SHARE
Joonas Rokka, EM Lyon and Lionel Sitz, EM Lyon

In management circles and beyond, companies are rushing to integrate, adapt and exploit big data in their organizations. Moreover, they are willing to recruit nearly anyone with a mention of “big data” or “artificial intelligence” in their resumes.

There are nonstop consultant talks and crowded workshops on big data, and academic journals are rushing out special issues with the magical keywords. Nearly absent before 2011, big data is well on its way to be the most talked about topic in the management press, including the Economist, Financial Times, Wall Street Journal and Forbes.

Business schools, too, are rushing to restructure their offering around big data and analytics – it seems as if nothing more is needed. Yet little is said about the kind of understanding and reflexivity that is needed when working with such voluminous data. We believe that an important lesson can be learned from ethnographic research, which should be taught to managers obsessed with big data.

Why Teach Ethnography to Managers Big Data Diagram

Big-Data Obsession in Management

Companies are wasting no time in leveraging big-data solutions to predict behavior, profile their customers, and to enhance their marketing effectiveness. For example, many use big data to target and develop recommendation algorithms.

Amazon suggests relevant products you are likely to be interested in buying, Netflix lists movies you are likely to want to watch, Spotify and Pandora propose songs you might enjoy listening, and Zappos optimizes its entire product selection accordingly. Through these and countless related developments, big data has already become a reality in our daily lives.

Big data offers great advantages, and detecting patterns in customer behavior is beneficial for companies. It offers a powerful way to refine customer profiling and to develop subtle, automated targeting strategies. Notably, it allows discovering correlations that would otherwise remain unnoticed.

For example, Walmart uses an ocean of data – the retailer analyses more than 2.5 petabytes per hour – to see if they can identify unknown correlations in consumption habits. And they have indeed found many: berries are likely to be consumed more when there are low wind and temperature is below 27° C, whereas cloudy, windy and warm weather prompts steak purchases, and hot, dry weather and light winds trigger hamburger sales.

Such correlations are not new – they simply were difficult (or impossible) to detect before. This is why companies reach out to combine and explore even more types and dimensions of data they can. Health insurance companies, for instance, ask their clients to give access to their Fitbit activity data, with the promise of a better insurance deal.

Many retailers have in-store traffic monitoring systems allowing them to track the location of customers within the store as they move in order to optimize their marketing effectiveness. In fact, big data is being proposed as a solution for just about everything. Burger King recently used it for designing their advertising – however, with somewhat disastrous results.

We argue that the obsession with big data collection and analysis risks becoming an end in itself. This would significantly narrow down the types of understandings that are being produced, valorized and acted upon.

Managers also need to foster sensitivity to sociocultural, contextual knowledge which, unfortunately, is largely erased by big data storing mechanisms. Pure correlations with weather and purchase behavior, for example, hide deeper level cultural processes at work – not least how the sunny weather and blue sky “call” for a barbecue precisely because of the inherent connection between lifestyles and consumption.

Big Data Limitations

Surprisingly few talk about the potential limitations. First, due to the thirst for ever more data, there seems to be no end to how much is enough data. At the same time, collecting, storing, updating, and curating big data is – of course – extremely costly.

For the record, many have also claimed that much of this data is hardly useful at all. But since they do not know which data could be interesting or not, managers have decided to keep on collecting it. In many cases, unfortunately, companies do not have resources to properly distill meaningful insight from it. These companies would thus arguably be better off not venturing in big data collection at all.

Second, big data relies on petabytes of what we call “decontextualized” data – in other words, data points extracted from the actual situation from which they were produced in the first place. The number of “clicks” or “views”, for example, are often closely measured and recorded but they do not inform managers about the immediate contexts, moods or situations in which users were clicking and viewing the website.

Despite technological progress, a significant part of this context will always remain impossible to measure because of its inherent complexity. Yet it is a crucial factor for the understanding and explaining the studied behavior at stake: doings and sayings become meaningful only in their immediate sociocultural context.

Third, we argue big data is unable to address embodied, sensory and affective experiences. When seeking to measure an emotion, for example, big data may only hope to measure the physiological reactions of the persons captured via sensors (muscles tension, sweat, heart rate, brain, etc.), but not the acute meaningful emotional states that people live through.

When analyzing tweets to determine people’s emotions, data analysts agree they could not address emotions themselves but only traces of their narration. This is a crucial caveat as the sensory dimensions are essential toward fostering understanding of the experiences that people actually live through. This is problematic, as human lives are essentially about experiences.

Finally, it is safe to say that big data alone is not helpful for developing a “deep” understanding. What big data scientists can find out are correlations between variables (what is or happens), not causation (why and how it happens). Hence, big data is an interesting and useful tool but should not be the only focus of attention. This is why we turn next to examining ethnographic thinking and research as a potential antidote for big data obsession.

