How To Think About Big Data For Marketing : Interview with Cory Treffiletti

Interviewer: Today I’m pleased to welcome Cory Treffiletti, Senior Vice President of Marketing for BlueKai which was just acquired by Oracle. Congratulations Cory.

Treffiletti: Thank you very much Glenn.

Interviewer: BlueKai does big data for marketing, so Cory has been a thought leader in the digital media landscape since 1994; and as author of weekly column on digital media advertising and marketing since 2000 for Mediapost. That in and of itself is pretty impressive. He’s a graduate of Syracuse University and is the author of Internet Ad Pioneers which you can buy on Amazon. So, welcome Cory.

Treffiletti: Thank you. Thank you for having me.

Interviewer: So, Cory’s going to help us understand what big data for marketing really means and why it’s so important to any marketer. So, Cory, how should we start thinking about big data for marketing?

Treffiletti: So if you go back in time a little bit, most marketing has always been about a big idea and about creativity. And a lot of people jokingly refer to the fact that they got into marketing and advertising because they hated math.

Interviewer: Right.

Treffiletti: But the fact is that in the last, I would say in the last 10 years, especially in the last 5 years things have start to shift to accountability in marketing. And more marketers are looking for ways to prove that the dollars they spend has a measurable ROI. And they’re all trying to figure out how they can, not only justify continuing to spend in marketing, but how do I justify increasing my marketing spend. And to do that you have to be able to show proof. So, –

Interviewer: That’s right. That’s right and they’re getting a lot of pressure from senior management to do that.

Treffiletti: Oh yeah, because otherwise marketing is an expense, and it’s one of the first things that gets cut when budgets are being reduced.

Interviewer: Absolutely.

Treffiletti: So what’s happening is that marketers are looking at different ways to prove the value that they bring to the table. And in the last four or five years what you’ve seen is this rise of programmatic media, and more technology being applied to marketing and advertising that can enable the automation of buying and selling for media, which is a really great place for things to start. Now what that’s done is that brought us to where we are just in the last probably 12 months and where things are heading in the future, which the development of the enterprise marketing solution set. And when people in marketing and advertising read about technology and enterprise solutions and all those things, it’s very, very complicated very, very quickly. But what I talk to marketers about; and what I talk to our clients about is really about how to simplify the understanding of the landscape is for the enterprise solution set.

Interviewer: Okay.

Treffiletti: So the best way to do that is to think about it in three tiers. There are three tiers of an enterprise solution for marketing. The first tier is a data layer; and the data layer is really where all your [00:03:40 – Audio Break] goes where you can make sense of that data and you can activate that data, and you can put that data into use. The second layer of that enterprise solution set is the execution layer, and that’s a lot of the platforms and the channels that most marketers are most familiar with. Those are Yahoo and AOL. Those are ad networks and DSP’s. That’s Facebook and Twitter. It’s really where the message meets the audience.

Interviewer: Um hm.

Treffiletti: And that layer is very fluid. But when you spend money with Facebook now maybe you didn’t five years ago. And five years from now you’re going to go spend money with somebody who doesn’t exist right now. And DSP’s are going to change from time to time. So, there is a fluidity there. Whereas, that data layer you tend to pic a couple of different partners and you’re going to stick with those partners for the long term.

Interviewer: So, a couple questions. First of all define DSP’s for us. And then also when you talk about the execution layer I’m hearing you talk about companies other than the company the marketer’s responsible for. So, what about doing it internally as well?

Treffiletti: Sure, those are good questions. So, a DSP is Demand-Side Platform. And that simply refers to an aggregation of display advertising online.

Interviewer: Um hm.

Treffiletti: Now to answer your second question, why aren’t marketers talking about or thinking about their own platform. For the most parts there’s a way of breaking down the media landscape that people call owned/earned and paid media.

Interviewer: Right.

Treffiletti: So, owned media would be their own platform, so their own websites; their own email lists, etcetera. And they absolutely should be thinking about that as part of the execution platform for marketing. I think what we do is we tend to focus a lot on the earned and paid components because they’re significantly larger for every brand in the world.

Interviewer: Um hm.

Treffiletti: You could be the largest brand in the world and you cannot completely rely on only your own properties to be able to interact with your audience.

