Segmentation insights

MM: If I understand you right, Trae — the first order of business is really to develop some segmentation insights.

Right.

That entails building or acquiring a clean, enriched database of households or businesses. Then using that as one version of the customer truth — or one version of the market truth, anyway. Then mapping all of this irregular — although real — data from our existing data sources to this customer master.

Exactly.

MM: Then as a function of that, once we’ve established the notion of a customer, a buyer and a stakeholder that’s part of that overall buying organization — be it a household or a business. Then we can start to lay in additional sets of data — vis-à-vis — web analytics or e-mail messaging analytics and so on — so as to make that overall customer profile clearer and clearer. Or, as you said, the fingerprint goes from a partial now to a fully-developed fingerprint, in terms of who this individual is.

Exactly. That’s exactly right.

About multichannel marketing analytics

MM: Before we get into the e-mail messaging and personalized messaging like that, I’d like you to talk to us a little bit about the notion of multichannel marketing analytics. Specifically speak to some of the web analytics that are now coming into this overall customer insight. Search analytics. Social media analytics. And perhaps even voice-of-the-customer content analytics.

Absolutely.

Again, as you know, that’s where I think just philosophically, we align so well with Alterian. One of the things that we’re so excited about in terms of the technology that they’re developing is that alignment.

I think that again — back to this idea of tracking consumer engagement. Everybody knows and has known forever now that… Allow me to just step into “social” for a second. Everybody knows that word-of-mouth is real. It has been real for a long time. Marketers have known that.

But how, historically, do you leverage that or are you even aware of that? It’s very, very spotty — at best. So much so that traditional word-of-mouth efforts have been very limited in scope, typically. Their measurability is highly questionable.

But the exciting thing about social, for example, is that now that’s happening increasingly online. It’s happening in a context where it can be… Depending on the exact execution, understood — listened to — tracked — even controlled, in some cases.

Again, back to that philosophy of data-agnosticism. I don’t just mean purchase data — third-party data — I even mean target-based views. Channels, agnostically. In the sense that, “How can they help us understand the consumer better?”

We certainly need to realize that, “This came from Channel A, and this came from Channel B.” But it’s just…

MySpace gave WSJ polling tremendous advantage

MM: For example, in this last election, it was clear that citizens under the age of 30 would play a pretty significant role in the electoral outcome.

Right.

MM: All of the polling organizations — Gallup, Harris and so on… They all said, “Well, these guys don’t have phones! They all have mobile phones, and we can’t call them!”. So it threw a whole variable into the tried-and-true polling methodology. It turned out — at least according to the stories that I read — that Fox/Wall Street Journal/MySpace had the most accurate polling data, as a function of all of the hyper-segmentation or micro-segmentation they did on MySpace. They were able to — as a function of their engagement with that youth market — were able to get very, very accurate polling data in terms of who was going to vote. And of the people that were going to vote, the probability of them voting, and for whom.

Absolutely. Yes. That’s a great example.

One of the actual products that was developed at Targetbase many years ago is a thing we refer to internally as “Channel Selector.” The idea again is that starting with the consumer… Let’s say it’s an individual on a client’s consumer database. We actually have a proprietary approach to identifying — again — with the use of various data sources. Identifying or — at the very least — modeling the channels that that particular individual consumes.

For example, whether it’s truly at the high-channel level of television, magazine, online, et cetera… All the way down to — in most cases — specific publications. Even some specific sites.

–

CRM and crap data

MM: Trae — you’ve kind of opened a can of worms. Perhaps by intent or perhaps unwittingly, but there it is. We’ve found in large, well-established brand-marketing organizations — specifically, where they have transaction data — that the CRM system is rudimentary in terms of its data model and its data collection.

Absolutely. I totally agree with that.

