GE, perhaps more than any other major company, is dedicated to the use of data visualization as a key part of its marketing and communications efforts. Stemming from last month’s Insight Center on visualizing data, I spoke with Linda Boff, GE’s executive director of global brand marketing, about the benefits and challenges of this approach. An edited version of our conversation is below.
What’s the history of data visualization at GE? How did your strategy around it develop?
GE specializes in complex challenges in solving the toughest problems in the world: Infrastructure, renewable energy, affordable health care. Things you have really have to get your mind around.
In trying to do that, the marketing communications brand group is always searching for compelling ways to bring these challenges to life. Five years ago or so, we started using data visualization.
One of our first was back in 2009, and was about causes of death. We separated them in male versus female, and via age spans. So if you’re 24 to 36 or what have you, these are the three things you were most likely to die of. Now it seems so simple, but it was really compelling:
We got a tremendous response to it. The media loved it. Our different stakeholders — be it customers, employees — everybody thought, wow, what a great way to tell a story, and it was sort of born from that.
How do you think about using data visualization when it comes to different audiences and stakeholders, both within your company and outside of it?
As a large multinational company, we do have many audiences. And they range from employees and retirees to retail investors and thought leaders. Initially we thought about this — and I think to a large degree continue to — as a way to do external storytelling, but we have found that it works on so many different levels.
As a result, we have used data visualization in places as diverse as our annual reports or our annual report app, which is obviously geared toward investors. We’ve used it with thought leaders. When we released a white paper last fall on the industrial Internet, data visualization was a great way to tell that story.
It really works across different audiences. That’s one of the things that’s perhaps most exciting about it. …
How do you staff for digging through all of that data, doing design work, and other digital elements?
The approach we took — and it’s an approach we often take — is that a couple people inside the brand marketing group spent a lot of time on it, but we also partner with the best of the best externally. And these are folks like Ben Fry, Lisa Strausfeld, Carlo Ratti at MIT, and Jer Thorp, then at The New York Times.
We didn’t say, OK, we’re just going to work with the design studio Pentagram or we’re just going to work with The New York Times or what have you. And that was a fabulous approach because it gave us the eyes and the sensibilities of folks in a number of different areas.
Also, we’re GE. We’re involved in everything from transportation to health, curing people to energy to building things. We wanted a diversity of points of view on a diversity of subjects.
What projects have been the most successful for you? And how do you define success?
Because we have approached this largely as storytelling, we’re always looking to experiment. …
We also paid a lot of attention to the kinds of things that content publishers and marketers do for engagement, comments, news coverage. Over the years, we’ve had great pick-up by people at publications, bloggers, all of whom are influential. That’s meant a lot to us because it’s a way for us to tell the GE story, and the amplification of that story is really, really important.
And it’s also been a way for us to “double click” on certain things. Let me give you an example of what I’m talking about. We’re an Olympic sponsor, and there are not that many Olympic sponsors. GE’s in there, with Coca-Cola, with McDonald’s, with their marketing machines so to speak.
I was looking back at what some of what we did for the summer games last year in London. And data vis was a transformative way for us to talk about the data surrounding the game. We made this wonderful visualization that was 100 years of world records for the summer games. And you could sort of click into it obviously and see by country and time, etc.
So what I mean by double click is that it was another way for GE to talk about the importance of the Olympic games and give a bit of a perspective on them over time, but using a tool that a decade ago we never would have. It’s enabled us to tell deeper, richer stories.
Another example, one that I really like, starts with the fact that GE generates about a quarter of the world’s electricity. That’s a lot. So we have a visualization right now that shows 713 turbines and the power generated over two weeks:
I can sit there and say, until the cows come home, “We generate a quarter of the world’s electricity.” But when you see it as a visualization, I think it’s much more memorable.
What are some data visualization experiments or projects you’re working on now?
We are working on one that I’m particularly excited about. Not long ago, we did what we called Flight Quest, … an initiative we ran with Kaggle. We released some data from our customers [at airlines], as well as data from the National Airspace System on never-before released flight times, arrival times, flight numbers, origins, arrival cities, all of these different elements.
We released all of it to the Kaggle community of data scientists globally and said: improve travel.
There were five winning algorithms that came up with a 40% improvement in flight arrival times. Eventually this could be software that could be incorporated into an airline’s system to improve arrival times. But what we’re working on now is a 3D interactive visualization of those winning algorithms. [Phase two of the project begins in June.]
What are your biggest challenges as you build new visualizations?
One of the biggest challenges today is that people expect data is that is very real-time and current. … And then the other piece of it is, how do you make it relevant? I think The New York Times has done a fantastic job [on these fronts]. I think Wired has done a nice job on this.
But at the same time, if the point is to simplify a story or make whatever the topic is a story well told, if it gets overly complicated it defeats the original purpose. So I think that’s the line we all have to just watch out for a little bit. And we’ve learned this as we’ve gone, whether it’s the topic or the how pleasing the interface on. Some things are just more inherently interesting than others.
And you just have to experiment to figure out what works and what doesn’t.
I think so. And I have no regret in experimentation because I think we wouldn’t be where we are if we hadn’t experimented along the way.
What advice would you have for other companies, be it big companies or small companies, about to why they should take data vis seriously? And what lessons would you impart to them?
The power of a good story well told in any sort of medium cannot be overstated. Data vis has allowed us to do storytelling at its best. Experimentation is also key, getting in there, understanding a medium and a technique, and not being afraid to experiment with it and be open and collaborative. We have had data marathons with many universities where we’ve brought in students, given them a problem, and said, hey, let’s work over the next couple of days to solve this.
This is an open space. This fact is incredibly important. Open experimentation is a great way to bring to life challenges through vivid storytelling
Dan Meyer: Math class needs a makeover
Mae Jemison on teaching arts and sciences together
Liz Coleman's call to reinvent liberal arts education 
Shimon Schocken: The self-organizing computer course
Geoff Mulgan: A short intro to the Studio School
Tyler DeWitt: Hey science teachers — make it fun
Kiran Bir Sethi teaches kids to take charge
