by Freddy J. Nager, Founder of Atomic Tango LLC + Guy Who Had To Memorize Too Many Formulas…
“… fear of the unknown and our desire for certainty lead us to throw ourselves into the arms of perceived ‘experts.’ … We trust quantitatively flavored constructs to escort us away from the gloomy reality of unmeasurable uncertainty.” — Pablo Triana, Lecturing Birds on Flying
Quick quiz: What’s harder than trying to predict human behavior?
Quick answer: Trying to predict human behavior using mathematics alone.
Longer, sillier answer: Trying to predict human behavior using mathematics alone in a field in which you have no working experience.
That longer answer is the premise of Pablo Triana’s book, Lecturing Birds on Flying: Can Mathematical Theories Destroy the Markets? Triana’s field is finance, which he’s experienced as a derivatives trader and a university instructor. He criticizes “quantitative hodgepodge” — like the Black-Sholes-Merton model and the Gaussian copula model — churned out by “quants” (academicians, mathematicians and physicists). “Markets can’t be tamed with equations,” Triana writes. “Maverick, unlawful human action rules the markets, unexpected and unimaginable monstrous events shape the markets.”
I don’t know enough finance to verify Triana’s assertions, but a compelling article in Wired explains why the much vaunted Gaussian copula model failed miserably: “Recipe for Disaster: The Formula That Killed Wall Street.”
What I can talk about is marketing, which has also been infiltrated by mathematical delusions.
The Unholy Reign of Formula Marketing
As an MBA student, I took a so-called “Marketing Strategy” course hoping to learn how entrepreneurs could outwit and outmarket established megacorps. Instead, the professor — who, of course, had no real business experience — fed us one formula after another while ignoring creative branding, viral marketing (or any kind of media strategy), cross promotions/partnerships, alternative distribution models, or pretty much anything that uses the right side of the brain.
The gods forbid that a marketing strategist should actually have an imagination.
My classmates destined to become brand managers at P&G or Nestle found some of the exercises useful, since their jobs would entail crunching years of accumulated data to determine the price of non-dairy creamer in Des Moines. But for the future entrepreneurs and general managers in class, the useful lessons were few. And for the aspiring marketers looking to work in services and media (this being, after all, Los Angeles in the 21st century) the formulas were completely useless.
Take the Bass Model of Diffusion, for example:
Pretty, isn’t it? That’s supposed to tell you how a product will penetrate the marketplace, with the following elements:
St = number of adopters at time t
m = ultimate number of adopters
Yt = cumulative number of adopters to date
p = innovation coefficient
q = imitation coefficient
All those parentheses and sub-scripted letters lull users into a false sense of accuracy. But note: if just one number in that magical equation is off, your entire prediction is worthless. And what are the odds of one number being off? About 100%.
Indeed, with this formula, you literally need to mind your p’s and q’s: those two coefficients are based on the historic performance of similar products. Yes, the words “historic” and “similar” should be setting off alarms in your mind. And how do you determine which products are similar? By guessing, of course.
For example, let’s say you invent a high-definition DVD player and want to know how quickly it will penetrate the market — but, alas, there are no p’s and q’s yet for HD DVD players. No problem: just use the coefficients for the next closest product — oh, say, traditional DVD players in the 1990’s — and that should tell you EXACTLY when your HD DVD player will hit a certain level of market penetration today, right?
What’s that? Most consumers aren’t buying high-def DVD players because they don’t want to replace their standard DVD collections? And others are now more interested in downloading movies than buying more dust collectors for their book shelves? Those inglorious basterds! How dare human beings not follow the formula?!
Indeed, this Bass Model of Diffusion relies on an absurd number of assumptions:
- The economy never fluctuates. Oops, there’s a recession? Well, consumers don’t act differently during a recession now, do they?
- No competition. Blu-ray who? Hulu what? Not only does the absence of competition in this fantasy world mean there’s no one to undermine your plans, it also means there isn’t an advertising war to conversely bring attention to the category.
- No collaboration. Let’s conveniently ignore the role of video game consoles in promoting high-def DVD player adoption. We don’t need no stinkin’ relationships — we fly solo here!
