Some questions to ponder regarding your organization’s use of estimates….

1. Does your need for estimates accidentally cause premature solution convergence (“how can we estimate something unless we know the solution?”), which in turn limits outcomes? How might you limit that effect?
2. How do estimates figure into your approach to budgeting? Are there opportunities to fund initiatives incrementally based on outcomes (vs. upfront budgets based on an estimate and project plan)?
3. How do you make money? Do you make money by delivering things to customers “on time and to spec”, or something else? How does that impact your estimation needs — on both the value and “cost”/duration side? What would 20% “better” estimation get you? 20% “worse”? Define “better” and “worse”. What would you pay (in $) for 2x “better” estimates?
4. Is a focus on estimates unintentionally obscuring the need for smaller batches, more frequent integration, and more “continuous” value delivery? How so? How might you address that?
5. If estimating to plan “load” on a single shared individual/team, what is the net benefit of keeping that individual/team shared vs. hiring and embedding? Also, what is the context-switching cost for that shared individual/team as they try to to fulfill multiple projects? Are you estimating that?
6. At the lowest level, do you estimate things that will take hours/days, weeks, or months/quarters? What might “going up a level” (e.g. only estimate at the business initiative level using historical data) do to outcomes?
7. How much time are you investing in estimation relative to the value it delivers to your company and customers? Might lighter weight approaches work as well (or better)? How might you measure the relative “weight” of different approaches and their efficacy?
8. Do your customers directly benefit from your estimates? Would they pay you more for better estimates? How about indirect benefits?
9. Are you spending sufficient time estimating/forecasting expected value? Or are you focusing primarily on duration and estimated time to completion (“when will this be done”)? How do your initiatives (ideally stated in hard money) perform against your forecasts for expected value?
10. What is the return on investment for a typical product development effort? What does the range look like across all of your efforts? How is duration a factor in that calculation? How much value are you “moving through the system” (e.g. “we’re adding $2-$4 million in revenue monthly) to our ARR”). Have your “best ROI” decisions been a result of good duration estimates, good value estimates, or both?
11. What is the biggest current impediment to your organization’s “value throughput”? Is it estimation? In other words, is knowing something will take between nine and ten months currently more important than figuring out how to deliver value in three to four months? These things are not mutually exclusive (obviously), but consider how you are investing your continuous improvement energy.
12. Do you use historical data when forecasting cycle team and lead time? How so? If you were to split your team into 1) historical data only forecasters and 2) specific initiative forecasters … who would “win”? And by how much? Is the winning margin worth it? This might be a good experiment…
13. A common use for estimation is prioritization. Given your candidate initiatives/projects — and the distribution in the forecasted value of those initiatives — how “good” do your estimates need to be in order to zero in on the extremely high value opportunities (where duration is less important), and the “quick wins” of medium/high value (where being “small” matters).
14. Does your approach to estimation reflect the uncertainty inherent in your context? Are you communicating that uncertainty (e.g. confidence intervals)? Or are you accidentally conveying a false sense of certainty? If communicating a false sense of certainty, what is the adverse impact?
15. How open is your organization to altering estimates once new information becomes available? Is a change viewed as a positive (“we’re getting more information!”) or is it viewed as a deficiency (“I can’t believe we MISSED this!”).
16. Are you currently using other forcing functions to trigger the decomposition of work (e.g. fixed cadences, max increment length, story size, etc.)? Are those forcing functions successful?
17. How do delays financially impact your company? One month late? One quarter?
18. What factors influence duration? Might, for example, context switching and unplanned work explain some high percentage of duration? Or too much work in progress? If so, how would you know?
19. Are you punishing employees — either implicitly or explicitly — for “missing estimates”? On what grounds? How much “control” did they have over the outcome? How do you know? Do you punish employees for their value estimates as well? Why? Why not?
20. How might you “accentuate the positives” with regards to estimation in your organization, and dampen some of the potential abuses, confusion, malpractice?