James Lea, founder of Project Science, spoke at EVA 26 earlier this year. He talked about the psychology of estimating. “People,” he said, “are just as important as the techniques and data.”
He went on: “Plans and estimates are built by and used by people. Psychology matters.”
The talk was very interesting, and here’s what I took from it.
He started by asking us our experiences of estimating and the emotional responses we had at the time. Think about your own experience of estimating. Did you feel:
- Under pressure?
- Worried about how the numbers would be used?
- Unsupported by colleagues
- Unsure what the estimate was for?
That’s all (unfortunately) normal, and we all nodded along as he talked.
Challenge how will estimates be used
James talked about how we should challenge how estimates should be used. “Uncertainty drives variable reactions in our teams,” he said. “It drives emotions and responses.” If you are open about how estimates are going to be used and how they should be used, that can help people feel more comfortable with the process.
Make estimating positive
How can we enable our teams to experience planning and estimating as a positive, creative experience? Instead of the stressful, “I suppose I can give you a number,” experience that it is mostly?
It’s hard for an organisation to accept that it doesn’t know the answer, and that can sometimes lead to a poor experience of the estimating process for the people involved.
Here are some ways he suggested we could turn the experience into a positive one:
- Develop the models
- Turn the work of an individual into the work of a team with reviews
- Use the past as a guide to future performance
- Frame everything as a change
- Refine the estimates.
Creating a route to predict the future
James talked about asking the question about whether we have a route to predict whether the estimate is a robust one or not. We need to understand what is in and out of our control. Where things are out of our control, accept that and track it.
Estimates are only a guess without a map of how you got there and a set of viable routes.
We often hear that people can’t estimate where there is no historical data. Well, data science should make it easier now to estimate from past performance and the vast tracts of data we store about projects. If leaders can give teams the data, in a way that helps with estimating, that should make our estimates better.
Building defensible plans
James talked about showing your workings and documenting the bases of estimates. Steve Wake, the conference chair, shared his thoughts too, namely that the audit office regularly says people don’t know the basis of estimate and therefore the best ‘proof’ that your estimates are good is that you can justify them.
He talked about bounding your plans carefully, describing the world around the estimate as well as the estimate itself to provide rigour.
He suggested we quantify and compare with data science, applying risk appetite to the delivery methodology to round out what we know.
That, and the other points discussed, are ways to shape the emotional response and create a safe space for people to estimate their work.
What do you think? Let me know in the comments below.
