Free template
Software estimate template
Estimate a feature as a range, not a guess. List the parts, give each a best, likely and worst case, let the formulas roll them up, and end with dates at 50, 80 and 90% confidence. It also has a place for work found in review and for what would change the answer.
The CSV has live formulas and opens in Google Sheets, Excel and Notion. No sign-up.
The template
Seven sections, from scope to the range.
A software project estimate is only as good as the parts under it. Each section below says what to write, and the order is the order you work in.
- 01
Scope and assumptions
What the estimate covers, what it leaves out, and what you are taking for granted.
- What is in, and what will people assume is in that is not?
- Who is building it, and how much of their time is on it?
- Which decisions are already made, and which are not?
TipAn estimate is only valid for its scope. Write the scope down first, or the number will be quoted for a bigger feature.
- 02
Parts by layer
The feature split into parts small enough to estimate, grouped by interface, backend, data and cross-cutting work.
- Best, likely and worst case per part, in dev-days
- PERT per part: (best + 4 × likely + worst) / 6
- Who owns each part, and did they give the numbers?
TipAsk the person who will build the part. Split anything over about five days; big parts hide the unknowns.
- 03
Found in review
Parts nobody planned that a review turned up, kept apart so you can see how much it added.
- Which check or question produced the part?
- Is it security, billing, an edge case or a missing state?
- Does it have its own range?
TipTeams that skip this meet the same parts mid-sprint. A rate limit on outgoing email is a typical one.
- 04
Overhead and margin
The work around the work: code review, deploys, coordination, and room for bad luck.
- What percentage do you add for review, deploy and coordination?
- Is that rate your own, from past features, or a guess?
- How much safety margin for 80% and for 90%?
TipSay the overhead rate out loud. Hidden padding gets cut; a visible 15% gets discussed.
- 05
Total and dates
The result as dev-days and a date at 50%, 80% and 90% confidence. Never one number.
- What is the total at each confidence level?
- Which date do you commit to, and which do you only plan around?
- What focus factor and availability turn dev-days into calendar days?
TipDev-days are not calendar days. A 70% focus factor, as the sprint template uses, is a common starting point; check yours against past sprints.
- 06
Riskiest parts
The parts with the widest spread between best and worst, with the reason and the way to narrow it.
- Which three parts have the widest spread?
- Is it new to the team, or does it touch sign-in or billing?
- Would a one-day spike narrow it?
TipDo the riskiest part first. If it goes badly you find out in week one, while scope can still change.
- 07
What would change the estimate
Open questions, each with an owner, a date, and what it does to the range if answered differently.
- Which answers add a part?
- Which move a range toward its worst case?
- Who decides, and by when?
TipUntil these are answered the range stays wide. Say so, instead of picking the optimistic answer.
Example
An estimate for team invitations, filled in.
Workspace admins invite colleagues by email. These are the parts, ranges and dates of SprintProof’s sample workspace, including the part its gap review added.
1 · Scope and assumptions
- In scope: invite by email with a role, pending list with resend and revoke, accept flow, 7-day single-use links.
- Not included: Domain auto-join, SSO and SCIM provisioning, CSV import of invitees.
- Assumption: each range comes from the person who owns the part.
2 · Parts by layer, in dev-days
| Part | Layer | Owner | Best | Likely | Worst | PERT | Spread |
|---|---|---|---|---|---|---|---|
| Invite modal | Frontend | Sam O. | 1 | 1.5 | 3 | 1.7 | 2 |
| Pending invites list | Frontend | Lena V. | 1 | 2 | 3 | 2.0 | 2 |
| Accept-invite page | Frontend | Sam O. | 1.5 | 2 | 4 | 2.3 | 2.5 |
| Invitations API | Backend | Ken L. | 2 | 3 | 5 | 3.2 | 3 |
| Invite tokens | Backend | Jonas T. | 0.5 | 1 | 2 | 1.1 | 1.5 |
| Invite email | Backend | Ken L. | 1 | 1.5 | 3 | 1.7 | 2 |
| invitations table | Data | Ken L. | 0.5 | 1 | 1.5 | 1.0 | 1 |
| Permission rules | Cross-cutting | Jonas T. | 1 | 1.5 | 3 | 1.7 | 2 |
| Tests | Cross-cutting | Lena V. | 1.5 | 2 | 3 | 2.1 | 1.5 |
| Subtotal planned | 10 | 15.5 | 27.5 | ||||
3 · Found in review
| Part | Check | Owner | Best | Likely | Worst | PERT |
|---|---|---|---|---|---|---|
| Existing accounts join directlySkip sign-up, add membership, notify by email | AUTH-14 | Ken L. | 1 | 1.5 | 2 | 1.5 |
| Seat limit checkBlock over-limit invites, show an upgrade prompt | BILL-02 | Tom A. | 1 | 2 | 3 | 2.0 |
| Rate limit invite sends20 invites per admin per hour, then a clear error | MAIL-02 | Ken L. | 0.5 | 0.5 | 1 | 0.6 |
4 · Overhead and margin
- Planned parts16.4
- Found in review+4.0
- Overhead, 15% (review, deploy, coordination)+3.1
- Base total, the 50% case23.5
The review added 4.0 dev-days of work, before overhead. Planned parts and review parts are the averages of 10,000 simulated runs, so they differ a little from the sum of the PERT column.
