By Tim Girgenti
As a baseball-mad boy growing up in Houston, there were certain basics that I could count on seeing in the newspaper every summer morning. If I wanted to evaluate a hitter, I looked at batting average, RBIs and home runs. If I were looking at pitchers, I compared win-loss records, earned-run averages and maybe strikeout totals.
Today, in the age of sabermetrics—popularized in Michael Lewis’s bestseller Moneyball—savvy fans look with disdain at those antiquated stats. In a sport uniquely suited to in-depth statistical analyses, it’s all about “win shares,” “batting average on balls in play,” “on-base slugging percentage” and “value over replacement player.” In other words, these smart observers have looked at their own data and found that the business-as-usual metrics were insufficiently descriptive. And by applying data science to this growing mountain of data, they overturned conventional wisdom and gained competitive advantage.
It’s time to bring these same Moneyball principles to sales and marketing and, in so doing, prove or disprove the unchallenged assumptions and worn-out hunches that have governed marketing for too long.
There’s little need to repeat the stats we’ve all seen regarding the all-too-appropriately-named Big Data. Suffice to say that companies have become expert at capturing exponentially increasing piles of granular data. Applying it, however, has been another matter entirely. But lurking in that data is the ability to actually transform how businesses market and sell to their customers.
Too often, companies have relied on hunches or rudimentary statistics – the old-school baseball stats, if you will – to shape and execute their marketing campaigns. As a result, too many marketing campaigns are rooted in one-size-fits-all approaches or primitive segmentations with maybe three or five different “buckets” that haven’t been revisited for months or even years and that overlook the dynamic nature of today’s markets.
In today’s “moneyball marketing,” the keys are personalization and probability. That’s because, as the volume of Big Data grows, savvy marketers have an opportunity to create hundreds or even thousands of separate and distinct segments based on a far more granular understanding of buyers and their behavior. And when they do, they have a much higher probability of improving marketing effectiveness and driving greater revenue.
Previously it wouldn’t be uncommon to create lumped-together “segments”—if one could even call them that—consisting of only the barest of distinctions, such as “customers who have purchased in the past 12 months” or “prospects with more than $5 million in annual sales.” These segments might contain 20-30 percent of all sales prospects.
But data science, atop a growing pile of Big Data, changes the rules. Now we can create better segmentation and achieve better results. Imagine, instead, the ability to sift through exponentially greater data to populate 1,000 different segments that each contains a fraction of one percent of the target market. These segments can use dozens of different attributes and characteristics and combine them with dozens or hundreds of distinct offers, opportunities, promotions and prices. It can be as specific as: “customers who have purchased any of our five newest SKUs within the past six months, who have also called our help line, submitted a rebate claim, displayed a positive reading in social-media sentiment measures, are in the healthcare industry, and have revenues between $10-20 million.”
Suddenly, we have a much clearer picture of the audience we can use to create much more efficient, micro-targeted and micro-segmented marketing campaigns. Based on these demographics and attributes, we can personalize marketing messages, tailoring them to the needs and preferences of tiny subsets of buyers. And then we can test offers we present and the messages we use to continually refine our sales communications strategies, and achieve outsized response rates and revenue. That means marketing is always on top of changes in the market and continually delivering precise messaging.
Probability: “How Likely,” not “How Many”
The question then becomes: which segments should be the highest priority and receive the lion’s share of scarce marketing resources? Naturally, in virtually all instances, you’ll want to devote the bulk of your resources to opportunities that have the greatest chance of bearing fruit. Like a savvy baseball manager, you can’t guarantee success, but you can make all the right choices to put the odds in your favor.
Big Data – and the underlying data science – presents an even greater opportunity to stack the odds in the marketer’s favor by showing us the segments that have the highest likelihood of completing a sales transaction in the shortest period of time. That’s very high-value information for any sales/marketing organization. Note, this isn’t about “how many” prospects are in a segment or “how many” units they might buy – it’s about “how likely” those prospects are to buy, right now.
With the right algorithms and analyses, we can do more than segment and personalize. We can make smart, valid predictions. And, in so doing, we can often overturn the false hunches and assumptions that marketers have erroneously relied on for years. We can prove – or even more valuably disprove – the conventional wisdom. Those “logical” four or five segments that you’ve thought were more likely to buy often underperform the optimized segments that we identify through data science.
Personalization and probability go hand-in-hand to drive unprecedented gains in sales and marketing. Instead of one or two attributes to define four or five segments, we can use 10-12 attributes to create tens of thousands of segments. By subsequently applying data science, we can match optimized and personalized offers to the segments that are most likely to purchase sooner.
Finally, we can say goodbye to the old-school stats and successfully challenge conventional wisdom.
Tim Girgenti is CMO at PROS Inc.