Do not give one number. Give a range: the best case, the likely case, and the worst case. Say what you do not know yet. Break off the first small piece, do it, and re-estimate with what you learned. An honest range beats a confident guess every time.
"How long will it take?"
Your manager asks. You have never done this kind of work. You say "about two weeks," because it sounds reasonable and you need to say something. Five weeks later, everyone is unhappy, and you knew on day one that two weeks was a guess.
The problem was not that you guessed wrong. Everyone guesses wrong on new work. The problem was that you gave a guess dressed as a fact.
Step 1: Split it into parts you recognize
New work is rarely all new. A new kind of report still has research, writing, charts, and review. A new integration still has setup, testing, and documentation. List the parts. For each one, ask: have I done something like this before? How long did that take?
The parts you recognize get estimates from experience. The parts you do not get flagged.
Step 2: Give three numbers, not one
For the whole job, or for each part, write:
- Best case: everything goes smoothly.
- Likely case: normal bumps.
- Worst case: the things you are worried about actually happen.
"Best case one week. Likely two. Worst case four, if the data turns out to be messy." That sentence is more honest and more useful than "two weeks." It tells your manager where the risk is. The three-point estimate calculator turns the three numbers into one expected figure and shows how wide the range is.
Step 3: Say what you do not know
Name the unknowns. "I have not seen the data yet. If it is clean, this is a one-week job. If it needs cleaning, add two weeks." Now the unknown is visible, and someone might be able to resolve it today instead of in week three.
Step 4: Do the first small piece, then re-estimate
Pick the part you are least sure about. Do a small slice of it. One day, not one week. Then come back with a better estimate: "I looked at the data. It is messy. This is a three-week job, not two."
Nobody minds a revised estimate in week one. Everyone minds a surprise in week four.
Step 5: Write it down and track it
Save the estimate next to the task. When the work is done, compare. Not to blame anyone. To learn. After five or six pieces of new work, you will see your pattern. Most people underestimate by the same amount every time, and once you know your amount, you can correct for it.
How to say it to your manager
"I have not done this before, so here is my honest range. Best case one week, likely two, worst case four. The big unknown is the data quality. Give me two days to look at it and I will come back with a much better number."
That is a professional answer. It shows you have thought about it, it gives them something to plan with, and it sets up the conversation for the revised number.
Where M-Tracker fits
In M-Tracker, each task holds its estimate, and the timer records how long it actually took. Over time you can see your estimate-versus-actual pattern by type of work, by person, and by project, without anyone filling in a timesheet. New work gets easier to estimate because you can see what similar work really cost.
Try M-Tracker free14-day trial. No credit card. No employee surveillance.
Work you can trust, without redoing it.
Your team gets feedback before work reaches you, or your client.
Copyright © 2020 – 2026 M-Tracker. All Rights Reserved. Privacy Policy