MM: One of the most interesting things I’ve found is that they almost make no distinction between the customer, the buyer and the stakeholder. The customer, in a consumer context, is the household. The buyer is the person who actually paid the bill. The stakeholder might be the consumer that influences the buyer — but not necessarily is part of the transaction record. In parallel, we have a similar thing with respect to B2B interactions. I had a conversation with a fellow who used to work at Compaq, and then became part of the DEC merger, in terms of the thing. He said, “You know, Compaq has almost no idea who their customers are.” All they had was a PO with the buyer’s name, and 100 boxes went there. Then over at Digital Equipment — because they were a service firm where they made most of their money from professional services — they not only knew all of the departmental managers and admins, but they knew the birthdays and anniversaries and favorite foods. The whole Harvey McKay 55 questions that you want to know about your customers.

Yes.

Customer DB master

MM: Can you speak to the notion of first of all getting the customer master, and then what you do with that and how use a customer master to support deeper, more meaningful insight about when, where, and how to engage consuming cohorts?

Well, like you say, that’s a very sticky and difficult question. I would say that in terms of knowing your customer, what you describe is very common. We are oftentimes — whether it’s a new business pitch or a new client or what have you — we’re coming to the table because somebody in the organization… probably somebody who just came in from outside… is realizing that, “Oh, my goodness. We don’t even know who our customer is.”

That’s why I mentioned before that we oftentimes begin with something like a segmentation approach. A segmentation project. So we can help that client just understand even in broad segmented terms who their customer is.

Again, from a heritage standpoint, Targetbase certainly began there, and continues to live there. In terms of identifying who your customers are, what they want from a content and a product standpoint — how they want to be interacted with from a channel standpoint, and then also when. When do you want to engage them? The who, what, when, where, how and why types of questions. That’s kind of where we live.

I do think, though, one of the big challenges is, when clients do have an existing database that in many cases they spent a lot of money on, oftentimes — like you said before — the data model itself and the data capture is really lacking. Many times, unfortunately, it just kind of lives off by itself in the IT department. Marketing knows it’s out there. They know that they’ve paid for it or helped to get it funded. But in terms of day-to-day usage and understanding of the consumer, there’s less than you’d hope.

I think, though, in terms of how we try to overcome that barrier — and I could speak at least broadly, in terms of our approach… back to that foundational principle of ever-improving approximation. You always have to start with what you do know, from a database standpoint. Maybe I’ve only got a buyer’s name and address in my database. But there are things that you can do.

For example, similar to the segmentation process, in terms of appended data at the household level. Syndicated data. Trying to map a consumer’s database into a richer data source.

When I say, “Map,” a lot of that again is where I think Targetbase has really broken some ground analytically — in being able to statistically model and identify key variables and key data points that do exist in a consumer database – that also exist in a richer source like a syndicated data source.

Then with a great deal of accuracy, being able to map individuals over. So we can then infer or at least guess what an individual’s attitudes might be — et cetera.

–

Tracking pages like ICBMs from USSR

MM: I’d like to run by one scenario that I ran across a while ago. I’d like you to validate or challenge it to the extent that you have anything to say about it. A large car company was spending a tremendous amount of money creating these immersive, virtual-reality movies of their high-end sports car.

Right.

MM: It was clear from the site traffic that it was a very, very popular feature. Yet most of their preliminary research indicated that some 60 or 70 percent of the usage tracked 14-year old boys. The CMO said, “Hey. I’m all for branding the next generation and our share-determining market sector. But at the end of the day, I’ve got cars to move this month. So I’m really need help in correlating investments in not just these expensive immersive multimedia movies and other sorts of rich Internet applications. But across the 4,000 or 5,000 pages of content that I have on my site, which of those pages do 45-year olds with household incomes of greater than $120,000 consume? And why?

Right.

MM: Classic optimization type of question.

Yes.

MM: In this particular case, the engagement agency went to — I think — R Polk. I guess that’s the aggregator of DMV records.

Yes.

MM: So that would be car registration and drivers license records. They then went to a credit database provider. In this case, I think it was Experian — although it could’ve been any of the other credit-scoring data providers. They were able to extract or develop a database of all the households that had — in the last 7 years — bought or leased a Lexus, Cadillac, BMW, Porsche, Audi — a Lincoln, an Infinity and I think there were a couple more brands in there. They then put that into a high-performance database, underlying the website. When people came to this particular auto site, they basically had to put in their name and zip code. They were able to infer with a fairly high level of confidence that Michael Moon does in fact live in Oakland California, and drives a Lexus GS 350. They were therefore able to track my consumption of media, and able to identify the 1,100 or so pages of content and/or media objects that — in this case — a 55-year old who makes more than $225,000 a year — what pages of content I actually consume.

