementation.
- Reduced bias - independent selection limits the impact of the loudest voice.
- Better decomposition - oversized or unclear items are identified early and split sooner.
- Stronger collaboration - the team aligns on scope and approach, improving ownership and coordination.
- Earlier risk visibility - disagreements reveal uncertainty and dependencies while changes are still cheap.
Limitations and considerations for Planning Poker
Planning Poker has diminishing returns when used on very large items, when items are not comparable, or when the team lacks context. It also becomes wasteful when the session drifts into detailed design debates that should be handled through spikes, prototypes, or targeted discovery.
- Oversized items - very high estimates usually mean the item should be split or reframed.
- Poor inputs - unclear items produce estimates that reflect ambiguity, not effort.
- Session fatigue - long sessions reduce discussion quality and encourage premature convergence.
- False precision - treating estimates as deterministic promises undermines learning and trust.
Planning Poker in backlog refinement and Sprint Planning
Planning Poker is commonly used during backlog refinement to prepare items for Sprint Planning. When items are clarified and sized ahead of time, Sprint Planning becomes a decision about the best next slice of value toward the Sprint Goal, not an estimation workshop.
Planning Poker should be paired with flow awareness. If work in progress is high or the system has persistent blockers, improving estimates will not fix delays. Use Planning Poker to improve the quality of decisions, then improve outcomes by limiting work in progress, removing constraints, and keeping quality high.
Best practices for Planning Poker
Planning Poker stays effective when teams keep sessions short, calibrate regularly, and treat disagreement as a signal for learning.
- Time-box per item - focus on scope drivers and risks, not on defending a number.
- Use reference items - calibrate what “3” or “5” means by comparing to known work.
- Stop forcing convergence - if uncertainty is high, trigger discovery rather than debating.
- Split large stories - treat high numbers as prompts for decomposition to keep items deliverable.
- Capture assumptions - document what the estimate depends on so it can be revisited.
- Inspect and recalibrate - compare expectations with delivery evidence and adjust how you slice and size.
Misuses and fake-agile patterns
Planning Poker is frequently misused as a mechanism to pressure teams into commitments or to measure productivity. When estimates become performance targets, teams respond by inflating numbers or optimizing for the metric rather than improving value and quality.
- Velocity as a target - using estimates to demand more output rather than to improve planning.
- Overriding the team - manager-led estimates remove the learning purpose and reduce transparency.
- Estimating without refinement - forcing numbers onto unclear items instead of clarifying scope.
- Comparing teams by points - treating estimates as comparable across teams despite different baselines.
Related estimation techniques
Planning Poker is often combinead with Modified Fibonacci Sequence scales, Story Points, T-Shirt Sizing, Affinity Estimation, and forecasting methods that use observed throughput, cycle time, and uncertainty to support planning.
Planning Poker is a consensus-based estimation technique where the team uses discussion and a shared scale to size work items by relative effort and uncertainty