Optimise for visibility over predictability
From the Vis-à-vis-visibility Dept.
“Listening is being able to be changed by the other person.”
– Alan Alda
As an engineering manager, I have been in countless schedule planning meetings. Generally these are weekly and you are expected to have an expected end date for each of your current projects.
This usually takes the form of a big spreadsheet. These spreadsheets get more and more complex over time, as people add scheduling features like “original estimated date” or “developer hours per week allocated”.
The date is usually a single day, although sometimes you’ll be allowed to specify a month or quarter. The real world doesn’t give you dates - it gives you probability curves. Depending on the team, the particular day placed lovingly into the spreadsheet will be “best case”, “worst case” or “currently politically expedient case”.
There is a big paradox at the heart of this process - the leadership is looking for the outcomes of a solid and mature scheduling program, but isn’t willing to pay the costs associated with it. At the very least - a bespoke spreadsheet is no replacement for a dedicated project management software system.
As a team manager, you don’t have many good options here. Setting a date can leave you open to criticism when (not if) a project slips. Alternatively, you can strip functionality or push overtime to meet a release date.
My approach had these principles-
- Keep tasks sized to no bigger than 2-4 weeks.
- Regularly change dates, to show that track and adjust
- Prioritise visibility over accuracy
On this last point - I could have spent a lot of time attempting to improve my accuracy. Instead, I put effort into making our progress clearly documented and easy to access.
Some executives were only interested in the end date. But the better execs were curious about the accuracy of that end date, and we made it easy for them to self-serve the information needed for them to judge the confidence level.
It’s tough to beat the inertia involved with the simple human desire for “just give me a date for when it will be ready”. I’d suggest that organisations consider the “No Estimates” approach. It takes a lot of change management and culture work to move to this approach, but you end up spending less time on pretend-date-guessing and more time on delivery.
Further reading: