Estimates are awful; Estimating is great!
This thread definitely resonates with me. If you know 100% of the requirements up front, you could make a reasonably accurate estimate of total cost.
The problem is: you don’t!
In my experience, estimates stop being useful as soon as they escape the team. Even further: story point/t-shirt size estimates are near-valueless as soon as they’ve been created.
The value comes from the process of estimating itself, not the estimates.
It’s a Team Effort
By having everyone on the team participate, I’ve seen very useful discussions take place that wouldn’t have otherwise happened. To make an estimate, you need to feel you understand the work, which can result in folks asking excellent questions.
The best part is when folks on the team have significantly different estimates for a given piece of work.
"What do you know that I don’t?"
It’s a great opportunity for knowledge sharing, and often surfaces key details or important clarifications.
I’ve seen these discussions be incredibly valuable, but it’s important to remember that the output is still "numbers we made up".
Further: to estimate enough work to be useful outside the team requires the team do considerably more estimation than is useful inside the team.
The team is necessarily operating at a different level of abstraction than outside stakeholders. Estimates that are useful for one level are incomplete, and likely misleading when considered at another. To make things even worse: the bigger the thing we’re estimating, the bigger the margin of error.
But does it scale?
So: the kinds of estimation practices that are useful to individual teams don’t seem to "scale" across an organization. However, higher level technical estimation seems important to organizational decision making.
How can we think about ROI without some sense of the investment?
I think @RonJeffries is right about starting with a discussion of a defined investment level. Not "what will it cost", but "what are we willing to bet".
From there: build trust through sustained incremental improvement.
The more trust you have, the less detail stakeholders will ask for.
I am interested in how engineers should engage in higher level organization planning. What kinds of information should we bring to those investment discussions?
How can we best share the knowledge and perspective we have with decision makers?
Note: This post started as a twitter thread! Come say hi, I’d love to chat about it!