Anchoring in effort estimates

The first number sticks.

When someone says a change should surely be doable in two days, the discussion rarely starts from scratch afterwards. Two days becomes the anchor. From then on, people negotiate whether it might take three or four — even if an independent assessment would have produced an estimate of two weeks.

My own early estimates have the same effect. Once I’ve said a number out loud, revising it can feel like a personal failure. Instead of adjusting the estimate to reality, I start trying to squeeze reality into the estimate.

It helps to label an early number explicitly as a rough orientation and estimate again once the crucial uncertainties are known. An even better question is:

Which assumptions would have to be true for this estimate to be realistic?

That turns a supposedly objective number into a model that can be tested.