Right.

MM: Do you see practices like that in place today? If not, why not? If so, how has it evolved beyond that basic model that I’ve outlined?

Actually, in terms of my experience — I think that’s probably a little ahead of the curve for most companies. I think frankly, just the fact that they’re even thinking individual consumer-level identifiable and trackable, online engagement is a bit unusual.

However having said that, that’s clearly the direction that we and Alterian believe you have to go, and the market will be moving toward.

I mentioned American Honda is a client of ours. One of the areas that we’ve really — I think — advanced their direct efforts — their relationship efforts — has been in the predictive modeling area.

Again, I think that example you give is probably advanced, relative to most people. But if I were going to criticize or recommend some additional layers, it probably would be more in the targeting and the right-to-win area. Again, we’ve had great success with that in the automotive area.

Maybe it’s not a fair assumption that if it’s a major auto manufacturer, they’d have records or a database — a consumer database — that identifies their current or previous customers.

A lot of what we’ve done for Honda has been compiling — not just, as you mentioned — from 3rd-party available sources, but certainly connecting that to their known customers. Those on their database that have bought cars in the past.

Identifying the ones that are likely, from a predictive modeling standpoint, in-market for a vehicle — or will soon be in market for a vehicle.

Then, which particular car model best fits that consumer’s needs. Predicting, for example, someone who’s more likely to buy a Civic than an Accord.

When it comes to real-time engagement, sorting or ordering the content or the information that the consumer will be able to engage with — based on what they’re most likely to be interested in. To actually just pushing. Let’s say in my example of a Civic versus an Accord — pushing. “Hey. Here’s a Civic-immersive application you can engage with online.”

That’s really — again — if I were going to criticize, recommend or maybe tout the approach that Targetbase would take that’s a little bit different… It would be more around identifying those that are top prospects. Identifying those that you have a right to win with, and then targeting them specifically with custom content, et cetera.–

–

Revealing the true identity of unknown users

MM: What other techniques-either innovative or just standard-issue data analysis used in a new way-do you use in identifying the true identity of this inferred-unknown user?

I think maybe the only difference or uniqueness would be like tiers of knowing someone. Clearly, when somebody comes into the site and log in, they’re a member or they’re registered. We know immediately who they are.

Maybe the next step down or a lower level would be a cookie-based approach. Then I think probably beneath that — in terms of the level to which we know an individual… We go from trying to identify an actual John Smith at 1234 Main Street, to identifying individual patterns of engagement.

It’s not so much about identifying an individual’s name and address. That’s obviously ideal. We’d love to have that level of knowledge. But when we don’t, the point is that our approach at Targetbase and — again, very consistent with what Alterian does and the way they think… “Let’s know what we can and let’s infer the rest.”

Let’s say somebody comes to the site and there’s not a recognizable cookie on a machine. We have no idea who they are. But as soon as they start doing stuff — as soon as they start engaging with us — we start to get an understanding of who they look like, if you know what I mean. In terms of their pattern of behavior.

Again, back to the idea that we discussed earlier — of segmentation. What group of similar patterns of engagement would we lump this person into, based on the information we have, so far?

The other idea maybe that I would throw out there is  “retroactive identification.” When an individual comes to a site, they allow a cookie to be placed on their machine. Let’s say they engage with us over a period of a few weeks or months. Then they decide to register. Then they decide to sign up.

One of the things from a technology standpoint as well as from an analytics standpoint that we do that may seem simple and straightforward, but you’d be surprised at how many don’t think this way and don’t do this. We can connect that previously anonymous activity to this known person now.

MM: All that unknown data. Yes.

All their previously unknown data.

MM: Sure.

Now that’s rolled into it.

Here’s a brand new registrant — but now we actually do have longitudinal data on them.

MM: In fact, that would also give you a whole set of metrics or insights in terms of how many clicks — how many particular content-consumption cycles occurred by type or class that it took to get somebody at the register.

