Estimates are guesses.
How good a guess is affected by a number of factors including:
- How similar the context of the current work is to past work
- How resilient the nature of the work is to unexpected changes
- How many similar past instances with accurate data are available
- How recently these past work instances were completed
- How experienced are the people providing the estimates
- How many people are providing estimates for the same work
- What has been their recent track record for providing accurate estimates
Even when we are on the preferred side of these factors we shouldn’t forget that there is always the chance that an activity which you had thought would be fairly predictable could surprise you.
My car had to be taken in for a minor repair. I had discussed the work with service technicians before driving to the service center and they had informed me it would take between one and one and a half hours to complete. The service center is an hour’s drive from my house and I had booked the first appointment time slot of the day.
If we evaluate this activity against the factors I’d listed at the beginning, there were no red flags. My car is a few years old and there had been enough reports of the issue I was getting resolved on the Internet that I was sure the service center performed similar repairs on a regular basis. The work was being done at the center, so the availability of technicians, spare parts, equipment and other dependencies should not be a source of significant risk. I had also received similar estimates from two different technicians and both seemed very experienced.
Given this, my expectations were set that I would make it home by noon to enjoy lunch at my usual time.
When I arrived at the service center all seemed well. My car was taken to the service area within a few minutes of my checking in, and the service advisor told me that they had proactively ordered a part which might have been needed for the repair so I wouldn’t need to worry about parts being out of stock.
An hour passed and the service advisor came by the customer lounge areas and told me that the work was progressing well and my car would be ready “soon”. Forty-five minutes later I asked for an update and was informed that they were still working on it. Fifteen minutes later when I asked for an update I was informed it would be another forty-five minutes. Finally, after just over an hour, the car was ready.
End to end, the process had taken just over three hours.
Remember the upper end of the original estimate range was an hour and a half. Doubling the original estimate might happen when we are dealing with a rough order of magnitude but that was not how I interpreted the estimate I’d been given, especially for such a routine repair.
I was never informed of the cause of the delay. As there was limited visibility into the work area, I couldn’t tell if the technicians spent the entire time working on my car or whether they were multitasking the repairs to multiple vehicles. The service advisor also didn’t inform me of any surprises they’d encountered which might have accounted for the variance.
This did reinforce the importance of asking folks for their level of confidence in the estimates they’ve provided and the wisdom of building buffers in regardless of predictable the work appears to be.
(If you liked this article, why not read my book Easy in Theory, Difficult in Practice which contains 100 other lessons on project leadership? It’s available on Amazon.com and on Amazon.ca as well as a number of other online book stores)
Easy in theory, difficult in practice 