David Boyle - Creativity and Analytics for Strata 28th April 2019

So I've been lucky to have been asked to build analytics capabilities in a number of businesses over the years. Sometimes that has worked brilliantly. It's transformed the culture of the business. It's transformed the economics of the business as well. Sometimes, shall we say, not so much. So what predicts success when you're building an analytics capability? That's why I'm here. Is it the ability to hire the best data scientists? Is it the ability to invest in the data, the technology, the people and the systems required? Is it a chief executive's leadership of building a data driven culture? I struggled with all of those things in the past and I struggle with some of them today. But I'm here to tell you that actually, you know what? None of those things matter if you can build a culture that combines analytics and creativity. I'll tell you what I think that means on the analytics side. I think what that means is recognizing your work is necessarily imperfect, embracing the fact that your work is a creative interpretation of the world around you, and recognizing that you're designing for people whose job it is to come up with a creative response to your work in order to drive action. I'll explain that. If you're a business leader, it's about recognizing your job is to embrace imperfect analytics and to come up with that creative response. So I thought to myself, how could I find evidence for this? I believe it to be true. And fortunately, McKinsey also believe it to be true. They say companies that harness creativity and data in tandem grow twice as fast as companies that don't. So I've seen it and they say it. Let's find out what it means. I thought back over my career, all the business leaders I've worked with, all the people in different industries who've embraced analytics and creativity. And I tried to think, if I could give one award for a person that personified the combination of analytics and creativity, who would it be? The answer is Snoop Dogg. So, Snoop, this is for you. Let me tell you why I found myself at EMI Music about eight years ago. Our job was to use data to help the business. Yeah, that's it. That's actually usually my brief. How can data help? And of course, there are millions of ways that data can help business. We picked one to start with. We decided that the best opportunity was to use data to help drive the strategy for an artist, to help them to navigate the countries or audiences, to drive growth, both for an artist and the marketing teams as well. Step one, gather the data that you need. For us, it was A million interviews over a few years with customers, with audiences, with music consumers around the world on rich dimensions, trying to distill infinite complexity of people into ones and zeros. Next, it was crafting that into something that would be insightful, clustering on the dimensions we felt were most important, useful, actionable, to have simple clusters that we could use. And then predictive analytics. Step three, simple predictive analytics for each artist in each country, which type of customer, which segments would represent the best opportunities for growth. And for Snoop Dogg, there were a couple of segments that really stood out and one that he really latched onto. We called them gadgeteers. They were defined by their passion for technology, their passion for smartphones and social media networks and all things technology. Early adopters, it was one of their values. He really latched onto them. In the meeting, he kept saying, we've got to get them gadgeteers. We got to get them gadgeteers. And sure enough, in the weeks and months and years that followed, he went after them. He was an early adopter of social media, for good or for bad. He was all across tech startups and technology companies in partnerships and collaborations, and it worked. He built resonance in that audience and as a result, he's relevant to a whole new generation. The analytics to him was a bit like having a superpower. Thanks to his creative interpretation of the results of the analytics, and I would argue, thanks to the creative use of analytics to drive that interpretation. Creativity in analytics, but not just in music. I've seen this work in multiple industries. If you look at television, we build predictive models to understand which countries the BBC shows should be monetized in, using a bunch of different data sets, each of which was imperfect. Piracy data, social media data, engagement, customer preferences. Thanks in part to a company called Parrot Analytics. Combining these imperfect data sets, we were able to show, for example, for Dr. WHO in Korea, there was a big discrepancy between the revenues we should have been expecting and the revenues we were actually getting. The creative response by the business to have the cast and crew stop on the way to Japan, I think it was 4,000 tickets sold out instantly and 50,000 people applied for tickets. This is awesome. Soon the marketing was happening, the distribution was happening, and the show was. Was monetized the way it should have been in Korea, and it's happening in Harrods. We're doing exactly the same thing for brands that we stock in store and for departments and for divisions in store as well. It's early days. We'll see how this ends up. But One live example would be this place. This is the Salon de Parfum, a beautifully crafted space at the top of the store. Handcrafted scents from all around the world. The finest, really popular. But we've identified one type of customer. I won't tell you who. One type of customer who loves the store, shops. The store loves Harrods, but is not yet engaging in this department the way I think they should, the way the analytics suggest they should. Let's see what the creative response is. It might be new scents or new brands. It might be a change to the physical space, might be marketing or events. We'll see. But in the same way that the creative response by Snoop for the Gadgeteers and the creative response by the doctor who team to Korea, there'll be one here and it will be the same pattern again. When this comes together, when it all works, it feels to me like analysts