Absolutely.

Idea of behavioral targeting

MM: Does Targetbase pull any data in terms of these inferred-unknown people or users from ad networks?

Yes. We’ve certainly used that. Depending on the circumstances and the client need.

You get into the whole idea of behavioral targeting. That’s certainly very viable and will become increasingly so. A very viable approach to targeting — certainly ads, and perhaps other things.

One of the criticisms that I have of the standard network approach — behavioral targeting approach — is it’s such a limited view of an individual’s online behavior. Even with really large network properties online. You’re getting a potentially very skewed view into an individual’s behavior — which can still be very powerful from an ad standpoint, depending on your objectives.

To answer your question, we have — in a couple of instances — mapped in some of that data, where it made sense for our clients on a particular ad network or property that would gain additional insight into the consumers on the database, in terms of their online behavior.

I would say that’s more the exception. Simply because when we’re [looking at] that, we typically lean more toward panel data. Matched panel data — more like your [COM] scores. Nielsen Net Ratings, we’ve used. And/or we’ve conducted primary research, where it made sense.

–

Now we know you, dog!

MM: I’d like you to speak to the old Internet joke, “On the Internet, nobody knows you’re a dog.” It gets to the difference between the known person and inferred unknown person. How can you use longitudinal data about the inferred unknown individual and produce useful insights beyond “some one did x?” Can you take us through how you know that an inferred-unknown individual on the 12th of December 2008 was in fact same inferred-unknown individual of October 23rd?

Well, certainly from a unique identification standpoint, there are the old standby approaches. For example, cookies — which aren’t perfect, for sure. You put a cookie on someone’s machine, and they can delete it.

But many times, that’s a very viable approach. An individual — at least the same cookie — will return to the site and engage with you over a certain period of time.

I think though, in terms of your question of the black art itself… Pardon the short sales pitch, but one of the unique things about Targetbase is — again — our breadth and depth of analytical expertise.

I think if we excel in anything, it’s kind of getting back to that point of data agnosticism. To us, it’s all just data. We’re looking to apply methodologies and approaches. Many times, brand new — that nobody else has ever thought of — but oftentimes that have existed for some time. But using them in a unique context.

This is a good example. It’s one of our many approaches. It’s a good example of the latter — of using an existing methodology.

For example, one of the things that we pioneered with Southwest Airlines was the use of Bayesian Statistics. Bayesian Inference with web data.

Just a quick short version of what I’m talking about… Using a Bayesian approach as opposed to a classical statistical approach allows us to do a few things.

We can begin with subjective assumptions. To say that, “You know what? When somebody comes to our homepage, we’re going to assume that they’re there to buy a ticket.” For example.

But the approach that we take then is, everything that an individual does or doesn’t do — or how long they view a particular piece of content or what have you — is used as evidence to update the probability that they’re going to do what we expect them to do.

Take, for example again, Southwest Airlines. If you come to the home page, we can know just by looking at historical data that — let’s say hypothetically — 10% of people that hit the home page end up booking a ticket.

So when somebody hits the home page and they’re there for the first time, we say, ” Our expectation of you booking a ticket is a probability of 10%.” But then that person clicks on a particular piece of content, or they go to a different section of the site, and that probability is dynamically — in realtime — updated. To say, “We just went up to 15. Uh — we just jumped to 50,” et cetera.

So the approach is an ever-learning, ever-improving approximation of who an individual is and what they’re there to do. Again, that’s very much a foundational principle at Targetbase. That ever-improving approach.

I often use the analogy that we may start with a really grainy, fuzzy picture. But as we move forward, that picture becomes clearer, clearer and clearer. In fact, a quick side note… One of the things that impressed me the most in terms of our like-winded philosophy with Alterian was one of the early discussions we had.

They were running through some slides, and they used the analogy of a “fingerprint,” that was partially concealed. Slowly over time, it was being revealed. Of course, that analogy fits perfectly with this idea of, “You may only know a little bit right now. So you’re making an assumption with maybe relatively low confidence that it’s true.

