Bain uses cookies to improve functionality and performance of this site. More information can be found in our Privacy Policy. By continuing to browse this site, you consent to the use of cookies.

We have limited Spanish content available. View Spanish content.


Data-Driven Marketing: A Conversation with Geoffrey Sanders

Bain Partner Cesar Brea and Geoffrey Sanders, a member of Bain's Advisor Network, discuss how data and analytics are transforming marketing.


Data-Driven Marketing: A Conversation with Geoffrey Sanders

Bain Partner Cesar Brea sits down virtually with Geoffrey Sanders, a member of Bain's Advisor Network, to discuss the many changes that have taken place in marketing and how data and analytics have allowed marketers to work faster and more effectively.

Read a transcript of the conversation below:

CESAR BREA: Hello, everybody. Our guest today is Geoffrey Sanders, who's a member of the Bain Advisor Network. Thanks for making time, Geoff. Our mutual friend Ben Clark suggested we get acquainted. And I'm very glad that he did. You've had a very interesting set of experiences recently in top marketing, commercial and analytics roles at B2C online natives, and I wonder if you could recap those for us a bit.

GEOFFREY SANDERS: Sure. So I'm really glad we've had a chance to connect here, and I'll go a little bit further back in time. So, coming out of business school back in 2005, I knew I wanted to get a data-driven marketing role. And I actually found one in a marketing leadership development program at Citi, predominantly in Citi's credit cards business, and ended up staying there for almost 10 years. I spent the first five years in very traditional marketing channels, like call centers and direct mail, and then pivoted to spend the next five in digital marketing roles.

I spent the next two years at Gilt Groupe, where I ran digital marketing, CRM and loyalty, and was also part of the team that sold the company to Hudson's Bay. And after that, I found myself back in financial services, at Betterment, the online investment advisor, where I spent two years running acquisition and CRM—so really end-to-end performance marketing—and then went back to e-commerce and was at Casper for about two years, where I ran growth, analytics and North America e-commerce business, and left Casper in July.

BREA: So that's a really cool set of experiences, and I'm always interested in the logic behind these moves. Can you tell us a little bit about how they happened?

SANDERS: Sure, well, it certainly wasn't my intention to bounce back and forth between financial services and e-commerce. But there were two common threads in those transitions. Number one is finding roles where I could bring together data and technology to drive marketing, and therefore business performance. And then number two, following great people. And so for me, much more than the industry, that thread has been data-driven marketing as it existed in B2C companies.

But I've also had the good fortune of following some great leaders.

I followed a great leader from Citi named Michelle Peluso, who brought me to Gilt, where she was the CEO at the time. She's now the CMO at IBM. And then I followed Gilt's CMO to Betterment, following a great mentor and friend named Clay Cowan, who happens to be a partner at one of Bain's competitors.

And that's also something that's been important to me as a leader, to bring great folks with me to the new companies that I've gone to. And then in terms of Casper, I didn't actually know anyone there at the time I started discussing it. But as someone who really wanted to be a performance marketer at scale, the opportunity at Casper was just a great fit.

BREA: Now, you made the transition from a major financial brand to working at all these different insurgents. Aside from some of the obvious stuff, what was different for you?

SANDERS: Honestly, the biggest difference was one of process. At Citi, there was always a way that something was done and relatively clear answers on who needed to approve. At the insurgents, as you call them, there's not a 200-year history of how to get things done. And so actually—and this was one of the biggest changes that I had to go through—it wasn't clear who needed to approve or who had the final say. And so, as I got more comfortable with the situation, I had to decide more and more that it was me. And that is an uncomfortable transition. And so what I've tried to do is build some of the process that's existed in bigger companies into insurgents so that it's not necessarily the wild west.

BREA: That point about unclear decision rights, decision rules, is actually a really fascinating one that most people don't talk about. Aside from that, what else surprised you about that transition?

SANDERS: I think another big surprise was the speed at which I could work. This was sometimes thrilling, and literally I found, shortly after joining Gilt, that there were some things that my team could do in an afternoon that had taken months at a larger company. But that speed and nimbleness also bring some challenges in that sometimes, in cases, we were nimble simply because we could be.

And sometimes—and I think this is less marketing and more human psychology—sometimes bias towards action meant that it felt better to do something vs. doing nothing. And maybe at times doing nothing might have made sense. But that nimbleness and the ability to make decisions, move quickly has been really positive.

