Agile Estimation Beyond Story Points: T-Shirt Sizing, Item Counting and #NoEstimates
Story points are the best-known Agile estimation technique, and also one of the most misunderstood. Many teams struggle with inflated points, endless debates about whether something is a 5 or an 8, and managers who convert points into hours or compare teams. The Scrum Guide does not require story points; it does not mandate any particular estimation method. Teams are free to choose what helps them plan. This lesson reviews alternatives, explains when each fits and offers a decision guide. For the basics of points, read Estimation and Story Points in Scrum and Planning Poker in Scrum.
What Estimation Is For
Before choosing a technique, remember the purpose: to help the team decide what fits in a Sprint, to help the Product Owner order the backlog by value versus effort, and to give stakeholders a rough idea of timescales. It is not a performance metric and not a contract. The goal is enough accuracy at minimum cost; time spent on estimation is time not spent building.
Technique 1: Relative Sizing With Story Points (Recap)
The team assigns points on a scale such as the modified Fibonacci sequence (1, 2, 3, 5, 8, 13) to reflect relative effort, complexity and uncertainty. Strengths: avoids false precision of hours, encourages discussion. Weaknesses: drift over time, confusion with hours, misuse as productivity measure.
Technique 2: T-Shirt Sizing
Items are sized as XS, S, M, L, XL. The scale is intentionally coarse, which makes it quick and avoids the illusion of precision.
- When it helps: early roadmap-level estimates of large items such as epics, quick rough-cuts of a big backlog, and teams new to estimation.
- How to use it: choose a reference item for M; compare each new item against references. Optionally map sizes to numbers (for example S=2, M=5, L=8) if you want to use them for rough forecasting.
- Watch out for: overly many categories; the point is speed.
Technique 3: Affinity Estimation (Wall Estimation)
The team silently arranges backlog items on a wall from smallest to largest, then discusses and adjusts, and finally groups items into size buckets. It is excellent for estimating dozens of items in an hour.
- Write each item on a card.
- Draw a line on the wall from "smaller" to "larger".
- Team members silently place cards relative to each other; move cards freely.
- Pause to discuss any card that keeps being moved.
- Label the groupings with sizes.
It works well for a first-time backlog sizing, release planning or when a new team needs a baseline.
Technique 4: Right-Sizing (Slicing to "About One Size")
Instead of estimating each item, the team agrees on a target size, such as "small enough to finish in about three days", and splits anything larger. Then every item is treated as roughly equal, and the team simply counts. This approach moves effort from estimating to splitting stories, which produces better flow and clearer scope.
Technique 5: Counting Items and Using Throughput
If items are right-sized, the number of items completed per Sprint (throughput) becomes a stable measure. Forecasts then use counts of remaining items and historical throughput. See Forecasting in Scrum and Agile for worked examples and Monte Carlo simulation.
- Advantages: no estimation meetings, no points to inflate, easy to explain.
- Disadvantages: requires reasonably consistent item size, and some history.
Technique 6: The "#NoEstimates" Movement
The #NoEstimates idea, associated with practitioners such as Woody Zuill and Vasco Duarte, challenges the value of estimating at all. The argument is that, with small, valuable slices and continuous delivery, forecasting using flow data makes detailed estimation unnecessary. It does not mean "never forecast"; it means replace effort estimates with flow-based forecasts and focus on delivering the most valuable work first. It works best in mature teams with a steady flow of small items and stakeholders who accept probabilistic forecasts.
Technique 7: Ideal Days or Hours
Estimating in ideal days (time without interruptions) remains in use, especially for task-level planning in the Sprint Backlog. It can help teams plan their capacity within a Sprint, but is prone to false precision and to being treated as a commitment. Use cautiously and never convert story points to hours as a rule.
Comparison Table
| Technique | Speed | Precision | Best used for | Main risk |
|---|---|---|---|---|
| Story points | Medium | Medium | Sprint planning with a stable team | Inflation, misuse as productivity |
| T-shirt sizing | Very fast | Low | Roadmaps and rough sizing | Too coarse for Sprint planning |
| Affinity estimation | Fast for many items | Low to medium | Sizing a large backlog quickly | Group dynamics, anchoring |
| Right-sizing and counting | Fast | Medium (with history) | Flow-based teams, Kanban and Scrumban | Uneven item size |
| #NoEstimates style | Fastest | Depends on flow data | Mature teams with small items | Stakeholder discomfort |
| Ideal days/hours | Slow | Looks high, often false | Task-level capacity planning | Treated as commitments |
How to Choose
- What decision will the estimate support? Roadmap choices need rough sizes; Sprint planning needs finer detail.
- How stable is the team and its work? Stable teams can use points or counts; changing teams get little from either.
- What do stakeholders need? A range with confidence levels is often enough.
- How much does estimating cost? If the team spends hours each Sprint debating numbers, try a lighter technique.
- Experiment. Use the Retrospective to try a different method for three Sprints and compare forecasting accuracy.
Practices That Improve Any Estimation Approach
- Estimate as a team, including everyone who will do the work.
- Use reference items (a known "typical" story) to anchor comparisons.
- Discuss big differences between estimates to uncover assumptions; this conversation is the main value.
- Time-box estimation sessions.
- Re-estimate only when understanding changes materially.
- Never use estimates to judge individual performance or compare teams.
Common Mistakes
- Converting points to hours "so management understands".
- Estimating items that will not be built for months.
- Treating an estimate as a commitment and blaming the team when reality differs.
- Letting the most senior person set the number before others have voted.
- Spending more time estimating than building.
- Scrum does not require story points; choose what helps your team plan.
- T-shirt sizing and affinity estimation suit rough sizing; right-sizing plus counting suits flow-based forecasting.
- The conversation matters more than the number.
- Do not use estimates as performance measures.