Agile Estimation — Making Uncertainty Manageable¶
Glossary¶
Agile Estimation A way of predicting workload using comparison-based methods instead of fixed time guesses.
Fibonacci Numbers A sequence (1, 2, 3, 5, 8, 13, ...) used in estimation because the gaps increase as uncertainty increases.
Sprint / Iteration A fixed-length development cycle (e.g., 1-2 weeks) where the team completes a set amount of work.
Backlog The ordered list of all features, stories, or tasks waiting to be built.
Software projects rarely go as planned. One task expands, another vanishes, and someone always says, “It’ll just take a week” — which almost never happens. Estimation is hard because humans are bad at guessing absolute time (“six weeks”) but surprisingly good at comparing things relatively (“this one’s about twice as much work as that one”). Agile estimation uses this strength.
Story Points: Estimating by Comparison¶
Instead of thinking in hours or days, agile teams estimate with story points — abstract units that measure effort, not time. A story worth 8 points is roughly twice as hard as a 4-point one. The exact duration depends on your team’s speed and will be calculated later.
Tip
Pick one small, clear story as your baseline (e.g., “2 points”). Then compare others: is it 2× harder, 10× harder? Assign points relatively, not absolutely.
To make estimates easier and less falsely precise, we often use Fibonacci numbers:
1, 2, 3, 5, 8, 13, 21, 34, 55...
Each number grows about 60% larger than the one before — matching how uncertainty grows with complexity. You can’t really tell the difference between 34 and 35 points, but you can between 21 and 34: it's a bit more than 1.5× harder, so it gets a bigger jump.
Value Points: Measuring What Matters¶
Story points tell you how hard something is. Value points tell you how much it’s worth.
Story points come from developers — they know the effort. Value points come from customers — they know the importance.
Example
A feature that takes 13 story points but delivers 89 value points is probably worth doing before a 5-point story that only gives 8 value points.
By giving each story both scores, you can decide what to build next — not just what’s easiest, but what brings the most value earliest.
Bang for the Buck: Choosing Smart¶
To find the sweet spot, calculate: Bang for the Buck = Value Points / Story Points
Sort stories by this ratio. The higher the result, the more value you deliver per unit of effort.
Don't build the roof before the walls!
Start from the top of this list when planning your next sprint — but always check technical or logical dependencies before locking the order.
Velocity: Learning Your Real Speed¶
After your first iteration, you’ll see how many story points your team actually finished — that’s your velocity. If you completed 55 points in two weeks, then your “speed” is about 55 points per iteration.
Future planning becomes easy:
If we usually do 55 points per sprint, and the backlog has 165 points left, we’ll likely need 3 more sprints.
Danger
Velocity is not a target. It’s a measurement. Don’t “game” it by inflating story points — you only fool yourself.
When to Stop¶
As the project goes on, you’ll notice something: value per sprint drops. That’s normal. The most valuable stories get done first. When only low-value ones remain, the customer might decide the project is “good enough” and stop — saving money and time.
Quote
“Agile isn’t about doing everything. It’s about doing the right things first — and knowing when to stop.”
Quick Recap¶
| Concept | Who Estimates | What It Measures | Use It For |
|---|---|---|---|
| Story Points | Developers | Effort | Planning workload |
| Value Points | Customers | Business value | Prioritizing stories |
| Bang for the Buck | Calculated | Value / Effort ratio | Choosing what to build next |
| Velocity | Team result | Speed per iteration | Forecasting future work |
In Practice¶
- Pick a small, clear story as your baseline (2 points).
- Estimate all other stories relatively using Fibonacci numbers.
- Have the customer assign value points using the same method.
- Compute Bang for the Buck and sort the backlog.
- After the first iteration, measure your velocity.
- Use velocity × number of iterations to plan and predict realistically.
Tip
Estimation is a learning process. Each sprint improves your sense of scale — the goal isn’t perfect accuracy, but better conversations and smarter priorities.