BREA: Very cool. Now nimble and action—there's always objects associated with those descriptors and nouns and so forth. Just tell us a little bit about how you approached those opportunities. What frames did you apply for understanding what was going on? And how did you conceive of what specific actions you wanted to try to take to drive performance up?

SANDERS: So I think what this nimbleness means in an ideal world is that you don't need every hit to be a home run. If you're being nimble, you basically get more at bats. And while you might strike out a few times, you can hit a bunch of singles and doubles. Where if you only get one hit in the game or one hit in the inning or whatever the right metaphor is, to win, every one's got to be a home run. And so, given that, the approach that I tried to take was really testing things in market as much as possible and getting real consumer data vs. spending tons of time building a case a priori. If the costs of actually doing an in-market test are low, it's much better than taking the time to be a little bit more academic about it.

And then, frankly, if the costs of iteration are low, it's iterating in market, again, and trying to improve things, because I think the odds are small of hitting a home run every time on the first bat. This is a process change, but it's also a cultural change, as in general people are averse to failure. And so part of the obligation of a manager is giving people a safe place to fail. As long as they have a good hypothesis for what they're testing and as long as they're committed to learning from failure, I think the benefits of that are pretty significant.

BREA: Very interesting. Now can you be a bit more specific about, for example, what were some of the key insights you developed for actually improving performance along the way? What were some of the big hits that really made a difference in the different opportunities you addressed?

SANDERS: So a few things come to mind there. I'd say the first is, at a lot of companies, the worlds of acquiring new customers and getting value from existing customers or retaining them are pretty separate. And that's true even in some of the smaller companies that I've worked at. But there really is a deeper connection between those two parts of the customer life cycle. Particularly, as more and more companies think about LTV and think about LTV to CAC as a metric against which to manage the business, the more that those worlds can be brought closely together, the more you can create a virtuous cycle of performance to drive the business.

Another—and this is an area that I'm deeply passionate about—is marketing attribution. And actually there's a fundamental difference between marketing attribution and incrementality. To me, being able to measure something is attribution. Understanding how it changes consumer behavior is incrementality. And often those terms are used interchangeably, but they're really, really different. And I think it's critical for companies to focus on the incrementality approach vs. just attribution.

And then, finally, I'm clearly a data-driven marketer, and so my head naturally goes towards the numbers and the targeting and the segmentation. But great creatives can be just as powerful as all of those levers in some cases. And so that can be easy for folks like me to overlook. And so it's really important to either have those self-reinforcement mechanisms to focus on that or to be working with the right people to be the complementary skill set and that complementary approach to marketing.

BREA: Yeah, that self-awareness about what you bring to the table as a marketer is really so important, knowing what you do bring and what you're missing and making sure you complement yourself with the right kinds of talent—really important. How did you balance improving near-term performance with progress on the underlying capabilities that you'd need to sustain that over time? What were some of the trade-ofs that you made?

SANDERS: Cesar, it's funny, because at a lot of companies those are viewed as opposing sides of the spectrum. And I don't think they should be, I guess. I'm generally not a fan of monolithic releases of new capabilities. When you approach it that way, often the requirements or specs that someone figured out for that release might have been done nine months ago and aren't relevant anymore. And so that's frankly too long, right, in the world that we're living in. I mean, does anyone remember what it was like in the end of December, nine months ago? The world was completely different.

And so I think the approach to enhancing underlying capabilities should really be iterative, where those iterations are tested in market to see if they're driving performance and then that data drives some of the future roadmap. Now it's never going to be perfect, and there are certain projects that take months and maybe, in cases, years to build. But then it's just really important to build into the process a way to reality check with what is the state of the team, what is the state of marketing at some point after the requirements are quote-unquote "set in stone."

Folks who are familiar with Agile development vs. waterfall approaches, they're probably hearing things that they're familiar with as I'm talking about this. But this isn't just a development thing. It's the way the world is changing. And I'd say that, on average, 40% of what my team did at any point in time didn't exist the year before. So if that's the case, how would anyone have known to put the underlying capabilities to support those on the roadmap?

Again, don't take away from this that you shouldn't do long-term planning. But there needs to be flexibility and agility in that to make sure that you're actually building for the world that you're in, not the world that you thought you would be in a year ago.

BREA: Now near and dear to my heart, one of the places where we typically have this trade-off felt most acutely is in the world of trying to apply machine learning and advanced analytics to some of these opportunities, because you obviously want to see if you can get out of the gate and get some momentum going. But on the other hand, sometimes building big pipelines can actually take some time. Are there specific examples of how you applied those techniques that you talk about and what you took away and learned from them?