5 · Total and dates
| Confidence | Margin | Dev-days | Done by |
|---|---|---|---|
| 50% | +0.0 | 23.5 | Thu, Dec 10 |
| 80% | +3.1 | 26.6 | Mon, Dec 14 |
| 90% | +5.1 | 28.6 | Tue, Dec 15 |
Commit to the 80% date. Dates assume engineers work in parallel about 70% of the time; this sample has two engineers. Yours depend on who is available and your focus factor.
6 · Riskiest parts
- Invitations API 2 to 5Create, revoke and accept all touch the same tokens.
- Accept-invite page 1.5 to 4Four states, and it touches sign-in. Spike it first.
- Seat limit check 1 to 3Billing has not decided: block, warn or charge.
7 · What would change the estimate
- What if the invited email already has an account?Jonas T. · TueAlready counted as “Existing accounts join directly”.
- Can anyone with the link join, or only the invited address?Jonas T. · TueAdds a part: bind the token to the invited email.
- How does the admin find out an invite bounced?Ken L. · ThuAdds a part: bounce handling.
- Seat limit: block, warn or charge?Tom A. · WedMoves the Seat limit check toward 1 or 3 days.
How to use it
How to fill it in, step by step.
Estimating in ranges takes a little longer than guessing one number, and the conversation it starts is the point.
- 01Write the scope and what is not in it. Two lines each. If the scope changes later, the estimate changes with it.
- 02Split the feature into parts. Group them by layer so nothing is forgotten: interface, backend, data, then tests and permissions. Split anything over about five days.
- 03Ask each owner for best, likely and worst case. Best is if it goes well, worst is if the usual surprises happen. The PERT column weights the likely case four times.
- 04Review for what is missing. Walk the edge cases: security, billing, empty and error states. Add what you find as its own table.
- 05Add overhead and read the dates. Put a visible overhead rate on the PERT total, then the margin for 80% and 90%. Convert to dates with real availability.
- 06Name the riskiest parts and the open questions. Share the range with the person who will be asked for a date, and give the date at 80%.
- Working in story points? See how to calculate story points, then turn them into days.
- Setting the process around the date? Use the release management template.
Skip the blank page
Let SprintProof draft it with you.
Describe the feature and the planner asks the questions a reviewer would, then drafts the brief, done-when, not-now and parts with ranges. The check library adds the edge cases teams usually find mid-sprint.
FAQ
Software estimate questions
How do I estimate a software project?
What is the PERT formula?
Why 50, 80 and 90 percent?
What is the difference between dev-days and calendar days?
How do I open the template in Google Sheets, Excel or Notion?
Related templates
Templates that go with it.
- Markdown, CSV
Release plan and release management template
Milestones with ranges, dates at 50, 80 and 90% confidence, the rollout steps and a release readiness checklist.
- CSV
Sprint planning template
Sprint goal, capacity, a backlog with ranges, risks and a definition of done.
- Markdown
PRD template
The problem, who it is for, what done means, what is out, and the open questions, on one page.
- Markdown
Feature spec and feature description template
A feature broken into parts by layer, each with behaviour, states and a range, and a one-page feature description.
- Markdown
Technical design doc and documentation template
Context, the chosen design, alternatives, data, security and rollout, for review before code, and a technical documentation template for after it ships.
Turn this into a reviewed plan with SprintProof.
Describe the feature and SprintProof turns it into a plan: checked against a library of questions reviewers ask, estimated as a range with 50, 80 and 90% dates, and signed off before anyone starts.