Interviewer: Got it. Good, okay.

Treffiletti: So, just to finalize that first analogy though. So the third layer of what the enterprise solution set looks like, that’s the analytics layer. And I would say that’s probably the least developed layer so far. That can be comprised of everything ranging from your ad agencies who typically pull together performance reports for all of your marketing, to attribution vendors and media mix modelers, and look alike modeling companies, and anybody who is out there trying to figure out how to dive more into the analytics side of the business to enable you to reach your customer.

Interviewer: Um hm.

Treffiletti: So, what we’re seeing is that large-scale fortune five hundred marketers are now starting to invest in developing their own enterprise marketing suites. And what they’re doing is they’re looking at those three layers and trying to figure out which kinds of partnerships they need to look into, and how do they knit those partnerships together to create their own solution.

Interviewer: When you say partnerships are you talking about vendors that are selling them a solution?

Treffiletti: In many cases, yes. On an even higher level I would say a lot of brands are looking at the question of build versus buy.

Interviewer: Um hm.

Treffiletti: And what they’re seeing is that to build these platforms is a significant cost and you have to maintain a level of innovation because it’s an immature category. Meaning there’s still a significant amount of development to take place before these things fully mature.

Interviewer: Right.

Treffiletti: So, I think what you’re seeing is that marketers are realizing that the investment is futile, because they can’t keep up with the investment, so they’re all leaning towards buying or renting a this point. So, you’re seeing companies like ours and a lot of other companies provide SaaS models for marketers so they can utilize these in the cloud and that platform can be maintained at a level of innovation and a rate that can keep up with the market.

Interviewer: Right, right, and would you say that the cost of switching is lower if I am using a cloud-based solution?

Treffiletti: That’s a good question, and you know the best way to look at that is look at the three tiers. So, if you work your way backwards, the analytics layer and the execution layer are very fluid. So, the cost of switching there is actually quite low and you’re very right about that. If you do this right though, the data layer, the partners that you select in that; there’s actually a really high opportunity cost to switching, because what you’re doing is you’re creating a centralized set where all your data is going to go.

Interviewer: Right.

Treffiletti: You’re going to utilize that data to build custom audiences. And you’re going to utilize those custom audiences across all the execution that you do.

Interviewer: Right.

Treffiletti: So, there’s an investment there that doesn’t make it very easy to switch. You actually want to pick the right partners from the get go there, because you want to work with those partners probably for the next 10 years.

Interviewer: And on that point, I know I think about a single view of the customer which is something we’re hearing a lot from our clients. How do we gain that single view of the customer and if you operate and manage and focus on the data layer you can build that.

Treffiletti: Well, one thing I would say is where I think you were going is that the single view of the customer it’s very difficult to maintain that across channels. And specifically because of desktop and mobile, there are a lot of things going on in that market and it really revolves around the third-party cookie.

Interviewer: Right.

Treffiletti: And in mobile third-party cookies are not really reliable because applications don’t use them, and most browsers don’t use them. The only browsers that do are android browsers. And if you look at the larger scale picture, a third-party cookie is most likely going to be going away, especially in the next year, maybe 18 months. And it’s going to be replaced by different identifiers that different companies are going to put out into the marketplace. You’re already seeing that from Mozilla. You’ve seen Google talk about it. Microsoft has talked about it. A lot of people are trying to find ways to identify who their users are. So what happens is that data layer becomes even more important, because what you have to be able to do is knit together a complicated view of who your customer is across channels.

Interviewer: Yes.

Treffiletti: That’s where data management platforms and a lot of the core components of that data layer, that’s where they become increasing important, because it’s going to be more difficult to understand who your audience is across different devices and channels and platforms. And yet, that’s probably the single most important component of marketing accountability. You need to understand who your audience is so that you’re not showing them messages which are untargeted; they are not seeing messages which are redundant. You know advertising gets a bad rap because advertising is viewed of as being intrusive or annoying to some people. But it would be significantly more annoying if it was completely untargeted.

Interviewer: Perfectly true, yeah. I totally agree. So that raises the question of data integration in the data layer. Can you talk a little bit about that?

