Kevin Cuxil, Vice President of Marketing Technology and Marketing Science at Fox Corporation joined our podcast, In the Clouds, for a conversation about Fox's current MarTech stack, including why they have multiple CDPs, how his team works together, and what the future holds for them.
BOBBY TICHY: Well today is a very exciting episode because we've got a special guest who we've worked with for a little while, and gotten to know and understand his marketing technology stack. We've really appreciated a lot of the different elements that they've implemented and brought together. So Kevin, if you wouldn't mind do a brief introduction of yourself, we'll jump in from there.
KEVIN CUXIL: I'm Kevin Cuxil. I'm Vice President of Marketing Technology and Marketing Science for Fox Corporation.
BOBBY: What does marketing science mean?
KEVIN: We realized that in addition to the marketing technology stack that there's some heavy analytics that needed to happen on that backend data. And so what we've done is partnered with our data science team and we have some analysts, our group of analysts who do this specialized tuning. I'm going to talk about it a little bit more in detail as well, but things like predictive analytics in terms of engagement that then feed back into Salesforce Marketing Cloud. So really a group that's a hybrid between data analysis and data science that connects directly to marketing data.
BOBBY: What's been your overall strategy for building your MarTech stack?
KEVIN: I think we really had three main focus areas. The first is we had to be omnichannel. We have to do email push, in-app, paid, social, add server integration — all of that needs to be coordinated. I think the second piece for us was we wanted to build an in-house team of experts. So, while we did deployment and partnered with Lev and consultants from time to time on jump starting certain special projects, we made sure that knowledge transfer takes place at the internal team. We actually have a former Lev person on one of my teams right now. So that's always fun to talk about. I think that the last piece is the raw data and being available in our data mark for the marketing science team and then the data science team to start to leverage and then pipe back into the platform.
BOBBY: Just a follow up to that, on that first point, talking about the omnichannel piece of it and the orchestration. What tool do you guys use, if you can share, to leverage or to actually build out the orchestration?
KEVIN: We're using Journey Builder pretty heavily in that regard. Everything from onboarding journeys have gotten more and more sophisticated with some AB testing for things like Fox Nation to now, even our predictive analytics that we're doing. We're predicting user engagement based on their activity, but we're building out different experiences for those users based on what type of next best action we want them to take, whether that's downloading an app, or getting another video start in some type of news series.
COLE FISHER: I'm curious to see what you've seen change over the past 10 years, especially with your MarTech strategy and thinking, especially lately, with this big shift of third to first party data. How has that had an impact for you guys?
KEVIN: The first party data has become obviously a much bigger thing over the last couple of years, trying to solve for the third party cookies going away. I think we've gone from an approach where we're reacting to RFPs. Really in the old days, especially when it comes to the ad sales side of things, we're reacting to RFPs from our customers and simply trying to say, “Okay, what data can we get and how can we improve the value of that data or the reliability of that data from a third party vendor?” Now, I think we've learned that we need to stay ahead of the curve, and so we really took some time and really invested over the last couple years to go state of the art with our DMP, CDP, CRM, and all of that stuff together.
We're piloting things like Google FloC even though that's not a thing anymore. We are piloting that in their origin. We've started to pilot some of the stuff. We look at Topics and understand how Google Topics is going to affect the ad ecosystem, how we can blend that with our first party data. So we're starting to look at some of those things, so really trying to stay ahead of the game as much as we can there with both the strategy, as well as the technology.
BOBBY: Do you have a team that focuses on experimentation, whether it's in new features or within that data science group?
KEVIN: Yes. So we're looking at a lot of that stuff. Personalization and recommendation is really a big focus for us in this upcoming year. We've deployed a single-armed bandit model on a couple properties now, working on making that a multi-arm bandit, meaning multiple things, taking into account time of day, users, historical behavior, etc. So really trying to build that into our products. That obviously takes some time because we need to tag it with data. We need all the right data points coming out of here. And so the team's really been putting a lot of work into our product team to make sure that we have all the right data coming out into our platform so our models have the right signals in order to make the right decisions. So that'll be a focus area in the next year.
BOBBY: Within that experimentation team, like you mentioned, Google FloC or Google Topics, or rolling out certain tests, or even just start to learn certain new elements that come out, whether that be around first or third party data, or it could be a new CDP — do you have a team or a team member that focuses on keeping their eye on what's happening within the martech ecosystem and then aligning what we should be looking at? Because the way I've heard about it from other companies your size is they'll have what are called innovation sprints, where they'll take a particular tool and they'll spend two weeks evaluating it, and how it could apply to your business.
KEVIN: About a year ago we moved to a scaled agile framework or SAFe. And so at the end of each three month increment, we actually reserve two weeks for innovation and planning. The first week is really that innovation piece that you're talking about there, where we're working with our vendors to look at the ecosystem, see what's next, see what we can pilot. And then the second week, of course, is spent a lot of planning for the upcoming program. It's not necessarily one person, but as a group we come together, we have an architecture group, and talk about what we want to test and leverage during that innovation phase.
BOBBY: I think that's such a big thing. And Cole, we talked to so many different clients who are not doing anything like that. It's really just focusing on whack-a-mole. What's the next priority based on what people internally want. And I think a lot of times, and I'm sure Kevin you've seen this too, is marketing folks typically have a really hard time aligning to some kind of a methodology or framework.
Within marketing, there's a lot of creativity, there's a lot of content people, there's typically a lot of folks who focus on the humanized element of marketing, which is certainly important, but not on the development side of marketing. It's always that ongoing battle of who owns the technology? Is it marketing or is IT? Is it its own function? So it's really interesting to hear that you guys have a specific two week sprint essentially at the end of those three month increments, just for innovation and planning, which is pretty cool.
