I used to give estimates the way most engineers do, as an honest technical guess with a fog of uncertainty around it in my own head. "Probably two weeks." What I meant was somewhere between one and four weeks depending on what we hit. What the business heard was a firm commitment they could plan around, tell customers about, and hold me to. The gap between what I meant and what they heard caused more friction over the years than any actual technical failure ever did.
The uncomfortable realization is that an estimate is not a private technical opinion. The moment you say it out loud to someone who runs the business, it becomes a plan. They will build a launch date on it, promise a client, staff a marketing push. Your uncertainty does not travel with the number. The fog in your head stays in your head, and only the confident-sounding figure crosses the table. So the professional skill is not estimating more accurately, which is mostly impossible for real work. It is communicating the uncertainty along with the number, in a form the business can actually use.
Now I never give a bare figure. I give a range with the reasons attached. "If the third-party API behaves the way the docs claim, about two weeks. If it does not, and integrations like this often surprise us, closer to four, and I will know which world we are in within three days." That last part matters most. I tell them when the uncertainty resolves, so they know when to expect a firmer answer instead of pressing me for false confidence now.
This is not hedging to protect myself. Padding every estimate into meaninglessness is its own way of losing trust, and people see through it fast. It is giving the business honest inputs so they can make honest decisions. A leader who knows a date is genuinely uncertain can plan for the uncertainty. A leader who was handed false precision cannot, and will be blindsided, and will remember that you were the one who handed it to them.
The engineers the business learns to trust are not the ones who are always right about dates. Nobody is. They are the ones whose estimates carry their uncertainty honestly, so that "two weeks" stops being a promise waiting to be broken.
– Serguey Shinder
