TLDR
There’s no single best estimation technique. Planning Poker works great for sprint-level precision. T-shirt sizing is better for roadmap-level planning. Affinity mapping handles large backlogs fast. The right technique depends on what you’re estimating, how many items you have, and how much precision you need.
This post covers 7 proven estimation techniques, when to use each one, and a comparison table to help your team pick the right approach.
Your Team Keeps Arguing About How to Estimate
One person wants to use story points. Another thinks T-shirt sizes make more sense. Someone heard about “dot voting” at a conference and wants to try it. Meanwhile, the backlog keeps growing and nobody can agree on how big anything is.
Here’s the thing: the estimation technique matters less than having a shared, consistent approach. But some techniques genuinely work better for certain situations. Knowing your options helps your team stop debating and start estimating.
1. Planning Poker
How It Works
Each team member holds a set of cards with numbers (typically the Fibonacci sequence: 1, 2, 3, 5, 8, 13, 21). The product owner presents a story. Everyone privately selects a card. On the facilitator’s count, all cards are revealed simultaneously. If estimates diverge, the highest and lowest estimators explain their reasoning. The team re-estimates until they converge.
When to Use It
Sprint planning and backlog refinement sessions where you need granular estimates on 10-30 items. Best for teams of 3-9 people who are familiar with the work being estimated.
Pros and Cons
Pros: Eliminates anchoring bias. Surfaces hidden knowledge through discussion. Gives equal voice to all team members. Produces consistent, reliable estimates over time.
Cons: Time-intensive for large backlogs. Requires everyone’s full attention. Can feel tedious if the team already has strong consensus on most items.
2. T-Shirt Sizing
How It Works
Instead of numbers, items are categorized as XS, S, M, L, XL, or XXL. The team discusses each item and agrees on a size relative to other items they’ve worked on. There’s no need for precise numerical agreement, just a shared sense of relative magnitude.
When to Use It
Roadmap planning, early-stage estimation, and conversations with non-technical stakeholders. Works well when you need rough sizing on a large number of items quickly. Also useful for teams that find numerical estimates stressful or contentious.
Pros and Cons
Pros: Fast and intuitive. Low barrier to entry. Non-threatening for teams new to estimation. Easy for stakeholders to understand.
Cons: Less precise than numbered approaches. Can be hard to use for capacity planning without converting sizes to numbers. Different team members may have different mental models of what “Medium” means.
3. Dot Voting
How It Works
Each team member gets a fixed number of dots (physical stickers or digital markers). Items are displayed on a board. Team members place dots on the items they think require the most effort. More dots means the team collectively perceives higher effort. The dot counts create a rough ranking by effort.
When to Use It
Prioritization sessions where effort is one factor among several. Quick sizing of large backlogs where you need a rough effort ranking rather than precise estimates. Good for workshops where time is limited.
Pros and Cons
Pros: Very fast. Visual and engaging. Works well in large groups. Doubles as a prioritization tool.
Cons: Low precision. Doesn’t generate discussion about individual items. Can be influenced by placement bias (items at the top of a list get more attention). Not suitable when you need actual effort numbers.
4. Bucket System
How It Works
Create “buckets” representing different effort levels (similar to story point values: 1, 2, 3, 5, 8, 13, 20, 40, 100). The team starts by placing one well-understood item in the appropriate bucket as a reference. Then, remaining items are placed into buckets relative to the reference items already there. Team members can move items if they disagree, triggering a brief discussion.
When to Use It
Large backlog estimation sessions where you need to size 50-200+ items in a single session. Product roadmap planning. Particularly effective when you have many items and limited time.
Pros and Cons
Pros: Handles large volumes efficiently. Creates natural relative sizing. Encourages focused disagreement (only items that get moved trigger discussion). Produces numerical-equivalent estimates.
Cons: Less discussion per item than Planning Poker. Requires physical space or a good digital tool. Quality depends heavily on the reference items chosen initially.
5. Affinity Mapping
How It Works
Write each item on a card or sticky note. Without discussion, team members silently arrange items from smallest to largest effort, grouping similar-sized items together. Once the silent sorting is complete, the team reviews the arrangement, discusses any disagreements, and assigns size labels to each group.
When to Use It
Initial sizing of a new or large backlog. Works well when the team needs to quickly create a relative ordering of many items without getting bogged down in item-by-item discussion.
Pros and Cons
Pros: Very fast for large volumes. The silent phase prevents groupthink and anchoring. Creates natural clusters that can map to T-shirt sizes or story points. Physical and engaging.
Cons: Requires everyone to have basic familiarity with the items. Difficult to do remotely without a good digital whiteboard. Less precise than item-by-item techniques.
6. Three-Point Estimation
How It Works
For each item, the team provides three estimates: optimistic (best case), most likely (normal case), and pessimistic (worst case). These three values are then combined using a weighted formula: (Optimistic + 4 x Most Likely + Pessimistic) / 6. The result gives a weighted average that accounts for uncertainty.