Benefits of Ethnography for Managers

While big data analytics are quickly entering the curriculum of most business schools, ethnographic methods often remain reserved for social sciences departments of universities. However, some institutions have decided to make them a much more visible and foundational part of management learning – from consumer behavior, marketing research, branding, service experience, and strategy – where ethnographic reflexivity and methods are actively in use.

First, ethnography is all about gathering in-depth data about lived experiences and situations. Anthropologist Clifford Geertz famously described this kind of data as “thick descriptions”: long-term and deep reflections about the experiences that people live.

Expert on Balinese culture and rituals, Geertz crafted his insights on first-hand participant observation with the idea that the ethnographer needs to live through the same experiences as the studied people. Thus, she/he is committed to discovering and sharing a common phenomenological sensibility and understanding – in a way, in his attempt to get inside the “skins of others”.

The method has been a staple in anthropology and sociology for over a hundred years, but it is gaining acuity and relevance in understanding today’s fast-paced society and markets. Producing in-depth data is useful not least for managers wishing to grasp, for example, customers’ or employees’ experiences – from their point of view.

Why-Teach-Ethnography-to-Managers-Big-Data-Era-Leadership-Geertz
Clifford Geertz, an expert on Balinese culture and rituals, asserted that ethnographers need to live through the same experiences as the studied people to understand them. Above, a village ceremony in 2018. Artem Bali/Flickr, CC BY-NC

Second, ethnography insists on reflexivity. This means that the ethnographer seeks to question her/his own preconceptions about the studied phenomena – a sort of unlearning about “what we think we know” is thus required. Also, it means that the ethnographer is mindful about the way she/he participates in shaping the studied realities: the kinds of questions being presented and the power exerted over those studied. In practice, this means being sensitive toward ensuring that people indeed share their unique views, experiences, and narratives.

Ethnographers are taught to mistrust what they consider “natural”, “normal” behavior and “objective” evidence. For example, people who were born before the Millennial generation consider a cassette player to be a basic, ordinary object. But those who grew up interacting with smartphones and tablets such an object can be a mystery. Ethnography can thus help managers foster reflexivity about the “limits” of their own experience and being attentive to the difference and multiplicity of understandings and truths.

Third, instead of gathering mountains of data on as many (decontextualized) variables as possible, ethnography seeks a profound understanding of the situational context. Within it, the objective is to uncover the social processes that help explain the reasons why people are bound or likely to act the way they do. In 2013, Netflix worked with anthropologist Grant McCracken to understand the emerging online video-streaming phenomenon.

The company had no shortage of data about customers’ video viewing statistics but wanted to dig deeper into the social dynamic at stake. McCracken’s ethnographic work revealed the meaning and importance of “binge-watching” for contemporary consumers. For him, our “digital lifestyle, where storytelling is often reduced to bite-sized, 140 character conversations or instagrammable images, leaves us craving the kind of long narrative of storytelling”.

McCracken found that 73% of consumers feel good about “binging”, i.e., watching multiple series or movies in one viewing. This kind of analysis was indeed fruitful for Netflix toward better serving their customers.

Fourth, in radical contrast to big data approaches, ethnography is concerned with the building up of “embodied data and knowledge”. In other words, the building of analytical accounts produced by our very own bodies (by way of seeing, sensing, touching, hearing, tasting) – about life. Ethnography is particularly sensitive to the multisensory aspects of people’s experiences. Going to a live concert, a sports event, or political demonstration cannot be reduced to the spectacle or “show” itself. Something happens during these events that can only be felt in our bodies, for example, an atmosphere of thrill emerges from social interactions, sights, sounds and other impressions which can sometimes touch, even change us. There is something in the aliveness and vivid flow of experiences that can only be addressed via our senses and which cannot be captured by “dead” big data descriptions, cut out of their contexts and summarised in static charts or representations.

Toward Teaching a Reflexive Mindset

The above points emphasize a crucial fact: that for producing knowledge and insight about human behavior, we may need more than big data. Ethnography calls for a curious and reflexive mind that is open to exploring novel understandings and perspectives, challenging taken for granted assumptions and norms. It also insists on an economic principle: we need to gather new data until a “saturation point” is reached – when gathering any new data produces no further insight.

We argue teaching ethnographic thinking to managers is now more acute than ever. The world is changing with stunning speed, a flood of data is being produced by computing systems, and there is only little time to make decisions. Ethnographic mind-set enforces managers to:

  • continuously reflect on the “right” questions and perspectives they may adopt,
  • exercise participant-observation which can be a “lifelong” asset,
  • critically analyze the kinds of seemingly “objective” empirical evidence offered to them (no matter how voluminous)
  • take a few healthy steps away from the ocean of data they may easily drown in.

Joonas Rokka, Professeur associé en marketing, EM Lyon and Lionel Sitz, Professeur de marketing, EM Lyon

This article is republished from The Conversation under a Creative Commons license. Read the original article.

LEAVE A REPLY