SANDERS: Sure, machine learning has been becoming more and more prevalent on the paid media side of the world over the past few years. And I think it's driving a lot of the success that the digital platforms—Facebook, Google, Amazon—are generating. And they also allow companies to rely more on in-house teams vs. using agencies. And so I think there's a lot there that some folks don't even think about as being machine learning. But it's really powering all of the look-alikes and bidding optimizations in those tools.

So one of the areas that's been near and dear to my heart since I started in the digital world has been around website optimization. And so an example that I'll speak about is, all too often, there's an argument about what should be on the home page. It doesn't matter what the company is. It doesn't matter. It's always, what are people going to see? And that argument, if you think about it, is about determining what's right on average.

But the power of machine learning is supplementing that with, how do you service the long tail? And so there are some examples that I can think of where, as opposed to having one home page at a company that I've been at, we've actually gone into peak periods of time with having four versions of the home page, letting a tool optimize effectively in real time for which was driving the best performance.

And this is an out-of-the-box capability and a tool that is not that expensive. And so that's not the sort of thing that should take years for a company to build and years for a company to use. And we're able to see a lift in website conversion simply by letting the tool optimize, basically change every hour, what was shown on the website. And that is even a baby step, because that's still optimizing for what's right on average in that hour of time or whatever the period was.

Where it can start to get even better is, how is it optimizing differently for traffic coming from emails vs. traffic clicking through from search vs. direct load, because those are very different customer mindsets. And there's a lot of sophistication there. But it also means you don't have to get there right away.

BREA: Now that's a very natural transition to another topic I wanted to cover with you, which is, we get a lot of questions from clients that sound like, have I got the right adtech or martech stack? How did you go about assessing and changing yours?

SANDERS: Yeah, it's a great question. And I think it often assumes that there's a perfect or a right answer to that question. And there's certainly not a perfect answer, and I think there's better answers there. Now a lot of companies seem to think about checking the boxes. Do I have whatever acronym? Do I have a CDP? Do I have a DMP? Do I have a data lake? And I think it's important to remember that none of those have value in and of themselves. And actually there's costs, both financial and opportunity costs, to build any one of them.

BREA: Right, definitely.

SANDERS: But it's really more about, what are you going to do with those tools? And so the way I think about it is, I think companies should think about their key business objectives. What are their existing consumer behaviors of their customers or prospects and the behaviors they want to drive? And then what is the right mix of people, process and technology to meet those goals? It could be—and maybe this is heretical—it could be that it's less expensive to hire a junior member of the team one year or two years out of college than to undertake a significant cross-functional tech build if what you're looking to achieve is relatively straightforward.

And so I don't just assume that building an expensive and complicated adtech and martech stack is going to revolutionize your business. And maybe that comes off as a little bit too negative about technology. I mean, I believe that's a huge driver of marketing performance. But there are times where this attitude has led me to add tech to the stack, particularly around media measurement in worlds where I'm spending a lot of money. And frankly, there are times where I've chosen not to add technology to the stack.

And so for instance, for me, the use cases around DMPs are becoming a little bit less clear in a world where the cookie may be crumbling a little bit. So just checking the box on that acronym probably isn't enough anymore.

BREA: That's interesting because I think, back to this interplay between performance and capability, to your example about hiring the junior person to do—I'm putting words in your mouth—but a manual integration where maybe you would have done an automated one. The idea that sometimes if you are continuously focused on improving performance, you won't think of that hiring of that person as the end of what you've done—it's a step along a journey to actually continue to get better that may lead you down a path to automation. But you're fundamentally about performance. You're not confusing the means for the ends in this case. I think that's a big challenge for people right now because they're faced with this blizzard of things they could use and assume that, if they pick the right one, that all will be fine.

So shift gears a little bit. Tell us a little bit about your teams. What did they look like? How did you organize them? How did you develop them?

SANDERS: Sure. Maybe this is surprising but I've actually had fairly traditional structures on my team. And so, for instance, at Casper, where I was the SVP of growth analytics and e-commerce, I had five direct reports. I had a head of performance media, a head of CRM, a head of analytics, head of digital product and an e-commerce leader. That's a very, very basic sort of functionally driven approach.

But for me this worked because the skills that made people really successful with each of those disciplines were different. But what that made happen is, that put the burden on me and my leadership team to act as that integration layer between the disciplines. And so I think we did a pretty good job of that. Because again, particularly at those junior levels, there's such different skill sets it's hard to find generalists across all of those, until you get relatively experienced.

