Ask anyone how long a job will take and you get the time it takes when everything goes right. That is not a lie, it is how estimation works, and it is why schedules slip in the same direction every time.
The useful thing is that the error is consistent. Consistent errors can be corrected.
What gets left out
Estimates cover the work. They almost never cover:
- Travel and parking
- Setup and packing down
- The conversation with the customer
- Finding something that was not as described
- Collecting a part nobody checked for
- Waiting — for access, for a decision, for someone to move a car
- The last ten percent, which reliably takes a third of the time
- Cleanup
- The paperwork afterward
None of those feel like the job. All of them consume the day.
Measure the gap
For one month, record two numbers per job: what you estimated, and what it actually took door to door.
Then calculate the ratio. Most small businesses find something consistent — actuals running a quarter to a half longer than estimates, fairly reliably.
That ratio is now a correction factor. Estimate as you always have, multiply, and your scheduling gets dramatically more accurate without anyone becoming a better estimator.
This is the single highest-value measurement exercise in a service business and it takes a month of writing down two numbers.
Estimate by type, not by feel
Once you have a month of actuals, group them by job type. You will find that certain categories are consistently underestimated more than others.
From then on, estimate from the historical average for that job type rather than from scratch. That is faster, more accurate, and it does not depend on who is doing the estimating — which means someone other than you can do it. See our note on quoting faster.
The three-point trick
For anything unfamiliar, ask for three numbers instead of one:
- Best case, if everything goes right
- Most likely
- Worst case, if the predictable things go wrong
Then weight toward the middle with real respect for the worst case. The point is not the arithmetic — it is that asking for a worst case forces people to actually think about what could go wrong, which single-point estimates never do.
Why it matters beyond scheduling
Pricing. If you price from estimated hours and actual hours run 30 percent higher, your margin is 30 percent smaller than you think. Some job types may be losing money and looking fine.
Capacity. Planning against optimistic durations means systematically overcommitting. See our note on capacity planning.
Customer trust. Telling someone two hours and taking four damages the relationship even though the work was good. Telling them four and taking three is a small gift.
Team morale. Perpetually running late is exhausting and it is not their fault. It is the estimate.
Quote the honest number
The instinct is to quote the optimistic duration because it sounds better and wins work.
It wins the job and loses the goodwill, every time. A customer who was told the truth and got it is happier than one who was told a nicer number and did not.
Padding is not dishonest. It is including the parts of the job that are genuinely part of the job.
Keep it going
Once you have the correction factor, do not stop measuring. Job mix changes, teams change, and the factor drifts.
A rolling look at estimated versus actual, reviewed monthly, keeps it honest. It is one column on the scorecard.
The uncomfortable finding
Occasionally this exercise reveals that a whole category of work is unprofitable at current prices — the estimate was wrong, the price was based on it, and the job has been losing money for years while looking busy.
That is a hard month. It is also the most valuable thing measurement will ever tell you, and it is invisible until somebody writes down two numbers per job for four weeks.