3.0 KiB
Duration: 30--40 minutes
Definition
Estimation predicts the effort or time required for work. Forecasting combines estimates with capacity, dependencies, uncertainty, and current progress to predict delivery. Keep estimate, target, and commitment separate.
Simple Mental Model
Estimate the work. Forecast the date. Agree on the commitment.
How It Works
A target such as "leadership wants September 1" is not automatically an estimate or commitment.
Estimate milestone by milestone after outcome, scope, and major dependencies are understood. Then account for real constraints.
Requirements 1 week
Implementation 3 weeks
QA 1 week
Deployment 1 week
This does not automatically equal six calendar weeks. People have operational work, dependencies create waiting time, work may overlap, and uncertainty exists.
Use ranges when uncertainty is high. "6--9 weeks" can be more responsible early on than false precision.
State assumptions:
Six weeks assuming DevOps delivers the environment by Aug 10 and the owner has roughly 50% project capacity.
If assumptions change, the forecast should change.
Example
An owner proposes eight weeks:
- 3 weeks implementation
- 1 week QA
- 1 week deployment
- 1 week Product validation
- 2 weeks availability/operational interruption
Do not simply change eight to five. Challenge the assumptions. Can work overlap? Is the operational allowance supported by history? Can a dependency start earlier?
My Company / Real-World Context
Your projects are interrupted by tickets and production issues, so project effort and calendar duration differ. An 80-hour project can take many calendar weeks if only 10--15 hours per week are realistically available.
CTO Perspective
Aggressive dates do not create capacity. If a target must move earlier, identify the trade-off: scope, priority, capacity, sequencing, dependency acceleration, or increased risk.
Questions to Ask
- What is the estimate based on?
- Which milestone has the most uncertainty?
- What assumptions are we making?
- What capacity is actually available?
- What is the requested target?
- What can the team reasonably commit to?
- What would need to change to deliver sooner?
Key Takeaways
- Estimate, target, and commitment are different.
- Forecasts combine effort with real constraints.
- Ranges are useful when uncertainty is real.
- Important assumptions should be explicit.
- Challenge estimates by challenging assumptions, not by arbitrarily shortening them.
- Faster delivery normally requires a trade-off.
How It Fits Into the Bigger Picture
This lesson builds on Projects vs Operations, [[Outcomes and Definition of Done]], Scope and Scope Creep, and Milestones vs Tasks. It prepares the project system to become more predictable and scalable.
Related Concepts
- Milestones vs Tasks
- Dependencies
- Prioritization
- Capacity and WIP
- Risk Management
- Project Health
- Portfolio Management