Compare estimates with actual time once a month, by type of work, not by person. Look for patterns: which kinds of work run over, and why. The usual causes are review rounds, waiting and unclear briefs, not slow people. Use what you find to set better estimates, not to judge.
Why the gap is almost always there
People estimate the work they can picture: writing the draft, building the page. They forget the work they cannot: waiting for input, the second round of changes, the meeting about the changes. So most estimates come in low.
Step 1: Collect the numbers
For each finished task: the estimate, the actual time, and the type of work. A month of data is enough to start.
| Type of work | Estimated | Actual | Gap |
|---|---|---|---|
| Blog posts | 4 h | 6.5 h | +63% |
| Client reports | 2 h | 2.5 h | +25% |
| Landing pages | 10 h | 19 h | +90% |
Step 2: Look at types, not people
Group by type of work first. If every landing page runs 90% over, it is a planning problem, not a person problem.
Step 3: Find the causes
For the types with the biggest gaps, look at three or four tasks closely. Ask:
- How many review rounds did it take?
- How long did it wait for input or approval?
- Was the brief clear at the start?
- Did the scope change?
Often the actual work was close to the estimate. The extra time was rework and waiting.
Step 4: Fix the estimate or fix the cause
Two choices for each pattern:
- Fix the estimate. "Landing pages take about 19 hours with two rounds. Plan for that."
- Fix the cause. "Landing pages run over because the copy arrives late. Make copy a must-have before design starts."
Fixing the cause is better when you can.
Keep it blameless
- Talk about the work, not the person.
- Never use the numbers in performance reviews.
- Say what you learned, out loud, as a team.
If people fear the review, they will pad estimates or hide time. Then the data is useless.
Use ranges next time
For new or uncertain work, use three estimates: best, likely, worst. The PERT calculator turns them into an honest range.
Real time, by task, without a stopwatch culture
M-Tracker times each task as people work and shows estimate against actual, including time spent on work that came back. You can see which types of work run over and why. No screenshots and no activity tracking.
Where M-Tracker fits
M-Tracker gives every task a clear brief and every finished piece of work a first review, in plain words, before it reaches the manager. No screenshots, no scores.
Try M-Tracker free14-day trial. No credit card. No employee surveillance.
Questions people ask
How often should we review?
Monthly, for 20 minutes. Quarterly for bigger patterns.
What if someone always runs over?
Look at their tasks' briefs and review rounds first. Talk privately, with examples, only if the pattern is clearly about them.
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