But over time, you’ve increased the confidence. You increase what you know about an individual. You increase your predictive ability in terms of what they’re going to do.”

MM: Let’s move beyond cookies and other kinds of unique identifiers that a user may provide. I mean, it’s pretty clear that if I’m a Southwest customer and go to www.southwest.com, it doesn’t take more than a click or two to re-establish the fact that I’m a particular Southwest customer and link my current activity to a customer record. You’ve got now good longitudinal data in terms of what I’m doing and how I’m doing it.

Right.

–

CPG Innovations from Targetbase

MM: That’s perfect. Could you give me an example of some of the innovations that Targetbase has brought to the CPG area? Specifically where you’ve worked around the lack of POS sales data?

Yes. In fact, this gets specifically into what we refer to internally as our “pattern of engagement analysis.” Thinking that we actually pioneered for a direct commerce site — Southwest Airlines’ site. We then ported it over into a CPG environment, without the transactional data.

Really, at its core…

MM: Before we let that gem get away, I what you to expand on the strategy of first establishing a baseline pattern of engagement. Shall I assume that you first developed the data structure and approach using a very robust, high volume e-commerce site from Southwest?

That’s where the idea started.

MM: So, the idea of patterns of engagement came out of dealing with real customers, real services. Real content. Real transactions.

Exactly.

Patterns of engagement

MM: Again, shall I assume that you developed an evolving set of analytic themes and data-analysis patterns that you then brought back into an area, where you lacked one critical part of the overall analytic equation — real POS data. As a function of the analytic themes and patterns you developed where you did have it, did you find that ou could then infer or make really good educated guesses, in terms of what the other non-POS activity data was actually telling you?

Well, I would say yes from a methodology standpoint. The actual patterns of engagement themselves are unique to the separate clients.

But yes in the sense that the methodology that we developed — which I’ll describe in a second… We kind of cut our teeth in a very much hands-on transactional data sort of way, in order to move that approach over to other clients. Yes.

We tend to view engagement differently, I think, than most. Particularly webshops. But most, in general.

When you refer to “engagement,” or when you see it out there in the marketplace, most are talking in a very aggregated sense, and at a very point-in-time sense.

For example, I put a video on YouTube and I got a 125,000 people to view it over the last week. That’s my engagement metric.

Well, that’s not the way we think about engagement. That’s one of the reasons we refer to it as “pattern” of engagement at Targetbase and with our clients.

The way we view this idea is that each individual — known or inferred — as they come to our site or engage with an application on their phone… Whatever the case may be… as they visit a store… They leave clues. They leave evidence of their presence.

For example, if somebody logs onto the General Mills site for Betty Crocker. They go view their recipe box, where they’ve stored recipes that they’re a fan of. They go view a video on how to make macaroni-and-cheese or whatever the case may be. They engage with the site.

Each of those activities is tracked. And most importantly, each of those activities is then stored on a database at the individual level. So from a processing standpoint and from a data-capture-and-storage standpoint, it is sizable. But thankfully we have the technology to achieve that, now.

As you mentioned earlier in the call, we have the analytical tools to utilize that now, as well.

But each of those activities is stored at the individual level on a database. Then our approach is to basically — in the beginning — subjectively weight each of those activities. To categorize them in terms of their relation to our desired outcome. In this case, purchase.

For example, the fact that I open an e-mail is a much lower form of engagement than printing off a coupon. If the ultimate objective is that they’re going to purchase my product. So, e-mail opens are a much lower form of engagement.

The reason I say, “in the beginning,” it is subjective, is because a lot of times we have to work with our marketing partners to just have a starting point for what we think the impact of things is going to be. Then as we capture more data, we can actually model or regress in order to determine the actual impact of interaction with a particular piece of content on the ultimate outcome of purchase.

For example, there might actually be a certain class of coupons — or a certain class of e-mails, videos or what have you — that are more predictive or have a bigger impact on the likelihood of purchase.

Once we establish that paradigm — that weighting paradigm — then everything that every consumer does with this, that we can measure and track, is captured and stored and weighted and rolled up into this metric that we refer to as, “pattern of engagement.” It’s captured over time.

