The Reclamation of Strategy
Marketers can use big data to improve their strategic efforts.
As has come evident, big data is one of the most important investments a brand can make. The stakes are huge — consulting and technology firm Capgemini has suggested that data can improve performance by 41 percent over a three-year period. Given the potential returns, what board would hesitate to sign off on a significant investment in big data?
But using big data to improve a business means more than just collecting the information, or even analyzing it — companies must develop a strategy for how to use the information to build their brands. Unfortunately, many firms are using big data tactically, rather than strategically. Marketers, in particular, are not realizing the full potential of big data —they’re mainly using it to drive programmatic advertising.
Despite some success stories, the jury is still out on whether this method is effective. A clutch of recent studies has questioned the ROI brands are getting. Gartner, for instance, forecast that 60 percent of big data projects globally through 2017 will fail to go beyond the pilot and experimentation stage, and will be abandoned.
So what is a brand to do? To get real ROI from big data, marketers need to reclaim their strategic heritage and use big data to understand their markets in fundamentally new ways. Here’s what that might look like:
To get real ROI from big data, marketers need to reclaim their strategic heritage.
At the individual level: Social scientists have long been aware that different psychological attributes, such as our personalities, influence our purchasing decisions. However, this has largely been met with a shrug of the shoulders by marketers, not least because these factors can be hard to measure. Marketers have instead preferred attitudinal data, which typically has a more direct (and easier to understand) relevance to consumer activities that marketers can influence.
But big data not only tells us what customers do but also how they think. A study by researchers at Microsoft and Cambridge University demonstrated just how much of our inner lives is revealed by very simple pieces of data. They found that Facebook likes revealed a wide range of information about participants, including hints about their personality and their voting preferences (even though these were not explicitly identified in the likes). So marketers now have many more levers to play with from their big data assets. Technological advances have made it a lot easier to start applying this information to good effect — rather than blind A/B testing, marketing communications can now be shaped by a strategic understanding of what underpins preference. For example, many brands are starting to undertake persuasion profiling of their customer base to understand what types of nudges are most effective at shaping customer activity.
At the social level: Another way in which marketers can make more strategic use of big data is to start exploring how social relationships are revealed through data patterns, something very hard to do by other means, such as market research surveys. Many of our beliefs, attitudes, and behaviors are shaped by our social connections rather than, as classical marketing would suggest, our own individual preferences and experiences. A good way of thinking about this is to consider forest fires: The fire itself has its own properties in the way it spreads, which we can’t necessarily explain by examining the way in which individual trees burn. Big data allows us to look at the way in which social effects rather than individual preferences are shaping markets. A huge amount of data — phone logs, social media, messaging and so on — tracks exactly how behavior operates at a social level. Studies, including one by Microsoft researcher Duncan Watts, have demonstrated how patterns of relationships are themselves critical to preference formation in markets such as music downloads.
Network theory (which identifies the different patterns in the way we communicate with one another) is also relevant here, which Watts demonstrated with his “big seed” marketing approach. He used large-scale mailing lists for an initial “seeding” of viral messages to determine how social effects lead to sharing. This strategy never got widely adopted, which may indicate that marketers continue to resist seeing how a web of relationships can reveal how to build influence. Perhaps the time has come for marketers to reevaluate the use of network theory.
At the cultural level: An even more strategic opportunity for marketers is to explore data sets to understand how cultures are changing. A good example of this is Google’s Ngram service, a digital database of about 4 percent of the world’s books published since 1800, where users can plot usage rates of words over decades and centuries. This tool can help us to understand the ways in which ideas and language have evolved over time. Work by anthropology professors Alex Bentley and Michael J. O’Brien suggests that our use of buzzwords (which appeared in print) spread by social diffusion (copying) rather than reflecting the changes and developments in the topic itself. Hence, as Bentley and O’Brien’s write, “when humans are overloaded with choices, they tend to copy others and follow trends, especially apparently successful ones.” This finding can of course help brands shift the focus of campaigns toward cultural learning and away from emphasizing the soundness of their content. A new product may be verifiably better than previous versions, but if we ignore cultural learning as a means of communication, adoption rates may be weak.
It takes something of a leap of faith to see the full creative potential of big data for marketers. Senior decision makers generally don’t yet fully understand the opportunity, and data analysts don’t often have marketing expertise. The opportunities lie between these roles, and the prize for those brave enough to go looking is invaluable. Because personal attributes can be identified from data trails, marketing messages can now be much more effective. And as data analytics show us which trends are in decline and which are in ascendance, brands can put their best foot forward, anticipating consumers rather than reacting to them.