Effective Data Analytics Is as Much About People as It Is About Numbers

As technical as it is, data analytics is a fundamentally human enterprise with profoundly human stakes.

When people outside the industry ask me about my work, I often tell them that I am in charge of the “beeps and boops” that guide pharmaceutical marketing campaigns. I transform the data that is generated by consumers’ and healthcare providers’ online activity into coherent stories that help pharma decision-makers stay on — or, if necessary, adjust — course.

This, of course, is a major simplification of what I do. As someone whose background is in data engineering, I have spent a significant portion of my professional life tinkering with the complex set theory and statistics that are embedded within programming languages like SQL to make sprawling unstructured datasets usable for other analytics team members and, by extension, clients.

For the first 18 months or so that I was at Saatchi & Saatchi Wellness, this was exactly what I did. I camped out on the back end, focusing almost exclusively on SQL database design, report population, and so forth. This ended up being a great way to find my feet in the pharma space — I had come to Saatchi from the utilities sector — but in the several years since, I have become progressively more involved with a variety of other aspects of Saatchi’s work.

This transition has been illuminating in countless ways, but perhaps the most compelling insight I have gleaned from being forced to think beyond the beeps and boops is that data analytics — even the most technical aspects of it — is a deeply human enterprise. Here are three reasons why:

  1. Analytics Is Always Performed with an Eye Toward the Client

There is always going to be a “language barrier” between a team of analytics experts and pharma decision-makers. Our clients are highly intelligent individuals, to be clear, but people are seldom well-versed in the language of analytics unless they have made analytics their career. As such, helping our clients maximize the value of their data is often an exercise in reading the room.

Each of our clients has a different level of data literacy, which means each of them needs something different from us. I have had clients request that I refrain from including any numbers whatsoever on presentation slides, just as I have had clients insist that I explain the nuances of the training data we used for a machine learning project in painstaking detail. At either extreme — and at every point in between — our job as an analytics agency remains the same: frame our findings in whatever way will ensure our audience is not only able to understand them, but able to put them into action.

There are a number of approaches to doing so, but one that I have found to be particularly effective involves making sure clients feel in control from day one of an analytics project. As I am preparing to launch into my presentation during an initial client meeting, I make it clear that what I am about to lay out represents a thoughtful, carefully crafted first pass — but a first past nonetheless.

While, as an agency, it is our responsibility to manage the technical aspects of an analytics project, our client is the ultimate arbiter of the contours of the project and the nature of the reports that trace its progress. Without going so far as to downplay our expert opinion, our goal is to deliver what our clients want, not what we want — the agency-client relationship should not be a paternalistic one. This can be a difficult balance to strike, especially when a client’s data literacy is limited, but it is part and parcel of exceptional service in our line of work.

  1. When It Comes to Analytics, Two (or More) Heads Are Better Than One

Just as collaborating with a client increases the likelihood we will set our sights on the right target, collaborating internally increases the likelihood we will hit said target as precisely and with as much force as possible. As one of the first half-dozen members of the Saatchi analytics team, I have seen our offerings expand and mature over time, and the lion’s share of this organizational evolution has been the byproduct of internal collaboration.

Many of our larger, more complex offerings were not only developed collaboratively, but are now operationalized through the collective efforts of multiple analytics team members. Advanced analytics solutions are rarely turnkey technologies; rather, they are frameworks that must be adapted to clients’ objectives, budgets, and data. To produce the best results, every aspect of a project — from the data sourcing and active template library processes to the training data manipulation and report population — must be tailored to a client’s unique needs, and this almost always requires input from analytics team members with varied areas of expertise.

The more I have been asked to interface with clients over the years, the more I have come to appreciate having a team of such varied skill at my fingertips. I have a fairly extensive background in data science and engineering, but when a client asks a nuanced question about, say, data strategy, both the client and I are going to be best served by consulting a true expert. The collegial culture at Saatchi is such that I can respond to such a question with, “I understand the question and am able to give you an answer, but let me bring in my colleague who is going to be able to give you the best answer.”

Ultimately, approaching analytics as a team sport leads to the development of more innovative solutions, the achievement of better outcomes, and the forging of stronger bonds both internally and externally.

  1. Analytics Speaks to Human Concerns

Most sweepingly, data analytics is an enterprise that has fundamentally human stakes — particularly in recent years, and particularly in the pharma space. The digitization of nearly every element of our daily lives has enabled analytics professionals to paint fuller, more detailed pictures of what patients want and need. In our industry, this can mean healthcare providers learning about new therapies that might change the way they approach their job; it can mean helping healthcare providers better understand the relationships between their patients’ comorbidities; or it can mean patients getting access to the treatments they need more quickly than ever before.

Data analytics — and especially data science and engineering — may be comprised of beeps and boops, but as an agency, our job is to take these and transform them into insights that can be used to some human end. Every data point corresponds to some action that some person took or some characteristic that some person manifests, and while it takes a great deal of training to understand these correlations — just as it takes a great deal of training to decipher a second language — at the end of the day, data is simply another way to represent human idiosyncrasies.

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.