From a development perspective, my personal philosophy is one of servant leadership. And so I can only be successful if the people on my team are successful. And so something that I spend a lot of time on is being accessible to the team. So having weekly office hours that anyone on the team could show up at, building team culture, working to proactively connect the dots between disciplines on the team.

And I also spend a lot of time getting to know the folks on my teams, not just my direct reports, so that I can have a sense of what their development needs are when additional projects or opportunities come up and can send the right things their way.

BREA: Let's shift gears a bit and look ahead. So the pandemic is driving e-comm channels from basically being one of the channels available to be kind of the dominant place in the line. It used to be that a physical retailer would say, well, e-comm is my biggest store, right? But now it's much more central than that. What are some things that you think multichannel retailers can still learn from the digital natives? And conversely, even if they don't have physical stores, what are the things you think some of the digital natives with booming businesses can actually learn from the more traditional folks?

SANDERS: Sure. Well, for multichannel retailers, I think it's particularly important in this period in time for the digital experience to be top notch. And these days your website, maybe more than your stores, needs to be the face of your brand. And frankly that applies to your mobile app too. And that's the experience by which customers will judge your brand.

I just saw some news this morning that customers are much less likely to go into a store for Black Friday, Cyber Monday than they have in previous years.

BREA: Surprise, surprise. Yeah.

SANDERS: Already I'm not sure that I would have wanted to run into a store at 5:00 in the morning but if that's ...

BREA: Right.

SANDERS: If that's the consumer mentality, that digital experience has to capture all of that demand that normally would have been waiting for doorbusters on the curb. And for digital natives, it's important to think about the purpose of a store, a physical store. A store is transactional, but for a lot of folks a physical store has been a place where people want to come and spend time, right? And do so again and again, and you see that in some of the experiences that traditional retailers are putting into their stores. And certainly what some of the mall owner and operators are putting into their malls.

And so that's getting people to come in and spend time and do it again and again and again. And so, how are you going to do that with your brand, in a world where you don't necessarily have a physical experience? And yes, it's possible to do an experiential pop-up, but I don't know that that should be sort of the future of the business.

And so if it's not going to be a promotion, and it's not going to be a Facebook ad, what is it going to be? And I don't claim to have the answer, but I think it's an important question.

BREA: That's a great point. One of the things I'm struck by on this point is how small businesses actually sometimes lead the way. They include things like videos that show the product and use. They explain how the product is assembled or can be repaired. Or any number of different sort of things that sort of show it in the wild.

And that sometimes is actually a form of entertainment for the people who come to shop, because they can sort of see themselves in that experience. They can share that, they can get comfortable and familiar with the product in a way that actually, when it comes time to actually press Buy, they do.

So I think that's a really interesting point you raise about the role of the store. So a question I always ask, how do you keep up? What do you read, what do you watch, what do you listen to?

SANDERS: So from a marketing perspective I tend to be maybe relatively old school. So I read a lot of newsletters in the space.

BREA: Yeah.

SANDERS: And some of which I've been subscribing to for years. And then the new ones coming in or filtering out. So some of my favorites are some of the MediaPost newsletters, AdExchanger, Digiday and even Harvard Business Review's Management Tip of the Day. That's someone who thinks a lot about the team, that's important. In terms of what I watch? A lot of football. And I'm excited for football season to start in two days.

BREA: And a very, very strange and different way this year, that's for sure.

SANDERS: For sure.

BREA: I'm glad you mentioned AdExchanger, that's a favorite of mine too. The podcast series, I think Zach does a really great job with that. It's a little inside baseball-y, but if you really want to keep up with what's going on and understand it, that's the place to be. So finally, we're really grateful to have had the chance to work with you as an adviser. What's next for you?

SANDERS: Yes, so I'll be starting a new role in the near future and I can't wait to tell everyone about it.

BREA: Very cool, we'll come back and hopefully, if you're game, do another one of these conversations when you're settled there. But advanced congrats on that.

SANDERS: Thank you.

BREA: Geoff, thank you so much for taking the time to do this. I know you're very busy and we really appreciate it, but I think just hearing from people with your very hands-on, practical frontline experience is incredibly valuable for everybody. And we really appreciate your making the time.

SANDERS: Well, thank you, Cesar. I appreciate it and always happy to do one of these.

BREA: Fantastic. All right.

SANDERS: Thanks so much.


Ready to talk?

We work with ambitious leaders who want to define the future, not hide from it. Together, we achieve extraordinary outcomes.