For example, in the case of General Mills, we have 3-plus years for those that have been there that long, of everyone on their database. Their pattern of engagement. In this case, we store it in monthly buckets. We roll it up into months. Then we’re able to perform time-series analysis on it. That allows us to trend it and to model it.

We can then say that at the individual consumer level, “Here’s somebody who is high in terms of their level of engagement, and they’re decreasing in activity from a directional standpoint.” That vector or that combination of volume and direction tells us how we need to interact with them from a relationship standpoint. “Here’s somebody who’s mid-level engaged, but they’re increasing in their engagement.” That again dictates how we interact with that person.

I think the uniqueness in the approach is the sophistication of the analytics, certainly, and the strategy and actionability it provides. But there is also a big technology component. Ensuring the proper tagging and tracking is in place, not to mention the sheer size of the proposition to capture all the engagements each individual consumer has with us, and to store them on a database, over time.

–

Traditional segmentation

MM: If I understand that right, Trae, traditional consumer-segmentation efforts use census and other forms of compiled data such as credit histories and other forms of household data, creating a cluster or basically a neighborhood. Now that entails the presupposition that everyone in that neighborhood will buy a similar set of things; that consumers that share similar socioeconomic backgrounds and motivations will buy the same stuff. However, in reality, each household of a particular neighborhood often represents a huge difference of consumer appetites, criteria, and mind-styles. So the idea of Cohorting takes another approach: rather than working from the physical data (households of a neighborhood) to develop a data set, you now work backwards consumer appetites, criteria, and mind-styles, creating logical set that you call a cohort. This approach of cohorts or logical groupings of buyers makes neighborhoods like Swiss Cheese-where each hole represents a distinct set of buying criteria and, when group together, create a cohort.

Absolutely. In our approach to Cohorting, we incorporate attitudinal data (from surveys, interviews, focus groups) demographic data (from consumer or business databases) and behavioral data (from websites and syndicated sources). We certainly see big differences within neighborhoods along attitudinal and behavioral lines.

A lot of it has to do with the methodology we employ. For example, we may begin with broad and nationally-representative syndicated data like a Simmons or an MRI, enabling us to incorporate many of those things. Not just geographic, demographic data, but attitudes and behaviors across a broad swath of areas including brand usage.

Other times, where available, we’ll start with our client’s customer database-if they have a database-mapping those customers into syndicated or third-party data, or direct primary research.

By the way, since we’re talking about Targetbase and our approach…we define ourselves as data-agnostic and make it a point of differentiation. We’re not trying to sell data; it’s not part of what we provide. We don’t want bias our results and outcomes. It’s really more about the outcome, about the solution that we’re trying to deliver, and not about selling data.

Brand-portfolio mix optimization

MM: As you said, you can take one way of looking at the data in standard cluster analysis, using data from syndicated research or compiled data. Or you can segment along segmentation themes that reflect mindsets of various cohorts, supporting brand-portfolio mix optimization. So with a large portfolio of brands, segmentation themes enable you to say, “Okay, what is the optimum mix of brands that we should market, can market, have marketed, to this particular profile of behavioral and attitudinal data?”

Exactly. Like I said, there are certainly themes that crop up across clients, across verticals. But one example for a particular client is a segment that we have dubbed, “Maxxed Moms.” That particular group — some of the themes or points that really stick out are…They’re mothers and certainly time-crunched. Their time is important to them. However they also want to feel like they are pampering their family and taking care of their family in a very traditional sense.

Not only does that level of attitudinal behavior and understanding drive how we message to those members of the segment — the types of imagery, et cetera, that we use… but the cohorting analysis also tells us what products the client has a right to win with.

For example, if you’re going to that segment of Maxxed Mom, products that are geared toward time-saving and/or pampering your family. Home-cooked meals in this case might resonate very well with that segment.

So then we incorporate scientific test and control methodology to determine if — in fact — we can win with that particular segment, with that set of products.

–

Case studies on cohorting

MM:  Would you just walk us through one or two case studies?

TC: Yes. We have some well-known clients for which we’ve done some great work, such as P&G, GlaxoSmithKline, Pacific Gas & Electric and American Honda, and in the area of consumer segmentation or cohorting-an area for which Targetbase is most well-known.

