Nearly everything in life, as well as in business, can be represented by mathematical models. Even if you’re uncomfortable with math, your brain still creates models for a variety of activities, and the decisions that you make are significantly influenced by this modeling behavior. We do this instinctively, since there were distinct evolutionary advantages for our ancestors who were able to correctly gauge the chances that food (a good thing) was just around the corner, or that a lion (a bad thing) was lying in the tall grass. Unfortunately, our capacity for correctly gauging the probability of the occurrence of good or bad things is skewed by the way our primate brains are wired, which in some cases diminishes our ability to understand and anticipate the likely outcome of certain actions.
Take, for example, the way that many people react to state sponsored public lotteries. The possibility of a good outcome is so seductive that many of us make an extremely poor bet, buying lottery tickets even though the odds of success are essentially zero. The advertising message “You can’t win if you don’t play” is truthful, but misleading. In actuality, the odds of success are so low that any expectation of doing so is completely irrational, but our ability to visualize the carefree life that a winning lottery ticket would create (and actual studies of lottery ticket winner suicides demonstrate that even this assumption is in error) overwhelms our common sense. This shows that our ability to conceptualize a good outcome actually interferes with our innate mathematical modeling, and we wind up exercising poor judgment repeatedly, and keep buying those lottery tickets.
Similarly, while the route to weight loss is extremely clear (move more, eat less), the obese population in America continuously embraces fad diets, because the vision of our potential selves as thin overwhelms the mental circuitry that recognizes that our prior experiences with fad diets have been unsuccessful.
Our ability to correctly gauge undesirable outcomes is also manifestly inaccurate when we desire to engage in activities that carry a high probability of doing us harm. Smokers, for example, smoke in spite of undeniable evidence that smoking is directly linked to a variety of health issues. Unmarried teenage girls repeatedly risk pregnancy by engaging in sex without birth control, even though they experience direct evidence (in the form of their friends who have become pregnant) of a high likelihood of an undesirable outcome.
I mention these examples because our inability to grasp the probability of likely outcomes can sometimes manifest itself in our business decisions. When I speak with retailers about advertising, they are nearly always interested in creating traffic, so that they can, in turn, drive revenues. But even when the methodology for creating that traffic is laid out before them, this incapacity to grasp the mathematics sometimes corrupts their reasoning.
As an example, allow me to outline a recent discussion with “Frank”, one of our customers who is participating in a test that we’re doing for Valentine’s Day. We know that Valentine’s Day is a low-end, last minute holiday, and actually, this knowledge is a good thing, because we’re using it in over 100 independent jewelry stores to initiate a dialogue with pre-bridal customers. In this way, we’re seeding their future bridal business, and while it’s hard to make money selling $99 items, the link to the subsequent engagement ring purchase makes the effort extremely worthwhile.
What we’re testing with Frank- and nine other Gems One customers - is something quite different. We are investigating whether we can attract affluent males and engagement ring customers (remember, February ranks second on the monthly engagement ring sales list) on Valentine’s Day with the offer of a Kindle Fire, or a pair of tickets to a major league sporting event, with a minimum purchase. In order to maximize the response, I’ve set the purchase threshold pretty low, at just $799. Taken out of context, this would mean a 25 percent discount on every $800 purchase, clearly a heavy price to pay for the traffic we hope to generate. And Frank, having just realized that he’s giving away a $200 Kindle Fire, instead of a $79 Kindle, had just made the connection.
The discussion started as follows: “George, how can I make money when I’m giving away half my profit?” Of course, Frank was correct... he’s unlikely to make any money if he’s giving away half his profit, and in this case, if we assume an 800 dollar purchase, at a 50 percent gross margin, then Frank has correctly built the mathematical model for what will occur. But are Frank’s assumptions accurate?
First, let’s consider what happens when a retailer offers a gift with purchase at a specific price. Yes, there will be a certain number of “bottom feeders” who will spend exactly the threshold amount to get the free item. This is particularly the case with very low price-point offers, like $25 off a purchase of $50 or more. Make that offer, and you’ll spend all day helping customers looking for $50 items. However, in our experience last November featuring a consumer electronics item with a purchase of $2500 for Black Friday, we found that the actual average purchase was just over $4200, meaning that the real purchase level was about 60 percent higher than the amount required. Higher ticket purchasers just behave in very different, predictable ways.
In fact, in looking at past g/w/p promotions, we find that the same percentages apply, so we can reasonably infer that at a threshold purchase offer of $799, we’re likely to have an actual average purchase of about $1300. In this case, the $200 gift-with-purchase represents an average discount of about 15 percent, a scenario that would be completely within Frank’s comfort zone. But just as the primate brain overestimates the chance of winning the lottery because it’s so easy to visualize being a winner, it underestimates the average purchase in this scenario because it’s so easy to visualize getting “bottom fed” to death at the minimum purchase level.
We also need to recognize that having targeted two specific groups with the advertising (in this case, affluent males and pre-bridal males), we’ve loaded the dice so that we’re much more likely to get much, much higher tickets. After all, Frank’s typical engagement ring sale is about $5,000, so it won’t take many of these sales to skew the average well above $1300. In addition, since the gift-with-purchase offer excludes discounts, and Frank typically discounts items about 20 percent off the tag price, it’s entirely possible that his real gross margin, after the give-away, will actually be higher than usual as a result of the promotion. Frank, of course, will remain concerned until the evening of February 14, at which point a review of the actual sales data is likely to show that everything has worked out just fine.
At the end of the conversation, Frank seemed at least a bit more comfortable. Plus, he told me he would be buying a lottery ticket on his way home that night, so when he won, all of these things really wouldn’t matter anyway. I smiled, and wished him good luck.