Quite often on my current project, I hear conversations like, “That was estimated at three days, and it’s taken five. Why do you think that was?”
This is never asked with any sense of blame; rather, it’s a learning experience. What’s holding people up? Where are the bottlenecks? Was the story more difficult than expected? Was information unavailable? What could we do differently next time?
There’s another question which I hear less often. “That was estimated at five days, and it’s taken three. Why do you think that was?”
Strangely, this isn’t because we never deliver before the estimates. The estimates on our project are balanced, so even when they’re not realistic, there’s about the same amount of underestimation as overestimation. (If you’re always coming in under your estimates, your estimates are too large.) So why concentrate on the shortfall?
I’m not an expert at this, but I can guess. There’s a certain budget allocated to every project, be that time or money. Whenever a task takes longer than was expected, it’s now eating into the contingency (assuming there is any). The funders, who have allocated time and money to the project, want to know where their money has gone, whether it’s in good hands and how much more they might have to dish out. They’re asking uncomfortable questions of the managers, who in turn are asking uncomfortable questions of the devs. You have overspend, and must beg for more funds. How humiliating!
Conversely, when a task takes less time than expected, this adds to the contingency (or creates some, if there wasn’t any before!) Where’s the motivation to report that to the managers? They’ll only expect you to repeat the miracle. Where’s the motivation to report that to the funders? They’ll only take away your budget for next time! So there’s every reason to keep this secret, unless your funders are starting to fret. You have underspend, and until someone finds out, you get to keep it. How lovely!
I can’t think of any way to prevent that game from happening, but concentrating on the overspend produces some interesting results.
1) We game the system to come in under target.
No one likes uncomfortable questions. The estimate was wrong; why? Did my colleague make a mistake when he argued us down to 3 in the planning game? Or was I just slower than expected? Chances are that I’ll try to argue the estimate up next time, to avoid having to answer why.
2) We learn from our mistakes.
We didn’t make the estimate because we tried something out that didn’t work. We didn’t ask for help when we should have. I had a hangover all week. Let’s not do that next time.
3) We’re sorry.
That low-priority feature that you asked for isn’t going to be ready.
4) We’re not as good as we thought we were.
Now I’m sad. What’s the point of working so hard, when it’s so hard to work?
Concentrating on the overspend is at best a learning experience, and at worse demoralising. Let’s try something a bit different. What happens when we concentrate on the underspend, too?
1) We game the system for accuracy.
No one likes uncomfortable questions. Am I too cautious? Did I not believe my colleague when he tried to argue the estimate down? Or am I just faster than I expected? I’ve learnt something about our capacity, and chances are that next time, my estimate will be more realistic.
2) We learn from our success.
Henry cleaned up the code last week, and we were able to add the change in just 10 minutes. We found an excellent manual on writing SQL. Rebooting the cruise server on Monday stopped it crashing on Wednesday. Let’s do that again!
You know that story you thought you weren’t going to get? Well, it looks like we’ll have time to do that too.
4) We’re better than we thought we were!
Happy, happy, joy, joy! I believe in myself, and my abilities. I feel like I can take on anything! Even that bug that no one else wants to touch!
So, next time you’re looking at the overspend on a story, take a look at the underspend on other stories too. It will make you smile, you’ll have plenty to talk about in your retrospective, your estimates will be more accurate, and they might even get smaller – along with the real time to deliver.