MM: Would you give us a quick primer on cohorting?

Certainly. If you look at our packaged goods clients like P&G and General Mills, they have a tremendous portfolio of brands. P&G came to us many years ago with the question of how they could leverage consumer insights-a better understanding of their consumers and what mix of brands to promote to each person on their very large database.

This resulted in turning the old model of brand marketing on its head: it was not about the brand; it’s about the consumer. They asked us to develop a mix of brands that a consumer would be interested in-that P&G has a right to win.

We often find among large portfolios of brands natural groupings or segments of consumers-what we refer to often times as a “cohort”—a group of consumers that tend to utilize a group of brands or a mix of brands in a particular way.

Originally for P&G and later for General Mills and others, we developed a methodology for identifying those unique segments or cohorts within their consumer base.

We identified them in such a way that P&G and General Mills could then know not only who they are from a behavioral and attitudinal standpoint, but then what products they were likely to be interested in…thus how o drive relevant messaging, offers, etc. to optimize their sales at the individual consumer level.

Joe Six-Pack and Hockey Moms

MM: Could you give expand your definition of a cohort? We just went through a election, where candidates and pundits used cohort-like terms of Joe Six-Pack and Hockey Moms.

Yes. Absolutely. There are certainly broad conventions. But one of the services that we offer our clients is a truly customized approach to segmentation and cohorting, specific to their brands and customers-unique differences.

There are certainly themes that crop up on a regular basis, but it is actually a customized approach — as opposed to, say, the traditional prism cluster type of approach. Where every neighborhood in the United States is dubbed a particular name, to fit within that prism cluster.

It is certainly consumer-centric, but it’s in the context of a particular brand or group of brands.

–

Michael Moon (MM) conducts masterclass interview with Trae Clevenger (TC), VP, Analytics, a master practitioner of multichannel analytics at Targetbase, a customer engagement agency.

–

Professional background

MM: We’re here with Trae Clevenger of Targetbase. Trae, would describe your current position and a few career highlights?

TC: Sure. I’m currently VP of Analytics at Targetbase, traditionally a direct-marketing agency.

At Targetbase I focus on innovation, helping to drive innovation across clients, particularly in the areas of analytics and digital marketing. As an integrated marketing agency, we focus on outcomes; most of our approaches and innovation impact technology and creative, or content.

My career spans multiple industries, many in Fortune 500 companies, in the areas of strategy, analytics and solution delivery, emphasizing statistical analysis and modeling, predictive analytics, segmentation, optimization, web analytics, relationship marketing, engagement analytics and behavioral targeting.

About Targetbase

MM: Would you describe Targetbase, its typical clients, and solutions?

TC: Sure. Founded in 1979 as a spinoff of MARC Research, today Targetbase has a long-standing market research and consumer research heritage.

Our analytics foundation remains at the core of everything we do down through the years.

Our business model—if you think of a hub-and-spoke—put analytics and insight at our core as the hub with strategic outputs and outcomes as spokes, producing technology, creative, and strategy that we believe delivers superior consumer insight

Our clients span the global and cross many industries, including Travel, Finance, Insurance, Pharma, Healthcare, Automotive, Utilities, Packaged Goods, and Retail.

The length of our client relationships represents one of the unique more things about Targetbase. Most of our clients view as a trusted, strategic partner (as opposed to a vendor or agency). Our clients value their relationship and tend to maintain it over the long term.

Concerning solutions, we pride ourselves on integration. We provide strategy, analytics, technology, and creative. However, the real differentiator is how all of those pieces come together to provide superior solutions for our clients.

For example, technology and analytics combine to produce more actionable business intelligence. Creative and technology combine to produce dynamic, customized content delivery. Analytics and Creative combine to produce communication planning and more targeted, impactful content.

We often use the phrase “Database. Digital. Direct.” So much of what we’re doing these days is increasingly in the digital space and the online world, using that term broadly, to incorporate not just web but mobile and every other touch point really that’s available out there. Very much a 360-degree view of the consumer is our philosophy.

–