Deadlines drive most decisions in business, either because they drive revenue or they are one domino in a chain of events that drives revenue. Given that deadlines are important, adequate attention needs paying to estimation quality.
Fundamental steps can be followed to make estimates reliable, whether teams use techniques such as guessing, Planning Poker, Delphi, historical data, or a mathematical model.
Possible Causes of Unreliable Estimates
Here are four (of many) to consider:
- People are unsure of how to estimate: Instead of robust estimates based on analysis, estimates and commitments are rough calendar predictions without factoring in resource availability, current commitments, dependencies or variations in scope. They are doing their best with the estimation knowledge they have, but it’s not what you need to run a business.
- People tell you what you want to hear: When you ask for an estimate (while reminding the project manager of the existing hard deadline), you might be just hearing back, “Sure, we can do that.” This is not an estimate. It is a way to avoid confrontation or avoid looking weak in the moment.
- A single estimate with no assumptions is provided even when project scope is variable: If the scope of the project has not been nailed down yet (and it may never be), then any single estimate, with no accompanying assumptions, is unreliable from the outset. For example, “The (undefined) data translation system will be finished June 15th at 4.05 PM.”
- Risks that might impact the estimate are not assessed or mitigated: Assuming the project will proceed with no problems immediately makes an estimate unreliable.
There are some fundamental steps upon which all estimates can be based.
- Educate people how to estimate so that they are able to provide good data. An estimate should at least include:
- What is in scope (and what is out of scope)
- Uninterrupted time needed
- (Optionally) size estimates such as the number of story points or reports
- The definition of the units
- Project assumptions (conditions that must remain true for the estimate to be valid)
- Estimate and date options, based on scope options, actual resource availability, and risks.
- Remind people that when you ask for an estimate, you want more than a, “Next month is possible” response. You want an estimate that is reliable, contains effort and calendar time, and options.
- Ask for an estimate range with assumptions when the project scope is clearly ambiguous:
- The ambiguity of the project can be used to your advantage to generate options, e.g., scope A, B, C is 100 days; A, B is 50 days, A is 25, D is unknown and will be estimated after A is complete.
- There are two choices with item D. Work with the customer to define D in the way they want and build credibility or provide a date, pretend it will work out and lose credibility.
- Ask for risks (potential problems) that might make the estimate unreliable.
Your staff can estimate given some guidance and support, and they can use the data to manage their own work and priorities. When they do so, you will have less stress, less surprise, and be able to meet commitments.[Forward this email to your boss! Subject: Here’s a cool tip for you] Quick Link
If you have questions or comments about this article, or would like to discuss your estimation or other challenges, please contact us.