Treffiletti: Sure, the way that I would approach it, and the way that we tend to approach data integration is that you want to look at data from multiple sources, bring it into a single platform, and then be able to identify which of the data points are the most valuable. So there are priorities within the data world, because you can amass a ton of information about [00:10:28 – Inaudible] users, but not all of it is relevant. You know there are things like recentcy. There are things like repeatability that have to be factored into it. So what happens is that you have to be able to find a way to prioritize all that data, and that comes from what we call custom classification; and being able to classify and prioritize all the data that’s available. Once you’ve done that, and once you’ve been able to build audience segments that are based on the highest priority sets in that data, it becomes important to be able to push that out to all the execution channels that you work with. And that is one of the things I think marketers are really starting to wake-up to, is that I don’t want to just use data and integrate for the use in banners. You know banners are a small portion of my overall budget. I want to use it to optimize my website when a customer comes.

Interviewer: Right.

Treffiletti: I want to use it to optimize the experience with my CRM. I want to be able to utilize it in mobile. I want to be able to utilize it in social and on video, and eventually on platforms like connected TVs and even in the dashboard of your car; and all the other things when they refer to the internet of things. Anything that’s going to be internet enabled should become, theoretically, addressable to an audience. And you need data to be able to do that.

Interviewer: So, it strikes me that at the data layer while I might rely on a lot of vendors to help me, ultimately I will customize my view of the customer for my particular way of going to market and impacting that customer.

Treffiletti: Um hm, I think that’s fair. I think it’s all about customization. It’s actually, customization is funny because it’s on both sides. You need to customize how you talk to them, and they need to be able to have customization of what they see, because when they come back to you, their responses and their interactions with you are really a one-to-one interaction. And you need to be able to take that and utilize that at that level.

Interviewer: So, can you talk a little bit about how some companies are using this effectively?

Treffiletti: Sure, there are a lot of different categories of companies that are utilizing it. I can tell you that a lot of travel companies. A lot of the Telco’s, you know the mobile carriers; those folks are doing a lot of great things as far as data. Retail companies are doing a lot of great things that we’ve seen as an example, a large Telco in the U.S. And that Telco realized that they were doing their acquisition-oriented advertising through banners and display online and what they were doing is they were showing acquisition messages to everybody. And fundamentally what they were doing is they were plastering they’re current customers with acquisition messages, which is a real waste of money.

Interviewer: Right.

Treffiletti: So, one of the things that they’ve been able to do using data was to suppress their existing customers right off the bat, which immediately gave them an efficiency in their media outreach. So, they were being able to optimize the dollars that they had against new customers, not just crushing their existing customers with the same message over and over.

Interviewer: And probably annoying them while they’re doing that.

Treffiletti: Oh, definitely. But what happened even going beyond that is that we found a secondary use case which popped up as a result of that; was as they started to dive in they realized that on their new customers that they were acquiring there was a level of churn over the course of the first 30 days of that customer coming on board. Where someone would buy a phone and they would be less than impressed with it and they would go return it. And that was actually costing them $20 or $30 million dollars a year in lost revenue, because people were returning the phones and cancelling their contracts within 30 days. So what we did is we worked with them to take that suppressed customer list; identified those customers that had only bought within the last 30-days and target them with retention messages either through display or through CRM. And what they were starting to see was accessory ads, or increased data campaigns, or they would see different things they could be doing with their phone by accessing an app store and stuff like that through their phones. Well, what they saw was their churn rate decreased dramatically. So, they started to save money by making sure that they were turning down churn and retaining customers, in addition to having increased the efficiency of their acquisition outreach in the first place. So, they kind of got a double whammy, and that’s a really great use case, because you’re starting to see that a really simple execution of what data can do on both ends, not only bringing new customers in, but helping you retain and increase the value of existing customers.

Interviewer: That’s a really powerful story.

Treffiletti: Yes.

Interviewer: Cory, can you share any stories from the B2B world?