When to Use It
Items with high uncertainty or risk. Projects where stakeholders need to understand the range of possible outcomes. Budget and timeline planning where a single number isn’t sufficient. Common in industries with regulatory or compliance requirements.
Pros and Cons
Pros: Explicitly captures uncertainty. Produces a confidence range, not just a single number. Forces the team to think about risk scenarios. Useful for communicating estimates to stakeholders who need to plan around uncertainty.
Cons: Time-intensive (three estimates per item). The formula can feel overly mathematical for Agile teams. Teams sometimes game it by making optimistic and pessimistic numbers artificially close. Better suited for project-level estimation than sprint-level work.
7. Team Estimation Game
How It Works
Developed by Steve Bockman, this is a structured group exercise. One person places the first item on a table and assigns it a relative value. The next person either places a new item relative to the first (bigger, smaller, or same size) or moves an existing item they disagree with. Play continues around the group until all items are placed and nobody wants to move anything.
When to Use It
New teams that need to calibrate their estimation scale. Teams that find Planning Poker too slow or too formal. Backlog grooming sessions where you want estimation to feel more collaborative and less procedural.
Pros and Cons
Pros: Fast and engaging. Naturally builds team calibration. Low ceremony. Discussions happen only when someone disagrees enough to move an item. Fun.
Cons: Less structured than Planning Poker, which can lead to less consistent estimates. Doesn’t work as well remotely. The first few items placed can anchor the rest of the exercise.
Comparison Table
| Technique | Speed | Precision | Best For | Team Size |
|---|---|---|---|---|
| Planning Poker | Medium | High | Sprint planning | 3-9 |
| T-Shirt Sizing | Fast | Low-Medium | Roadmap planning | Any |
| Dot Voting | Very Fast | Low | Prioritization | Any |
| Bucket System | Fast | Medium | Large backlogs | 5-15 |
| Affinity Mapping | Fast | Low-Medium | Initial backlog sizing | 3-12 |
| Three-Point | Slow | High | High-uncertainty items | Any |
| Team Estimation Game | Fast | Medium | Team calibration | 4-10 |
How to Choose the Right Technique
Ask three questions:
How many items need estimating? Under 30, Planning Poker works well. Over 50, use the Bucket System or Affinity Mapping. Over 100, use T-shirt sizing first and refine later.
How much precision do you need? Sprint planning needs granular estimates (Planning Poker, Bucket System). Roadmap conversations need rough sizing (T-shirt, Affinity). Budget planning with high uncertainty benefits from Three-Point Estimation.
How experienced is your team with estimation? New teams benefit from Planning Poker’s structured discussion or the Team Estimation Game’s calibration effect. Experienced teams can use faster techniques because they already have shared mental models.
Most teams end up using two or three techniques across different contexts. T-shirt sizing for quarterly roadmap planning, Planning Poker for sprint refinement, and Three-Point for high-risk items. That’s perfectly fine. The goal is useful estimates, not technique purity.
Get Practical Agile Insights Every Week
Techniques like these, delivered to your inbox. No certification fluff, no buzzwords. Just what actually works in real teams.
SubscribeFrequently Asked Questions
What’s the most popular Agile estimation technique?
Planning Poker is the most widely used, particularly for sprint-level estimation. It strikes a good balance between speed, precision, and team engagement. T-shirt sizing is the second most common, especially for roadmap and portfolio-level planning.
Can you combine multiple estimation techniques?
Absolutely, and most mature teams do. A common pattern is using T-shirt sizing or Affinity Mapping for initial backlog grooming, then switching to Planning Poker for sprint-level refinement. Use the technique that matches the level of precision you need at each stage.
Do estimation techniques work for non-software teams?
Yes. Marketing teams, design teams, operations teams, and content teams all use these techniques successfully. The principles of relative sizing and group estimation apply to any kind of work. The key is establishing shared reference points for what different sizes mean in your context.
Should estimates be in hours or relative units?
Relative units (story points, T-shirt sizes) are generally better for Agile teams. They avoid the false precision of hours, account for the fact that different people work at different speeds, and reduce the pressure of “getting it right.” If your organization requires time estimates for budgeting, convert from relative estimates after the team has established a reliable velocity.
What if my team hates estimating?
That usually means past estimation sessions were unpleasant, unproductive, or both. Try switching techniques. If Planning Poker felt slow, try the Team Estimation Game. If story points caused arguments, try T-shirt sizing. Some teams thrive with no estimation at all, using throughput-based forecasting instead. The goal is useful planning, not estimation for its own sake.
How accurate do Agile estimates need to be?
Estimates are forecasts, not commitments. They need to be accurate enough to support planning decisions, not precise enough to predict exact delivery dates. If your sprint estimates are within 20% of actual effort over time, you’re doing well. Focus on improving the trend, not perfecting individual estimates.
What’s the fastest estimation technique for a large backlog?
Affinity Mapping is typically the fastest for large volumes. A team of 5-8 people can rough-size 100+ items in under an hour using silent sorting. The Bucket System is a close second, offering slightly more precision with a bit more time investment. Both are dramatically faster than estimating items individually.
Leave a Comment