Your Estimate Is a Wish

The date you gave was the best day you could picture

Someone asks how long it will take. You think it through. You see the steps. You see them going fine. You say, “Two weeks.”

That number did not come from the work. It came from the best version of the work you could imagine. Call it what it is. It is a wish with a date on it.

Then the supplier is late, a key person is out sick, the first fix does not hold, and the date slips. Everybody acts surprised. Nobody should be.

What the research says

Psychologists named this the planning fallacy decades ago. In one well-known study, students guessed how long their senior thesis would take. Their average guess was about 34 days. Their worst-case guess was about 49. They took about 56. Only about a third finished on time. Read that again. The real average was longer than the worst case they had imagined. That was students and a paper, not a plant, but anyone who has run a rollout has seen the same thing.

Why smart people keep doing it

Researchers call the story you tell yourself the inside view. You look at your project, your plan, your team. The inside view feels like knowledge because it is detailed. The detail is the trap. A detailed plan has no room in it for the things you cannot name yet.

The fix is the outside view. Ignore the story. Ask: how long did the last five jobs like this actually take? Planners call this reference class forecasting.

Picture a second-shift lead

This is an example, not a memory. Picture a second-shift lead asked to retrain the whole line on a new changeover sheet. They think about the sheet, the crew, the time. “Three weeks.” It sounds fair.

Now picture them pulling the last four rollouts like it. One took five weeks. One took seven. One got dropped in week four and picked up again after the holidays. One finished on time because nothing else happened that month. The outside view says six weeks, with a real chance of more.

They tell the plant manager six weeks. It feels like a loss. They are not being negative. They are the only person in the room who just told the truth about the schedule.

The cost is not the late date

The late date is annoying. The cost is what happens on the way there. When the real clock runs out, people cut the steps nobody sees. Teach-back gets skipped. The check-in after week one never happens. The change goes live half-built and the crew learns that dates are decoration.

A wish date also teaches your people something. If they know the date is fiction, they stop planning around it. Next time you say “two weeks,” they hear “maybe a month,” and they quietly adjust. Now nobody’s number means what it says.

One thing to do tomorrow

Before you give your next estimate, take ten minutes and do this.

1. Write down the last three jobs that looked like this one. Use the real start and finish dates, not what you remember.
2. Write down how long each one actually took, from first work to done and holding.
3. Take the middle one as your starting number. Do not take the fastest.
4. Ask one other person: “What always happens on jobs like this that we never put in the plan?” Add that time.
5. Give the date as a range, and say what would move it. “Five to seven weeks. It is five if the supplier ships on time and nobody gets pulled for coverage.”

If you have no record of the last three jobs, that is your first finding. Start one. A page with the job, the promised date and the real date is enough. In a year it will be the most honest thing in the building.

The objection

“If I give them the long number, they will say no, or they will hold me to the short one anyway.” Maybe. Then you are in a different conversation: what do we cut to hit the short date? That is a real conversation about real trade-offs. The wish date skips it and sends you the bill later.

An estimate built from the best day is a wish. Build it from the last three jobs and tell the truth about the range.

Every week the Briefs bring one problem like this from the floor and one thing to try on your next shift. Sign up at bigbeardedginger.com.

Sources: Buehler, Griffin and Ross (1994), Exploring the “planning fallacy”, Journal of Personality and Social Psychology 67(3), 366-381. Kahneman and Tversky (1979), Intuitive prediction: biases and corrective procedures. Kahneman, Thinking, Fast and Slow (2011), on the inside and outside view. Flyvbjerg, on reference class forecasting.