Treffiletti: Yeah, it’s actually it’s a funny area, because as a marketer I know I work for a company that offers a big data platform. But I’m also a B2B marketer. My job is to reach out to Fortune 500, Fortune 1000 level companies and speak to them, so I can tell you from first-hand experience where we actually eat our own dog food. We employ a DMP [00:15:09 – Inaudible] off our own website. We utilize that to help us identify people who are coming to our pages and then we target them. We use that to cross-reference our sales force named customer accounts that we’re trying to go after. And when we identify someone who comes to our site from a domain that’s representative of one of our target accounts we take that information and we reach out to those folks and say, we understand that you might have been looking for information on this topic, how can we help you. So you start to see that B2B marketing is actually really, really effective because whether it’s B2B or its B2C, it’s really all about audience. B2B audiences are typically smaller than an average B2C audience, so you can get more granular in the data that’s available to those folks. And you can be just as targeted in your campaigns, whether it’s doing retargeting CRM or even social.

Interviewer: Excellent, excellent, yeah I read recently that the B2B marketer is actually the same kind of emotional buyer as a B2C buyer. And we haven’t always thought of them that way, and if we can appeal to them through the right kind of messages and make them feel good about us as a brand, then we can have a real impact if we can communicate exactly what they need, when they need it. Almost like the Holy Grail here.

Treffiletti: It’s funny too. You think about it, the B2B audience is the B2C audience. They are consumers. They just happen to be consuming it in a business setting, so the motivations are the same. The decision making process is very, very similar. You know people talk about in a business setting, spend money like it’s your own. And I think that comes through in more ways than they even thought possible, because what you’re doing is you’re tasking the person in a business environment to spend their corporations funds, whether it’s marketing, or it’s IT, or whatever it is, in the most, I guess most ethical and honorable and effective way possible, which is exactly how they would spend their own money.

Interviewer: So, talk to us again, you talked earlier about how cookies are essentially going away. And so how can companies effectively know who the business visitor is? Whether I’m going to the company website or I’m going to a third-party website, how do you really understand both who I am and where I am in that buyer’s journey?

Treffiletti: Well, one thing to keep in mind is that it’s not that cookies are going away, its third-party cookies are going to go away.

Interviewer: Ah!

Treffiletti: So, there’s a big difference there. First-party cookies are still totally approved, very valid, and there’s really no threat against first-party cookies being utilized in the future. So, when you have a website and or CRM list that you work with and a person comes to your site, they register with you; maybe they get access to some information; that’s all first-party data. And that’s well and good and will continue to be utilized. And I think what you’re seeing is that more and more that’s the basis for how either B2B or B2C marketers move forward.

Interviewer: Okay.

Treffiletti: So, the way to think about that is that every brand in the world is looking for a competitive advantage. First-party data is the first place that you start, because what you know about your customers and your prospects is exclusive to you. Now most marketers will then go look at third-party data. And third-party data is a non-exclusive advantage. Everybody has access to the same data. So what happens is that between first and third-party data is this realm of opportunity that marketers, whether they are B2B or B2C, are starting to explore. And the best two examples there are what we call second-party data, which is two companies that are non-competitive sharing data about their customers to create a semi-exclusive advantage.

Interviewer: Um hm.

Treffiletti: So, you might see a credit card company and a hotel chain go and share their first-party data on an exclusive basis with one another, and create a hotel chain themed credit card.

Interviewer: Um hm.

Treffiletti: And so you see a lot of those out there in the marketplace, and those are co-marketing, or you know they come under other names, but we consider it to be second-party data that’s being shared. So, that’s a high value area.

Interviewer: Great example.

Treffiletti: Thanks, the second area that you are starting to see value is in modeling, which is taking first-party data then augmenting that with third-party data so you can create what we refer to as semi-exclusive competitive edge. Meaning that if you are building a model off of first-party data that you have that nobody else has, somebody else may create a model that’s similar, but it’s never going to be the exact same.

Interviewer: Um hm.

Treffiletti: So, what may happen is that you and your competition are going to know slightly different things about your audience and you’re going to utilize that to go reach out to them. And you’re each going to come to similar audiences and you’re going to be able to target them in different ways. And that’s where you start to see scale off of first-party data, because first-party data is finite. But if you can model off of that data you can get much, much larger.

Interviewer: Wow, so one of the things that strikes me is that most marketers don’t have the skills to implement this quickly and easily, because this is new and it’s not typically what they are hired for. And yet, it’s becoming extraordinarily important. So first of all, do you agree that that’s one of the key obstacles to implementing this, and, if so, how should companies go about addressing that obstacle?

