For most of August, I’ve been writing about topics that seemingly have little to do with marketing--sports for one, automobile cost of ownership for another. I suppose that I use August to clear my head the same way the French use August to clear Paris.
What traffic? It’s August.
However, in a spectacularly circumlocutory way, I’ve hit on a theme that matters dearly to marketers of all stripes: that data and data strategy mean two very different things.
Marketers don’t need another McKinsey report to understand that data management has become an important business-wide issue. Nor do they need anyone to tell them that managing data can get hairy. Rather, they need to understand that while compiling, storing, sorting and analyzing data take resources and expertise (read: money), those activities do not in and of themselves comprise a data strategy.
Of course, we cannot dismiss the care and feeding of data as a simple task. Merely maintaining extant data requires a great deal of hardware, software and know-how. To their credit, many companies have invested in their data infrastructure. Nevertheless, all this investment begs the question: why?
Data strategy involves knowing how to use your data, rather than simply trying to collect and sift as much data as possible. As such, marketers need to answer three basic questions to help them create a data strategy:
- What do you need to know about your audiences to communicate effectively?
From my days of building email preference centers, I remember the perennial challenge of telling clients that they simply couldn’t ask subscribers to fill out long lists of questions. When clients only needed to know the answers to a few questions, a drawn-out survey formed a burden to the subscribers.
The same principle applies to marketing data as a whole. Marketers need not know everything possible about their audiences, only the information that can help them develop communications. So don’t ask about family composition unless you have products that relate to kids. Don’t ask about home ownership unless you sell products that relate to the home. Rather, ask questions that help define your audience in terms of your product.
- What relevant distinctions exist among your audiences?
This question may, in fact, contradict the first one. While a brand may not sell child-specific products, it may notice key differences between how families with kids use the product vs. families without. For instance, a hotel may not cater to children, per se, but it can expect more traffic from families during school vacations and would probably want to use that information to target messages at the right time.
More to the point, marketers should think about key differences in their audiences and what data highlight those differences. A cosmetics brand might want to know age because older and younger women may buy different products. Or they might want to know about hair or skin type to help sell the right products. Marketers should have a good idea of how individual groups of their customers shop and buy.
- How do you know you’re right?
Perhaps measurement requires a strategy of its own, but marketers must take measurement into consideration when discussing data. Direct marketers generally have a leg up in this situation, because they can measure the ultimate result--sales--at the same level as communications. Similarly, loyalty marketers can generally measure sales on the member level and back out communication response using a little bit of simple math (e.g. did sales go up when TV spots were on air?).
However, even marketers who have no first-party relationships with customers, such as CPG marketers, should think about measurement. It gets trickier of course, but not impossible. For instance, let’s say a pasta sauce brand acquires consumers via a recipe email and website program. Obviously, it can measure interaction via the two communications channels (email opens/clicks, site visits, etc.), but it can also estimate consumption as well.
The brand could survey customers and ask them how many jars of sauce they buy each month. They could then use those numbers (plus more broad-based numbers of how much sauce households nationwide buy) to measure the correlation between communications with the brand and sales at the individual level. While caveats apply (notably the vagaries of what consumers remember), this approach should yield valuable directional data.
In a subsequent post, I’ll discuss what a data strategy document should look like. For now, however, think about whether your data can answer these questions. Feel free to suggest others in the comments!