Bookending the Outcome

by Edward Henkler on February 25, 2014

Chip & Dan Heath suggest the concept of bookending in their marvelous book entitled “Decisive”.  Bookending entails separately choosing a lowest and highest expected value for an object or activity, then a most likely value within that range.  This moves away from so many of the dangers inherent in predicting the future.  One of my favorite examples involves the typical approach I saw used in estimating the cost of a major capital project, such as constructing a new laboratory facility.  The estimators spent a lot of time employing a robust data set based on historical performance to estimate the cost of the new facility.  It was always my contention that if the forecast was met, it either meant that something was skipped if the budget was too tight, or we invested surplus funds in unnecessary add-ons.


Were the engineers dishonest or incompetent?  Neither characterization would be correct.  These were very smart individuals using excellent information.  The problem is that the future can be very hard to predict.  A steel shortage, building boom, or dearth of projects can all cause construction estimates to vary by 30% or more.  A severe winter or series of summer storms can delay a project by days, weeks, or even months.  If key decision makers don’t delegate their authority, purchase approvals can be delayed.  The bottom line is that intelligently and accurately developed estimates can vary dramatically based on circumstances beyond the control of the estimator and project leader.

Rather than saying the new facility will cost $125,000,000 +/- 10%, it would be much more realistic to say the new facility could cost as little as $90,000,000 if all factors are favorable and a reputable firm is desperate for work and underbids.  It could also cost as much as $170,000,000 if all factors are unfavorable.  The estimate would conclude by noting that the expected value is $125,000,000 +/- 10% (or even +/- 5%).  This approach changes the subsequent dialogue when costs are trending high or low beyond the control of the project team.

Please note that if you build10 similar lab facilities over a period of years, the cost will likely average out near that $125,000,000 estimate but individual projects may vary dramatically.

The next time you are asked to estimate how much, how long, etc., try inserting the range concept with well-developed rationale for the low and high bookend.

{ 0 comments… add one now }

Leave a Comment

{ 1 trackback }

Previous post:

Next post: