# [Feature name]: software estimate

**Owner:** [name] · **Status:** Draft · **Last updated:** [date]
**Estimated by:** [names of the people who will build it]

> Free software estimate template from SprintProof: https://sprintproof.com/templates/software-estimate-template
> Replace everything in [brackets]. Delete the guidance lines (starting with ">") when you are done.
> All numbers are dev-days of focused work. Use the .csv version if you want the totals to calculate themselves.

## 1. Scope and assumptions

> What is in, what is out, and what you are assuming. An estimate is only valid for this scope.

- **In scope:** [what the feature does]
- **Not included:** [what people might assume is in]
- **Assumptions:** [people available, existing code you rely on, decisions already made]

## 2. Parts by layer

> One row per part, small enough to estimate in under a week. Best, likely and worst case, by the person who will build it. PERT = (best + 4 × likely + worst) / 6. Spread = worst − best.

| Part | Layer | Owner | Best | Likely | Worst | PERT | Spread |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [part] | [Frontend / Backend / Data / Cross-cutting] | [name] | [d] | [d] | [d] | [d] | [d] |
| **Subtotal planned** | | | [sum] | [sum] | [sum] | [sum] | [sum] |

## 3. Found in review

> Parts nobody planned that a review turned up: security, billing, edge cases, empty and error states. Keep them in their own table so everyone sees how much the review added.

| Part | Source | Owner | Best | Likely | Worst | PERT | Spread |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [part] | [check or question it came from] | [name] | [d] | [d] | [d] | [d] | [d] |

## 4. Overhead and margin

> Code review, deploys and coordination are real work. Add them as a percentage of the PERT total, and say what percentage.

- **Planned (PERT):** [d]
- **Found in review (PERT):** [d]
- **Overhead:** [rate, for example 15%] = [d]
- **Base total:** [d]

## 5. Total and dates

> Never one number. The base total is roughly the 50% case. Add a safety margin for 80% and 90%, ideally from a simulation of the ranges, then convert to dates using who is actually available. Dev-days are not calendar days: apply a focus factor (the sprint template uses 70%).

| Confidence | Safety margin | Dev-days | Done by |
| --- | --- | --- | --- |
| 50% | [d] | [d] | [date] |
| 80% | [d] | [d] | [date] |
| 90% | [d] | [d] | [date] |

## 6. Riskiest parts

> The three parts with the widest spread. Say why each is uncertain and what would narrow it.

1. [part]: spread [d]. [why, and how to narrow it]
2. [part]: spread [d]. [why]
3. [part]: spread [d]. [why]

## 7. What would change the estimate

> Open questions and decisions, each with an owner and a date, and what happens to the range if the answer is different.

| Question | Owner | Decide by | Effect on the estimate |
| --- | --- | --- | --- |
| [question] | [name] | [date] | [adds a part / moves a range] |
