What a scientist learned from digital analytics

Early in April I found myself at a hotel in Downtown San Francisco, surrounded by experts in the field of internet marketing. I was not lost; I went there to learn how to be a better scientist. Let me explain.

Looking for industry examples of communicating with data led me to digital analytics. This branch of internet marketing relies heavily on data analysis and hypothesis-testing, which sounds like a scientist’s bread and butter.  Like scientists, analysts will be successful if they can demonstrate their approach is better than a gut feeling. Unlike most scientists, digital analysts routinely present quantitative data to non-technical audiences. And being part of  the marketing world, they take their presentations seriously.

In contrast to scientists, marketers are expected to explain what they do to people in all fields. At a meeting of digital analysts working in marketing, I expected to find data storytellers; experts at crafting knowledge into narratives to drive good business decisions. Actually, most analysts admitted struggling to communicate effectively. They were frustrated that their insights were not apparent to others, or often not implemented. The difference between this crowd and that at a typical scientific meeting was this: The digital analysts knew that their professional value depended on good communication, and they were there to learn the best techniques.

A large portion of the conference was on techniques for helping a company become data-driven. Good analysts know that data alone is never enough to drive a company’s decisions. Instead, they focus on the insights learned by analyzing the data. For example, to learn how to increase conversion rate (the portion of visitors to a website who take the desired action, like subscribing to a newsletter) analysts may look at time spent on the site, the path a visitor took to get there, and lots of other metrics. Like many scientists, they may love the technical challenge of  collecting, analyzing, and making sense of the data. But in the end, the outcome that matters is telling a company what actions to take.

As a scientist, there is a lot to gain from thinking like a digital analyst. Most people will never understand the details of our work, and that’s fine. Insights are how we create value. How many PhDs invest time helping nonscientists understand the insights from their work? How often do we give up and say, “They just don’t get it.”? The analysts at the conference took time they could have spent learning analytical methods, and invested it in learning to communicate the impact of their work. I can’t imagine that happening at a chemistry conference.

I learned many, many valuable tools from my two days at eMetrics. The links below show just a few that you may find useful too. If these take you out of your comfort zone as a scientist, that’s probably a good thing. As for myself, I’ll continue looking for ways to share insights from data. Maybe I’ll see you at the next convention.

I’d love to hear from scientists: Who are your role models when it comes to delivering insights?

Presenting Insights

Lea Pica (@LeaPica) hosts a great podcast on data visualization and presentation skills. It’s geared toward digital analysts, but I learn something useful every time I listen. When it comes to presenting data, scientists and analysts face the same challenges. In this episode, digital analytics pioneer Jim Sterne (@jimsterne) talks about the common mistakes analysts make when presenting data. To scientists, this will sound very familiar; and the take-home message is relevant to anyone who works with data.


Becoming Data-Driven

Tim Wilson (@tgwilson) talked about solving a common problem: how do we know what is important to measure in order to achieve business goals? This problem should be on the front of every professional scientist’s mind, and the slide deck here shows many of the key points from the talk. Tim also co-hosts a podcast called The Digital Analytics Power Hour.


Visualizing Data

Ryan Sleeper (@OSMGuy) showed some of the features of Tableau, a data visualization software that is widely used in the marketing analytics field. Compared to the science tools I’m familiar with, the graphs are visually dynamic and designed for interaction. Check out this graphic to show, both visually and statistically, the cause of concussions in pro football. Science visualizations have been improving recently, but we would do well to take some lessons from the marketing world.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s