Project Estimation with Brett Harned
Estimating project work is difficult because the number must be useful before all the details are known. This episode summary turns the discussion into a practical workflow: clarify the purpose, estimate at task level, learn from baselines and retrospectives, and negotiate scope when timing changes.
Start with purpose, not a number
An estimate is a working judgment, not a guarantee. Before naming a figure, establish what the work is meant to achieve, who is affected by the decision, and when the people doing the work are actually available. Those questions prevent a plausible-looking number from hiding missing context. Once the purpose is clear, the estimate can be explained in terms of the work required rather than presented as an unexplained deadline.
Build the total from task-level work
Divide the project into individual tasks and estimate each one separately. The resulting total is easier to challenge, adjust and convert into a schedule because every part has its own reasoning. If a task is too large to estimate credibly, split it further until the work, dependencies and likely effort are visible. Team participation matters here: people who help build the numbers are more likely to understand and stand behind them.
Use baselines and retrospectives to improve
Keep a baseline for comparable work and review performance during sprint retrospectives. Comparing a new plan with what earlier projects actually required exposes optimistic assumptions and recurring bottlenecks. The goal is not to punish variance; it is to carry better evidence into the next estimate. Over time, the baseline becomes a shared reference point for scope, effort and timing.
Negotiate scope when the date moves
When a stakeholder asks for a faster result, walk through the task breakdown and make the trade-off explicit. The useful response is not simply refusal: explain what would need to be removed, deferred or added to make the earlier date realistic. A tight estimate can also serve as a creative constraint, forcing choices about what matters most instead of allowing every request to remain in scope.
About the conversation
Nathan and John discuss project estimating with Brett Harned, whose background includes Razorfish and Happy Cog and who founded the Digital PM Summit. The conversation focuses on common estimating mistakes and practical ways to make future estimates more reliable. The episode runs 28:31 in the supplied episode information.
| Check | What to confirm | Why it matters |
|---|---|---|
| Purpose | Goal and intended outcome | Prevents estimating the wrong work |
| People | Stakeholders and decision owners | Makes assumptions visible |
| Availability | When contributors can do the work | Connects effort to a realistic timeline |
| Breakdown | Task-level estimates with dependencies | Makes the total defensible |
| Baseline | Comparable past work and variances | Improves the next estimate |
| Trade-off | What to cut or add if timing changes | Turns a faster request into a clear choice |
Use as a pre-estimate checklist; it does not replace project-specific evidence.
Frequently asked questions
What should an estimate be treated as?
Treat it as a reasoned approximation, not a promise of an exact outcome. Clarify goals, stakeholders and availability first, then build the number from smaller tasks so the reasoning can be inspected.
How can later estimates become more accurate?
Break the work into task-level pieces, compare the plan with a baseline from earlier work, and review what happened in retrospectives. That evidence makes the next estimate easier to defend.
What should you do when someone asks for delivery faster?
Show the task breakdown, explain the trade-off, and offer a choice: keep the scope and accept the timeline, or remove or add work to change the timeline. A flat refusal without alternatives rarely helps the decision.