- Framing Your Focus Beyond the Answer Itself
- Three Perspectives for Marketing and Sales Analytics
- Applying the Three Approaches
- Reconciling Organizational and Cultural Perspectives
- Packaging for Balance: "The Analytic Brief"
- Visions for the Analytics Capability to Serve These Needs
Three Perspectives for Marketing and Sales Analytics
The conversations for this book suggest three main orientations executives have for marketing and sales analytics. For short, I call them “Venus,” “Mars,” and “Earth.” “Venus” is primarily a way to tell “outside-in” customer experience stories about how people connect to your firm, and how your sales and marketing efforts map and execute against the contexts these customer experiences define. By contrast, “Mars” is an “inside-out” numbers-first way of thinking about how you should shift marketing investments across different channels. Today, “Attribution Analysis” (the Big Data Brother of media mix modeling) is the most well known specific technique for this angle on marketing thinking. Finally, “Earth” describes an “infrastructure-up” order for building properties and presences that is essentially reasoned from a blank slate (“Ok, first we need a web site, then we need a mobile version of it, then we need to be on Twitter...”). The sequence the Earth perspective suggests is generically sensible, but typically untethered from an appraisal of how your specific customers want to navigate toward purchases and of the relative performance of your current investments to help them.
Different senior executives have, through their backgrounds and inclinations, a dominant framework among these three that influences them. Creatives like to tell stories. Analysts like to talk numbers (and with the glorification of quants on Wall Street, marketing analysts now also talk in terms of “arbitrages” and “trades” as well). Builders express themselves in terms of “foundational capabilities” with “end-state” visions, phased plans, and “process designs” for managing them. Because of these biases, senior teams with similar backgrounds run the risk of groupthink and missed opportunities, while other teams with widely differing backgrounds run the risk of confusion and paralysis. So in addition to the intrinsic value each framework offers on its own, using them collectively is very helpful so team members can tune into each other’s points of view.
Various analytic disciplines apply the same fundamental Venus perspective using different terms. Strategists talk about “customer segments,” the “pathways” each customer travels, through a grid defined by “buying process” stages on one axis and “channels” on the other. Business (or “functional”) analysts talk in terms of “users” and “use cases,” also frequently documented using process maps or requirements documents for technical solutions. Creatives talk in terms of “personas,” “occasions,” and “touchpoints,” bundled into “narratives.” Regardless, Venusians like stories, and so the common denominator for them, regardless of functional training and experience, is the largely “left-to-right” narrative arc of the “customer experience.”
If numbers are introduced into the resulting Venusian stories, they are expressed as costs-per-something (for example, per impression, click, lead, or acquisition) at each step, or as a percentage yield from one step to the next (for example, conversion rate). But very often they are not included at all—a crucial miss. Even more rarely are they updated. A more common outcome is a beautiful flowchart, capturing the existing or desired experience in amber, displayed in a large framed print on an office wall.
Martians like data. So they take what various channels produce and look for significant answers to the question, “When I change x (and y and z), what happens to a (and b and c).” Then they divide costs by these results working toward the holy grail—a common currency for “pricing” investments for different options (“full attribution”). Next, they use these “prices” to “buy” and “sell” media, messages, and audiences (at the extreme, this could be in real time, such as real-time bidding in display ad markets).
Today, there’s lots of data. And, the relationship among different data is increasingly deterministic (I see search result X and click through to page Y) rather than probabilistic (I see BigCo’s TV ad X, and later visit BigCo’s store Y). Further, the tools for capturing and making sense of all this (cloud computing, Hadoop and its friends, modeling and visualization packages, and so on) are widely available and more cost effective than anything to this point.
On the other hand, data volumes can be overwhelming, the data itself can be very messy and hard to integrate, and with so many moving parts, analyzing and acting on it can be pretty complex. Skills to set up, use, and interpret the available and affordable tools are scarce. Plus, there’s the classic joke: “Q: Why does the drunk man look for his lost keys under the lamp post, instead of the bushes where he fell? A: Because that’s where the light is!” In other words, by starting with data we do have about the things we are doing for the people we do see, we miss opportunities to do things we don’t yet do, for people we don’t yet know.
