Finding the ‘Flaw’ in Projects
The next time your boss asks you for a number, a deadline date or another fixed value, remember anything you say will be wrong. The reason is ‘the flaw of averages.’
Discussed in a book of the same name by Sam L. Savage, the flaw of averages basically explains why everything is behind schedule, beyond budget and below projection.
For example, you’re developing two software modules. Both are expected to take between eight and 12 weeks to complete. When both modules are finished, your organization can start a new process, which requires four additional staff.
Your boss asks you for a completion date so the additional people can be brought ‘on-board’ and the new profitable line of business started. You say, “Eight to 12 weeks,” and your boss replies, “Give me a date!” You estimate that a safe date would be 10 weeks, the average of eight and 12 weeks.
Everyone is happy. But should they be? Let’s look at the flaw of the average:
Based on your projected date, your boss works out his profit forecast for the next quarter based on an estimated profit of US$ 1,000 per week. You take the first seven weeks developing the application, and the new team uses the application for the remaining five weeks.
This sounds sensible, but the estimate of US$ 5,000 profit is the best that can be achieved. If you finish early, there is no upside. The new people will not be available.
If you finish late, however, sales will be lost. The cost of the unproductive new staff will be an added expense until both modules are working. On average, the profits achieved are likely to be significantly less than US$ 5,000.
Plus, you’re more likely to be late than early. The probability of finishing each of the modules in 10 weeks or less is 50 percent. It’s like tossing a coin – there is always a 50 percent chance of it landing on ‘heads.’
Since two modules need to be finished in 10 weeks or less, think of the options when you toss two coins:
Tails + Tails
Tails + Heads
Heads + Tails
Heads + Heads
There is only a one in four chance of you achieving the ‘average.’ That 25 percent probability means there is a 75 percent probability of being late. Therefore, on average, you can expect to be late.
All you did was assess a reasonable number based on your expected average time to complete each module. The problem is not your estimate, but the way it is being used. This is the flaw of average.
The next time you are asked for a ‘number,’ use your skills in managing upward to build flexibility into the conversation.
For example, offer your boss a safe date with an option for improvement. “We will definitely be finished in 12 weeks, and there is a possibility of finishing sooner.” Point out the cost risks of being early and late. Keep the boss updated as you work through the project.
Or, do some serious analysis. Offer your boss a range of dates with different levels of probability. You need tools for this, but you want a target date with an 80 percent probability of achieving.
Effective stakeholder management needs more than simply complying with a request — however reasonable it may appear on the surface.