Misleading by the Numbers: 5 Traps to Avoid


How do I mislead thee? Let me count the ways.

If sales grew by 40% last year, but fell by 30% this year, how would you feel about the trend?

This comparison is typical of many dashboards and metrics reports and year-over-year financial comparisons. It would leave most of us feeling that sales were still up overall by about 10%. But before you leave the room feeling comfortably ahead, consider that this, in reality, means that sales have fallen 2% over the last two years.

The fact of the matter is that most business reporting is unwittingly designed to mislead. That’s certainly not the intent in the majority of cases. It’s more a matter of not knowing what we don’t know.

As you review your first quarter results with a thoughtful eye for the months ahead, use this checklist to make sure you’re not misled and that your decisions are on target:

  1. Two Plus Two Does Not Equal Four. People tend to treat percentages like whole numbers, making systematic and predictable errors in calculation. If your production statistics show efficiencies have reduced costs by 20% last quarter, and your production manager tells you they’ve found an additional 25%, chances are you’ll be looking for a 45% improvement in the next report. You may even be slightly disappointed by the impressive 40% gain.

This is the same phenomena responsible for the 10% growth versus 2% fall in the example above. When it comes to percentages you can’t always trust your gut, so make sure you’re doing the extra checking necessary to make sure you’re not misled.

  1. A 300% Rise, Or An Increase < ½ %? Did you see this week’s (04/23/08) headlines? “Home Foreclosures Quadruple in State, Bay Area” or, “Bay Area Home Foreclosures Soar Over 300%”. It’s doubtful that, “Bay Area Home Foreclosures Reach 0.04%” would be as effective in grabbing your attention, although it describes the same data points.

Most people tend to be good at simple percentages, but we’re not good at taking those same results and putting them into different perspectives. Even when we have the necessary knowledge at our fingertips, we often don’t make it to the next step. For example, if I tell you 40% of all sick days are taken on a Monday or Friday, would you be suspicious of long week-ends? (Monday and Friday represent 40% of the days in a typical work week.)

Which would you support?

Would you rather support research for a disease that affects 30,000 Americans a year or one that affects just .01 percent of the U.S. population? Charities, health researchers and public officials often use this principle to present their cause in ways that will motivate us to act. Rather than dispassionately informing us, they persuade by choosing presentations emphasizing their relevance, such as describing an increased risk of cancer in terms of percentage increase in likelihood you’ll have the disease (i.e., a 50% increase in your risk of cancer) versus the actual change in the incident rate (e.g., from 0.0050% to 0.0075%).

So even if the trends you’re seeing are accurate, are they presented in the most appropriate context? (Note: This is a significant issue with graphs, dashboards and other visual presentations, the subject of an upcoming Grey Matter article.)

  1. Anchoring Bias Away For Your Expectations. Did you know that just writing down your social security number is enough to bias you toward that number?

Bidding for Bias

In a study by Ariely of MIT, participants were asked to write down the last two digits of their social security number before being asked what they would pay for items. In a real auction at the end of the study, the half of the participants with social security numbers ending in higher digits paid 60% – 120% more than those with lower ending numbers.

This bias, called anchoring, occurs in all aspects of life, including when you’re negotiating, evaluating prices and reviewing your financial results. After reading in today’s paper that Starbucks says the economic environment is the weakest in company history, as a coffee shop owner you may be pleased if your sales are stable or growing in the low single digits. On the other hand, if you read that McDonald’s attributes much of its double-digit sales growth to its new line of coffee as direct competition to counter the Starbucks juggernaut, you may wonder instead why you’re not able to achieve the same kind of growth with your personalized service.

  1. Confirming Your Own Conclusions. Frank Abagnale, the real-life character in the Steven Spielberg/Tom Hanks/Leonardo di Caprio film, Catch Me If You Can, often tells about how he took advantage of the assumptions people make once they are given a cue to reach their own conclusions. Late one evening as the airport was closing, he noticed ticket counter staff bagging up the cash from their drawers, walking down to the end of a hall, and pulling open the drop-box at the end to deposit the day’s earnings.

The next evening when ticket counter staff walked down to the end of the hall, they saw Frank Abagnale in a security guard uniform with a garbage bag next to him and a sign taped to the drop-box door: DOOR BROKEN. PLEASE LEAVE BAGS WITH SECURITY GUARD. Seeing the sign, and Mr. Abagnale in his freshly-purchased and pressed security guard uniform, everyone dutifully dropped the day’s earnings into his bag. Mr. Abagnale tells of sweating bullets at the time, thinking, “Somebody’s going to figure this out. It’s a hole in the wall with a door over it. How could it possibly be broken?” But no one stopped to question how a drop-box door could really be broken. The consistency of the sign and the security guard uniform were enough to stop everyone from considering any other conclusion.

In the last decade, Kodak repeatedly made the same mistake, continually convincing themselves that their sales were about to take off, allowing them to reclaim their leadership position of an earlier era in the new digital photo market.

Recently General Electric made the same error, announcing to the market that they were still on track despite numerous threats to multiple core businesses in their portfolio. Just three weeks later, instead of earnings growth, GE reported a 6% drop in profits, somehow surprised by the effects of the real estate market and slowing economies. (This prompted former GE CEO Jack Welch to comment on CNBC that Jeff Immelt, his handpicked successor, “Has a credibility issue.”)

  1. The Risky Shift. How you evaluate your results will also depend upon who you’re with at the time you do it. When we are with a group, we tend to moderate our individual opinions to be more like those of the group, resulting in a group opinion that is more extreme than each individual would have reached on their own. Originally known as the Risky Shift, the affect of this tendency has been especially well-studied for its impact on jury decisions and the effects of prejudice. And in teen-agers hanging out with their friends. Beware of its effect on your judgment.

Despite all these biases, one thing we do know is that being aware of tendencies and biases can help us to be vigilant in mitigating their effects. So keep your eyes open, always be vigilant, and remember to be aware of information bias. You know… the tendency to want more information, even when it won’t affect your actions. Any questions?