In the age of big data, we’re awash in data, technology to analyze it, and new channels to communicate through.
However, as McKinsey reports, the shortage everyone talks about is the lack of unicorn data scientists who can collect, clean, calculate data, and then communicate insights. According to McKinsey, the United States faces a shortage of 1.5 million managers and analysts with the skills to understand and make decisions based on the analysis of big data.
The need for skilled data scientists is particularly apparent in marketing departments, where the top two most sought-after specialized skills are digital marketing and analytical thinking, according to Korn Ferry’s 2014 Marketing Pulse Survey. But concern over this gap masks a more important problem: The biggest challenge keeping brands from capitalizing on big data to improve their marketing capabilities is the lack of an orientation to results.
In response to the Korn Ferry survey, 57% of CMOs cited an “inability to directly connect marketing efforts to tangible business outcomes” as the biggest contributing factor to low CMO tenure relative to other C-suite roles.
Brands may be tempted to bring in new talent to capture the potential of big data. But management teams would be better served redirecting existing resources to tackle three common challenges to ensure their marketing efforts contribute to overall business results.
Challenge 1: Poor focus. Too often, senior executives investing in marketing analytics teams and associated capabilities abdicate their leadership roles in suggesting where those capabilities should be focused. This is especially true when the investment has multiple sponsors. Their frequent inability to reconcile ways of looking at things and to develop joint hypotheses about where high stakes and uncertainty should drive analytic focus means they prematurely punt those questions to the analysts. And, in turn, the analytic centers of excellence they charter too often focus on demonstrating excellence through analytic depth rather than through providing frameworks to help pick the right problems.
Solution: To improve focus, use an “analytic brief” to facilitate your senior executives’ deliberations about what’s going on in the business and where analytic attention needs to go. Such an artifact goes target segment by target segment, examining how well marketing investments are mapped to perform for customer experience. Most importantly, it helps to parse for where it’s possible to take action--or at least experiment with it--to avoid having everything look like a nail for your big new Hadoop hammer. For example, as P&G plans to divest 60% of its brands in the next two years, it’s doubtless applying perspectives like these to re-prioritize analytic attention away from attracting new customers to these products.
Challenge 2: Over-analysis. If you put meat in front of a wolf, the wolf will eat it. If you put a complex problem in front of a highly sophisticated marketing analyst, she will not stop when the answer's good enough; she will eat all of it. For example, marketing data scientists often seek confidence levels and intervals that are beyond what are practically necessary for a conclusion, despite the higher costs, longer lead times, and diminishing returns of doing so.
Why? Sometimes because they want to exercise their skills. Sometimes because they feel they need to live up to their resumes to reinforce their street cred. And sometimes because they were trained that way, “In the High Church of the Normal Distribution,” as Scott McDonald, former SVP of research and insights at Condé Nast calls it. And if you outsource the analytic challenge, remember that the more complex the model, the bigger the bill it justifies.
Solution: To guard against over-analysis, proceed iteratively. At the strategy consulting firm Bain & Company, case teams I worked with adapted the hypothesis-driven problem solving approach the firm used by adding what we called an “answer slam.” The idea was to try to see if we could start by answering the questions posed above (“What’s going on?” “What can/should we do about it?”) in just a few minutes. Then, based on how satisfied we were with the answers, we would decide whether we should spend a day, and then a week, and so forth to refine the overall answer, or particular elements of it.
By doing this, we worked both questions to their middle. We developed an early sense of what causes, options, and costs might be. At the same time, we got a sense for where and how much uncertainty about our conclusions might lie and a sense for how hard it was going to be to resolve that uncertainty. Another tip: Thinking about how you would sustain attention to a worthy insight--for example, thinking beyond a mix modeling project to a platform for continuous mix adjustment--can help simplify your analytic ambitions.
Challenge 3: Under-delivery. Most marketing analytics projects take their scope from their name; they stop at the analytics part. For a number of reasons, too many marketing analytics teams function as service bureaus that passively take questions into their in-box and publish answers to the department out-box. They don't integrate well with the execution part, like making sure that so-whats are practical and that tests are instrumented to capture data necessary to validate any insights informing them.
Solution: To keep from taking too long, run your collection of analytic projects like a VC's portfolio. At the investment management firm T. Rowe Price, head of US client and market insights Paul Musante calls his portfolio a “learning agenda.” Says Musante, “It’s crucial that we allocate scarce resources not just for insight, but for where we can practically act as well.”
So review all your holdings every 90 days. Manage the portfolio via an expectation for "3-2-1." Nominally, each quarter the portfolio should generate three news-you-can-use insights, two tests based on those insights, and one initiative producing results at scale based on past insights and tests. If you're not generating these returns, dig into the mix of what you're looking at and examine your process and thresholds for moving things along.
Note that none of these mitigation techniques foreclose going deep and long. Rather, they help to build confidence that such efforts will be the right ones, they will keep them accountable, and most importantly, they’ll help to pay the freight with results as you go. So as you analyze your marketing strategy and goals for the year, think twice about turning to the head of HR and recruiting for all of your needs. Instead, get to the root of the problem by shifting your focus from the need for big data capabilities to the need for results.