are building maps. But as analysts, we sometimes make two fundamental mistakes in building our maps and doing our analytics. The first, I think, is to assume that the analytics, that the maps are striving to be an objectively true representation of the world. I don't think that's possible. And the second is to assume that our job is to craft analytics, which tells the business what to do instead of inspiring in them the creativity to come up with solutions. So I tell you, I talked earlier this week with a mathematician called William Byers, a mathematician and author on this subject. He likes to talk about how the data is not the situation. The map is not the territory. Right. If the map were the territory, it would be actually the territory. It would be completely unusual as a map. If your analytics were the situation, it would be the database that it's represented from is completely unusable. By definition. Maps or analytics are abstractions. They're summaries. They're a perspective that we have crafted on the world, that we think as analysts are useful to business leaders. That's a creative expression. We should embrace that. Also, it's not our job, it's not the role of analytics to tell businesses or people what it is they should do. It's our job to inspire in them the creative decision. If you're out hiking, it's not the map's job to tell you which path to take. You need to decide your ambition for the day. How long do you want to be away from home? How much food do you have in your belly and in your backpack? What does the weather look like? How adventurous are you feeling today? That will decide your path through the mountains? Snoop should decide his path through the customer segments and the BBC should decide on how to monetize Doctor who in Korea. Our job is to provide those opportunities to provide the maps, but not to tell people what to do. And much as I love inspiring examples where this has gone brilliantly, I feel it's on me to share with one example of where this has gone horribly, horribly wrong with big implications for all of us. And it's something I have to say I feel at least a little bit responsible for, albeit because of some things I did many years ago. When I first moved to Washington D.C. it was 2005. It's a pretty grim place actually. Republicans had the presidency, the Senate, the House, and they were far more advanced than Democrats were in analytics. We were working hard to catch up. We invested millions of dollars in analytics and people and data was awesome. By 2006, the Network of people competing and innovating, inspiring and challenging each other on the Democratic side had built enough capability to win back the Senate, the House, majority of governorships up and down the country, and a majority of state legislatures up and down the country as well. We were starting to win. Analytics was working. I left. But the a couple of years later, the Obama team went on to use those same people, analytics and data and infrastructure, plus a bit of innovation in order to win the presidency. Democrats were now masters of data. We were beating Republicans. So how did it come to be that Hillary lost the presidential election by 80,000 votes in three different states? Each dot here is 20,000 votes and the middle four represent the margin of loss in those three states. How did that come to be when we were such masters of analytics? Well, I've got a strong hypothesis and it's about this topic. It's about the rise of analytics over creativity. It certainly wasn't a lack of analytics. Here's Hillary talking about how her campaign was so data driven, how it made decisions for her, and how she reacted to the data so strongly. It wasn't a lack of analytics, I argue it's a lack of creativity in analytics and in use of analytics. It's the central lesson. How do I know that from thousands of miles away and not being involved in this process? Well, that was exactly the culture we were trying to build back in 2005, 2006 and 2007. I hadn't met Snoop Dogg yet. I didn't realize the necessary imperfections in data. I think they spent 10 years trying to perfect polling, trying to perfect predictive models based on polling. Instead of embracing that uncertainty and embracing the uncertainty of all the other data sets that could have helped to understand what was happening in Wisconsin and the other swing states. I think they designed analytics to say, here is what will happen with a margin of error. You should go here or not go here. Instead of designing analytics like we did with Snoop, to say, here are the next opportunities. Here are the types of customer who are most excited about you. Here are the types of situation we're seeing on the ground to inspire creativity and a leader to interpret that and come up with a creative response. I think more creative analytics and more creative use of analytics would have changed the campaign. It would have changed the result of the election. I've seen business leaders like this as well, paralyzed without perfect analytics, unwilling to accept the imperfections and unwilling to embrace the responsibilities. Come up with creative responses to that necessarily imperfect analytics. So let me end by talking to two audiences here. If you're an analyst, I think you should hide behind algorithms no more. Embrace the fact that your work is a creative representation of the world. Embrace its imperfections and design for business leaders that you can inspire. Design your work to challenge them and give them opportunities, but let them make decisions. Don't try to tell them what they should be doing and how they should be doing it. And seek out business leaders who respect that and embrace it. If you're a business leader, accept the imperfections in analytics. Find analysts who can craft stories that represent the world and celebrate that. Embrace the imperfections and recognize your responsibility to come up with a creative response just like Snoop did. When it comes together, when it works brilliantly, it feels like a superpower doctor who should know a thing or two about that, I guess, particularly in Korea. And maybe, just maybe we should ask Snoop to sit down and explain that to Hilary. Thank you very much.