Marketing remains as much about storytelling as ever, but the language used to craft these stories is beginning to change.
Historically speaking, most marketers have conceived of themselves first and foremost as storytellers. Marketing has always been about crafting a narrative that is at once couched in truth and as compelling as the best fiction. Though the digital revolution hasn’t necessarily displaced this cornerstone industry principle, it has forced marketers to start telling stories in a language in which they are not entirely fluent: data.
Granted, marketers still deliver their messaging in English and Spanish, Mandarin and Arabic, but the tone and tenor of these efforts are becoming increasingly dependent upon insights derived from data science. Consequently, basic data literacy — or the ability to understand, use, and communicate data effectively — has become absolutely essential for everyone working in or adjacent to the marketing sphere.
Learning a new language is not always easy, but fortunately, the average marketer need not become a bona fide data expert in order to forge a strong working relationship with their more experienced analytics partners. Here’s what marketers need to know to communicate intelligently and effectively about data — no interpreter necessary:
Differentiating Between Data Types
The most important thing for marketers to realize about analytics is that the underlying data are often tied in some way or another to the “real world,” and thus are inherently subject to both change and imprecision.
Data analytics are grounded in hard science, but there’s no “right” or “wrong” way to approach gathering, organizing, and analyzing information. Certain approaches will more often than not produce better outcomes in certain circumstances, to be sure, but good analytics are highly sensitive to contextual variables.
It is therefore helpful for marketers to calibrate their expectations according to the kind of data an analytics project is using, the complexity of the questions being asked, and the amount of experience the analytics team — whether in-house or external — has with the targeted consumer segment.
To drill down a little deeper, data scientists make a fundamental distinction between deterministic data and probabilistic data, each of which offers a different kind of value to marketers. Deterministic data is willingly and knowingly submitted by consumers — think consumer survey responses, login information, or transactional data — and tends to be highly accurate, at least within a certain timeframe. This data enables a marketer to, for instance, track a unique target across their various personal devices – smartphone, tablet, laptop, and so on. The drawback of deterministic data is that it is rarely updated and is only available from consumers who are already engaged with a brand in some basic sense.
Probabilistic data, on the other hand, is far more speculative, but with this greater level of uncertainty comes a much broader scope of insight. In short, probabilistic data emerges from repeated analyses of large data sets that eventually begin to uncover patterns that can be statistically transposed onto the general population. Once a marketer manages to collect a sufficient amount of probabilistic data, they can use it identify trends in how a specific kind of consumer will behave in the marketplace, even if they are unable to pinpoint the behavior of a unique target. This becomes incredibly useful when the desired deterministic data is too difficult to access – or simply doesn’t exist.
Understanding the Building Blocks
Understanding the type of data that an analytics team is using for a given project — and adjusting expectations accordingly — is an important first step, but this is just the beginning of true data literacy. Like any other language, data is comprised of a variety of constituent parts, and before a marketer can comprehend the ins and outs of complex analytics, they must become familiar with the “grammatical” building blocks underlying the datasets being analyzed.
Just as spoken languages have nouns, data includes “entities” which refer to people, places, and things. Everything from the mechanisms of a marketing strategy — various channels, activities, and campaigns — to the assets placed in front of consumers — ads, apps, sites, products — to the consumers themselves will show up as entities in a marketer’s datasets.
The interactions that occur among entities are called “events,” and like verbs they capture anything and everything that “happens.” Each event takes place at a specific time and in a specific place, and analyzing the sum total of events related to a particular set of entities can help a marketer gain an understanding of the causal relationships that are at play.
Finally, metadata function like adjectives and adverbs insofar as they describe the properties and characteristics of entities and events. While entities and events determine whether or not a marketing strategy is working, metadata provide a window into why. Descriptions of consumer entities in particular are incredibly valuable to marketers who are attempting to pick out behavioral trends among consumers in general, as such metadata inform marketers of what kind of people are the most likely to take what sort of actions.
Entities, events, and metadata are meaningless in isolation — and it’s a data analyst’s job to combine them in a coherent way — but marketers who are conversant in the differences among them are going to have far better relationships with their analytics teams than those who are not. Beyond this, the most valuable elements of data literacy will vary depending on the situation.
Regardless of the circumstances, however, the more a marketer learns about data, the more refined and nuanced their campaigns will become. Ultimately, the modern marketer need not speak data fluently, but if they can listen to data-speak with a baseline level of comprehension, they are well-positioned to cement a substantial competitive advantage over their less multilingual competitors.