Become an Estimation Leader
Several years ago, I had a senior VP as a client. He wanted his managers to work as more of a team. He thought they were not working together for the good of the company.
I asked him why he thought that. “When they estimate, they don’t work together on an estimate. They each estimate their own parts. I have to add the estimate together.” He paused for a second. “But I found a way to get a better estimate. I add everything together and lop 20 percent off. They have to scramble to make the date.” He laughed, and it sounded like a cackle to me.
I spoke to the managers. Each one of them said, “I provided a reasonable estimate the first time. In fact, we all worked together to provide a reasonable estimate. Then he took 20 percent off our estimate. We decided to stop working together to estimate—and to pad our estimates by 50 percent.”
That’s one kind of estimation leadership. It saves the teams from having to work under significant pressure, but it’s not honest.
What if you could provide a different estimation leadership?
I recommend you avoid providing a single-point estimate. Single-point estimates, whether they are for a budget or a schedule, are never correct. You start a project, thinking you’ll be done in six months. Things happen. Some parts of the project you finish faster, and some parts proceed slower. That’s the way things go in software projects.
You can provide a different estimate, not a single-point estimate. You can provide the “first possible, likely, and pessimistic” dates (or budget). You can spiral in on a date. You can work to a target date. You can provide a percent confidence for given dates. Each of these options is more accurate than a single-point estimate. Even better, they show that you take the deadline seriously and you want to provide the best information possible.
When I worked with that client long ago, we decided the senior VP was not going to listen to a three-date (possible, likely, pessimistic) estimate. He would take the first possible date and not consider how risky the project was. Instead, the managers got together and presented him these options:
- Tell us your date and provide a ranked backlog, and we will work the best way we know how to help the teams make that date. This is working to a target.
- We’ll tell you our confidence level for any date once we start the project. We’ll keep working until you want us to stop.
- We’ll provide you our target quarter and continue to spiral in on a date as we proceed.
The VP was a little disappointed. He told me, “I can no longer take 20 percent off their estimate. How will I hold their feet to the fire?”
I asked him, “Do you want to hold their feet to the fire, or do you want completed projects that work?”
He sat back and grinned. “Both. But that’s just me being petty about the feet, isn’t it?” We both laughed.
You don’t have to be a manager to use these approaches. When you provide estimates that explain your risk management, you are a leader.