Among the three perspectives, Martians are the most likely to test as well. Because of their primary orientation to data and not to design, they don’t fall in love with their creations and confuse these means for ends. On the other hand, their relatively lower engagement with design of infrastructures and experiences sometimes limits their orientation to data-rich and test-friendly channels, recapitulating the “drunken man” dynamic described above. Smart executives see this bias, and while they respect and embrace the discipline of an analytic, test-oriented mindset, they don’t allow it to dominate their options. Speaking of La-Z-Boy’s analytically-inclined culture, CMO Doug Collier says, “It’s almost a joke...‘If you don’t bring me a control group, don’t bother talking to me!’” He explains the firm’s insistence on a “double-delta”: “Don’t just tell me how we compare with past performance—I also need to understand the lift over control.” So now, “Even the PR and brand guys will come in with numbers to support their suggestions.” Yet at the same time, speaking of the firm’s significant investment in its brand awareness campaign through a relationship with the actress and model Brooke Shields, “Signing Brooke wasn’t something you could do as a test. We had to go, and we felt combining Brooke’s appeal with our prior brand equity would lead our target customers to give us a chance.” However, speaking to the data-driven discipline balance he tries to strike, he added, “On the other hand, we didn’t just jump; we did a lot of careful qualitative and quantitative research on our brand platform to shape our advertising. We didn’t want to make Brooke the brand, but rather have her be the vehicle through which our target customers could discover that our product line fits their lifestyle, and that our stores are great places to shop.” The firm has done seventeen spots with Ms. Shields so far, and has the analytic infrastructure in place (including the use of extensive quantitative and qualitative testing of the spots in primary research, and the YouGov BrandIndex service extended with some custom questions focused on the awareness and effect of the brand platform) to understand, scientifically, which spots, in what combinations, are effectively conveying which messages at different times.
People of Earth are more architects and engineers than operators. In the absence of good Venusian stories and Martian models to guide their paths, they document the “current state,” envision an “end state,” and then plot and launch phased programs for getting from A to B. But if you look more closely, very often the starting and ending points are described more in terms of capability inputs than performance outputs and milestones for deciding when what you’ve built is good enough and it’s time to move to the next part of your plan. Or, if implementation of capabilities is tied to related operating metrics and not just schedules, it’s sometimes hard to make investment trade-offs without the kind of common currency that Martians might prefer.2
The “builder bias”—the tendency to use this lens—is especially prevalent in digital marketing, because many of the senior people in this domain came up at a time when their priority was to field basic capabilities, rather than optimize pre-existing ones. Rob Schmults, VP of Online Commerce at Talbots, says:
- If you asked a physical store manager, “How’s business?” you’d never hear, “Well, this month we installed new lighting, and the talking price tags go live next month.” They’d be less about pixels, and more about stockout rates in popular sizes. By contrast, ask an ecommerce person, “What’s on deck for next year?” and he or she tends to focus on “Design and launch site x or feature y” and relatively less on the performance of what they are operating. This means time and time again you see insufficient attention given to inventory levels, to basic usability and shopability. A physical store manager is going to raise bloody murder if she’s hollowed out on core sizes in core product. She’s also going to be absolutely certain her store is easy to get around. The online counterpart all too often gets distracted by shiny keys.
At minimum, even with weak ties to operating metrics such as feedback loops, one thing our executives suggested looking for, to distinguish more- and less-effective applications of the “Earth” perspective, is how iterative these applications are. In other words, ideally they try to keep implementation progress in balance across your “experience”: make a little progress on the website, then shift to promotion, then back to your mobile site, then perhaps some attention to social media presence, and so on. This requires, of course, some coordinating mechanism and a strong team culture in which senior folks responsible for different pieces know when to “lead, follow, or get out of the way.”