KEVIN: Yeah. It's been really great. We put a lot of focus on personal development training also for our teams to make sure that they're upskilled and up to speed on all the latest stuff. And so I think for years we had exactly what you described. We're chasing the mail truck. What's the next thing? Where do we get? And we made a decision a couple years ago that this needs to be a priority, and we needed to get ahead of that. That was really in partnership with our ad tech team as well, to make sure that we're on the leading edge. Otherwise, we're not there. We're simply subject to everyone else's movements in the marketplace, which is not where we want to be.
COLE: In your current stack you have multiple CDPs. Can you dig into, A, why have multiple CDPs, and then, B, what different functions those serve?
KEVIN: Yeah. I think you have to always step back and define what a CDP is, right? Because it's such a broad thing. I think so many companies sitting on a lot of data said, oh, well, we're close to this, we should be a CDP. And others are more purpose built, I think, in that regard. So yeah, we have Tealium, we have Segment, we have Lytics all as CDPs in our platform, but we really focus on the strength for each one. Segment is really an amazing data orchestration tool. It has some really advanced features in terms of even some live data transformation, schema enforcement. So really that collection from all of our platforms, it's a really great tool for that.
But when it comes to data collection around the user and we wanted to ask them explicit questions and surveys and polls and quizzes, it wasn't the ideal tool for that. We use Tealium tag management. It's an amazing tag management platform. It's very scalable. We get in there and all of our pixels are fired from in there in a privacy compliant way. So that's really an amazing tool on that side. And we went with Lytics as our main, what I'd call CDP, basically because it was out of box functionality. It allowed us to very quickly deploy widgets across websites for data collection, whether that's an email address or a survey or a quiz. Those are all really easy experiences to get out of the box really quickly and it gives a lot of power to marketing teams where they aren't relying on dev teams to deploy code.
BOBBY: It's really interesting. And one of the main reasons why I wanted to have you on the podcast was around the CDP topic. Because, I think, there's so much confusion. It's like what marketing automation was probably 10 or 12 years ago, where it's still a fairly immature ecosystem at this point. There's a lot of different players, there's a lot of, what we would typically think of as marketing clouds saying that they have a CDP, there's a lot of immature CDPs out there. And I think too, I saw an infographic from Simon data around, there were four different pillars of what CDPs could potentially be, and so I think going back to your first point is really, you've got to start by defining what your business goals or objectives are by purchasing this piece of software.
To your point around data orchestration and being a huge fan of how Segment does that or user data collection with Lytics. It's really important to identify, first, like everybody should before they buy any piece of software, what is the business value or the business objective that I'm trying to accomplish with this piece of software, with this purchase? And then going from there. You've already mentioned it a couple of times with Journey Builder, and obviously we're a big fan at the Salesforce marketing cloud. But what's your favorite thing about Salesforce marketing cloud?
COLE: If budget, time to value, and implementation, weren't considerations, what would be the wish list item that you would implement right now?
KEVIN: Definitely more automation of our models. I would speed up our personalization recommendation stuff. We've built some really amazing models. The teams currently built predictive engagement based off of things like what devices have you installed on? Have you installed it on web only? Are you also on living room or mobile? What type of content are you interacting with? Are you interacting with some of our daily shows? Are you interacting with movies? So we've got these models that are actually predicting people's engagement over the next two months, and so what we can do is actually tell whether or not a cohort or an individual is healthier or unhealthy. And then we can send the signal into SMC as to the reason for that. And then they'll build the journeys. We're trying to automate as fast as we can. We're in process, but automating that stuff always takes a little bit of time. Obviously cleansing the data, make sure it's the right way so that we can adjust it in the Salesforce marketing cloud.
BOBBY: How big is your team?
KEVIN: That's a good question. I think we're around 13 people right now across both groups, 13, 15, somewhere around there.
BOBBY: Do you guys handle all of the development? Are you managing all of the development and running the business operations, or are there any other teams internally that you work with to help with that?
KEVIN: No, we pretty much own our own platform. So I partner very heavily with data services and my teams are in those Redshift databases on a regular basis, but that's all in house that we've done. Again, depending on the product we'll definitely partner with a Lev or consultants on different platforms to do improvements. But for the most part, we've really found a lot of value in bringing that knowledge in house.
BOBBY: I couldn't agree more. I think that's where most folks find the most amount of value is having folks internally who can help run the business or innovate upon whatever that might look like. I also think it's great that you actually have ownership over these platforms like Salesforce Marketing Cloud, or Lytics or whatever it might be so that way there's not this jockeying internally of having to reach out to another team, because that can be just as bad as having a third party consultant run those different things for you. So it's pretty awesome that you guys have full reign of not only the team internally, but also the platforms that you're interacting with.
KEVIN: Yeah. We're very fortunate in that regard. I think we made a conscious decision to make sure that we owned. We've had a lot of stops and starts using different scenarios where we didn't necessarily own all the data or the platforms, and I think we learned a lesson. So I think we've gotten to a good place.
BOBBY: Last question. What's your biggest lesson learned from your time in the martech ecosystem?
KEVIN: I think it's interesting because there's so much overlap in all these tools. One tool started with emailing and grew out, another tool may have started with mobile analytics and then added on. And so really understanding your use case and what you're trying to achieve and going deep on each of the features and functionality. The same as what we see with the CDP landscape, three different CDPs, really going deep to understand how do those specific features within each of those tools map to what you're trying to do, long term? And then trying to pick the best ones there because there's so much overlap in the space. I think that's the biggest challenge.
To listen to the entire conversation, check out the podcast episode below.