Posts

Story points are a simple, relative way to size work. We don’t size purely in hours, instead we compare one story to another: is this smaller, about the same, or much larger? Pointing becomes more precise over time. The key is to establish a tight feedback loop. Early estimates are imprecise, and that’s OK. As we work through our iterations we continue to deliver, measure, and recalibrate our point sizing. And that quick adjust-and-learn cycle transforms vagueness into useful predictability. ...

For big tasks I use trivariate estimates - three numbers: best‑case, nominal‑case, and worst‑case. Each number is a confidence estimate: Best‑case (5%): only 5% of similar tasks finish this fast. Nominal‑case (50%): about half the tasks finish by this time. Worst‑case (95%): 95% confidence the task finishes by this time. For example: “I’m 5% sure we’ll finish in 1 week, 50% sure in 2 weeks, and 95% sure in 3 weeks.” That means out of 100 similar tasks, ~5 finish in 1 week, ~50 in 2 weeks, and ~95 in 3 weeks. ...

An estimate is a guess about how long the work will take - made without actually doing the work. The estimate needs to provide us some value, so it can’t just come out of thin air. But it also doesn’t need to be exactly right and take weeks and weeks to determine. There is a balance. We want that guess to be cheap (spending as little time as possible to make it), so we can get on with the actual work. And because of that estimates are intentionally imprecise. ...

As a builder you have the right to accept work, not have it forced on you. It’s OK to kindly refuse it if you don’t feel it’s appropriate for you and it might better for someone else to take on. But! When you accept a task, you also accept the responsibility that goes along with it. That means you own the quality and delivery, you keep your estimates up to date, you tell the team the status, and you ask for help when needed. ...

We all get to make and update our own estimates. We know the work best, so no one else should be guessing for us. “Estimates” are educated guesses - useful, but not concrete “promises”. They’re open to change as we learn new things. And when we do, we can change the estimate and let the team know. Good estimating means: We give our best guess, honestly. We use ranges when unsure (e.g., 3–7 days). We update estimates as facts change. We treat estimates as planning tools, not deadlines to blame people for. Discussions for your team ...