- Similar products diffuse in the same way. Sure. That’s why Reeboks sell just as well as Nikes, and RC Cola sells just as well as Coke.
- All consumers behave the same way. Who needs market segmentation? Now, a professional marketer would identify favorable consumer segments or key individuals who could turn an unknown product into an overnight mass-market sensation — like, say, Oprah. But this is a formula, damn it, and that means all human beings are identical!
- Impact of price, place and promotion are negligible. Which do you think will penetrate the market faster: a $1000 product with no advertising that’s available only in high-end electronics stores, or the same product sold for $100 at Target and endorsed by that Oprah person? But why worry about any of that? Just use the formula and you’ll be a marketing genius!
Now, I know I’m gonna catch flack from my fellow academics who have devoted time, energy, and entire academic reputations to mastering tripe like the Bass model. They’ll say that such formulas exist only to explain how the “perfect” marketplace works in theory, and to provide a reference point on which marketers can craft larger strategies. But if that’s true, what’s up with the faux precision? If it’s just an illustrative estimate, shouldn’t the outcome of the formula be “likely to have slower than average diffusion unless you advertise the hell out of it” instead of 38.2918?
The Slippery Slope of Faux Precision
Faux precision cons actual businesses to use formulas to predict real results for real products in real markets. When the numbers don’t turn out as expected, analysts snicker, millions of dollars evaporate, and heads roll. Toshiba was the company that pushed the late great HD-DVD format, and I’m guessing that they had a lot of projections of how wonderfully profitable HD-DVD would be… only to see Blu-ray blow them out of the water when Warner Bros., Netflix, and other key participants stopped being team players.
Note also that the Bass model is for physical products. It cannot be applied to marketing a university, a social networking site, a new sitcom, an office building, a life insurance plan, red-eye flights to Argentina, or a gubernatorial candidate — you know, the products we American marketers are likely to promote here in century 21.0.
But the real world is beside the point. For the quants of academia, the prize is having your name attached to a widely used model that’s published in the Harvard Business Review, which in turn leads to tenure, publishing deals, and lucrative consulting gigs… Now that’s a killer business model! Want to get in on the action?
In my earlier post, “The Young Professor: How To Get Published,” I wrote a formula for formula fabrication:
Let’s quantify the saying, “Birds of a feather flock together.” First, create some variables:
CK = degree of avian Common Kinship
F = propensity to Flock
μ = a coefficient to qualify the theory and make it harder to judge
Then express the relationship as a formula: F =
Take the first derivative of this formula to find the maximum propensity to flock. This means absolutely nothing, but it sounds like you’re dropping mad science, you beautiful mind, you.
Naturally, when I came across Triana’s Lecturing Birds on Flying and a review describing it as “deliberately incendiary,” I felt a degree of avian Common Kinship, and my propensity to Flock kicked in.
Now I’m not saying that mathhas no place in marketing. To the contrary, marketers must quantitatively analyze the myriad choices they face. Who are your most profitable customer segments? Which media outlets will reach the most prospects at the lowest cost? If you’re advertising online, should you go by views or by clicks? What price will appeal to consumers, keep competitors at bay, and generate the profits you need? Can you afford a loss leader? Is distribution through Walmart worth cutting your price to the bone, or should you dis them for a Target exclusive?
I also endorse statistically valid customer surveys (not focus groups) and deep scrutiny of the numbers presented by media reps and distributors. Even so, you should provide equal weight to qualitative analyses based on experience and insight. How do your prospective customers really feel? How do you think your competitors will react? What’s happening in the real world that could warp your best laid plans? Is Paris Hilton really the best choice to endorse your product?
Qualitative thinking provides key sources of differentiation. After all, if everyone runs the same formulas and statistical studies on the same customers, won’t they get the same results? Differentiation comes when you say, “OK, those numbers are cute — but what will really make a difference?”
Above all, I discourage believing in any predictions of success. Such predictions might occasionally work for a few products, such as commodities that have years of accumulated data reflecting steady demand (Des Moines non-dairy creamer). But for a new product — particularly one driven by image or trends — don’t rely on a “quantitative hodgepodge”; that’s no more reliable than randomly asking people on the street, “Psst, how buddy, wanna buy a DVD player?”