Treffiletti: That’s a very valid concern and very valid question. And I’ve heard it emanate from the mouths of a lot of different marketers; CPG marketers, travel marketers, etcetera. Where they’ve literally sat down with us and they’ve asked how do I train my existing team on the skills that are going to be required, and how do I identify the right partners, and vendors, and platforms to be able to do this? And my response usually is that there’s an evolution that goes on, and it happened almost 20 years ago when digital started to bloom in the first place. It’s happened 10 years ago after the dot com bubble when digital started to take a strategic seat at the table. And it’s happening now at this renaissance for data driven marketing. And what you’re doing is you’re identifying that there are partners out there that you can lean on. Those partners can become an outsource extension of your marketing team. And what they do by proxy is they become a training grounds for your internal teams. So that’s one great thing is that don’t intimidate the existing teams inside of marketing organizations by telling them there’s a whole lot of stuff they don’t know. Find the right partner that they can work with, who becomes a valued partner over time. And you enable those partners to learn from one another; which means you have to tolerate some mistakes being made along the way in order for everybody to get to the same position.

Interviewer: Um hm. Um hm.

Treffiletti: Now the second thing is that the partners that are out there, you know the DMP’s and the DSP’s and all those things that we’ve kind of mentioned already, they were originally developed with IT people in mind. But what we’re starting to see is that as marketers are adopting these tools, the tools are becoming more what marketers are looking for.

Interviewer: Um hm.

Treffiletti: So, we’re starting to see a lot more of a user experience that looks like a marketing tool, a platform with a dashboard and graphs and visuals. And the kinds of things that are important to marketers. So, as your internal team starts to become familiar with what the landscape of these tools are and the tools themselves start to become more familiar with the lexicon, and the terminology, and the things that marketers are looking for, what you’re going to have is that time will progress and that marketing teams are going to become extremely comfortable with these tools. And I think that’s where we’re headed right now.

Interviewer: Yes, that’s what we’re seeing. It’s probably a combination of external resources with those tools who together can transfer knowledge and make it easier for the marketing teams to pick this up and run with it.

Treffiletti: Yeah, many of those partnerships are going to be either agencies or consultancies that know about these tools and platforms and can come in, train, and then when they walk away that existing marketing organization is better for having had that experience, because they understand it now, and they’ve been trained on it. So that’s definitely what can happen. You know if you go back to when TV was first bought and Donovan Systems came in, people had to be trained on that. But it just took a little bit of time and they build the expertise over a period of years. That’s exactly where we are right now.

Interviewer: Great, wow, well this is very exciting. Any last thoughts for us, Cory?

Treffiletti: I guess I would leave with one thought, which is that our industry has a horrible history of trying to intimidate everybody by making things too complicated to justify our own existence. And we’ve done that for years. For the last 20 years the digital media business, the digital market industry has tried to overcomplicate things. It’s really not that complicated. That’s why it’s easy to break down what an enterprise suite looks like on those three tiers; data, execution and analytics. And all the things that a marketer has been trained to do for the last 60 years is still very applicable. It’s always about knowing who your audience is. The only difference here is that 20 years ago you would spend money and you knew in aggregate it worked, but you were unable to identify which of the individual pieces of the pie were most effective. Where we’re headed right now is that maybe 75% of that pie you can deem if it was effective. And that’s not a scary thing. That actually should be like a light bulb, which is you should feel more confident that what you’ve been doing, it is working. But now you can actually remove the waste and you can focus on the components that work the best. And it doesn’t mean you have to get rid of the other 25%. It just means you have to feel more confident in the 75% that you’re able to track and utilize. And I think that if every marketer understands that, that this is going to help them in the long run it should remove the intimidation factor.

Interviewer: Well I hope it does Cory, and you’ve been very helpful in taking that first step for helping people understand the value of big data in marketing and how to take those steps.

Treffiletti: Well, thank you for having me. This has been fun and I hope that a lot of people will get to listen to it and I hope it helps.

Interviewer: Alright Cory, thanks so much.

Treffiletti: Thank you.