Site Search

How to Leverage Real Time Analytics to Improve Market Results


Paul McConville and Jason Beyer walk you through ways to use customer segmentation to improve your market results in this detail oriented webinar.



Analytic and segmentation to identify your best customers, and use that as a mechanism for identifying and prioritizing target-- then demonstrating how you can gain key insights from this analysis to execute a very strong marketing plan against your overall marketing strategy, opportunities that exist in the marketplace, your customer acquisition or retention initiatives, media planning and buying, message and offer creation. And throughout the webinar, I'll be giving real business use examples to illustrate how to do this within the ElementOne platform.

So I wanted to level set a little bit here with the group. I know we asked some questions of the participants prior to the webinar and the majority of the participants are using segmentation today or plan to use segmentation in the future, but just wanted to set some basics here in terms of what segmentation does and doesn't do. At its lowest level, segmentation models assign individual people, households, or geographies into clusters or segments based on similar attributes or dimensions that can be used to describe them. Commonly, there are two different approaches to segmentation. There is custom or syndicated models.

What I mean by a custom model is you're doing primary research to better understand your target consumer as it relates to your business. And traditionally this is done through primary market research, fielding a survey of some kind and going to a sizable portion of your prospect universe, and better understanding their needs and attitudes and behaviors as relates to your business. And then doing a cluster analysis against those responses to group those participants into like segments.

Typically, these are done by large research companies and they're fairly expensive custom projects that are done. The contrast to this approach is syndicated models, which takes large industries and develops a segmentation scheme, using similar techniques around understanding the consumers and grouping them together based on like behaviors as they relate to a particular industry. The benefit to a syndicated model is that you gain a lot of speed to market advantage, in that it's pre-done, you engage with a company that has a syndicated model and they can score your customer database against those syndicated segments. And then you have a good understanding of where they all. The biggest con to syndicated models is that you're using the same segmentation scheme and intelligence as the rest of your industry, and in many cases, your business is different.

There's differences in your brand, what you bring to market, your value proposition, that isn't the same as your next competitor. And so those nuances are often lost when you're going with the syndicated model. And we're going to talk about the TARGUSinfo* approach to this in a minute, but wanted to just lay some foundation here. And so common assumptions we have in doing any segmentation is that not all customers are the same and that consumer behavior isn't random and actually can be modeled against. And then any business or company that's interested in using segmentation also has to have this understanding that a different marketing strategy will be needed to go after different segments that are behaving very differently.

And so with those foundational components, I'd like to introduce TARGUSinfo* segmentation marketing analytics platform, which is called E1 or ElementOne. I know we have a few existing clients on the line here, so that's not new information. But for the remainder of the participants, ElementOne is our marketing and segmentation analytics platform. What makes ElementOne unique? And really it goes to what I was speaking to around custom and syndicated approaches to segmentation as really the blend and best of breed of those two solutions in one.

We make the segmentation actionable and relevant. And what I mean by that is, a custom model is going to be the most relevant to your business. And a syndicated model is likely going to be most actionable because in most cases, all US households are associated with those indicated segments. And so you can go out and buy direct mail lists against those segments and find those households out in the market, where a custom model it's very difficult to do that.

What ElementOne does is it brings those two together. We allow for household level targeting based on specific client's customers behaviors and buying propensities. We have a mechanism that allows this segmentation scheme to be easily applied to existing customers or prospects that are out there. It's a very cost-effective way to do segmentation because you're not having to go through a large primary research engagement to collect data, that data already exists.

And you're also getting the benefit of the tremendous data resources that TARGUSinfo* has that many of you that are existing TARGUSinfo* clients are aware of. What our segmentation schemes also allow for is knowing lifestyles, media preferences, online behavior, and product propensities for categories that extend beyond your primary business-- which is, I think, very unique to the platform that we have. And then what it allows for, finally, is finding those segments or targets that you've deemed important, those that are more likely to buy your product or services or have high value to you can be easily located within any geography in the United States. So how do we get to this?

And why is it called ElementOne, and what are elements? So elements are essentially the basic building block that we use to associate all US households. We have 232 of these base elements and they were created based on three primary dimensions. Starting with household level demographics, which many segmentation schemes are built on, we use age, income, presence of children, home ownership, as primary input. We also layer on neighborhood factors like the urbanicity or the population density in which a household lives, along with the cost of living of that area of the country. A segment living in New York City is very different from a segment living in a rural part of the country. And so that's taking into account in our base elements in our platform.

And then finally, these elements were optimized by lifestyles, purchase propensity, online behavior, and shopping behaviors that we also have at the household level. These factors all were used to define these 232 base elements. And so how do we use this to build a segmentation scheme, particularly a segmentation scheme that's custom for our clients? Well, we start with that bottom layer, which is all US households that I mentioned that we have data on. And then we roll those US households into the 232 distinct elements which I just talked through, the building blocks of those. That's our base layer. And that's where we kind of separate from most syndicated segmentation models. That's a very granular view and micro-segmentation that can be applied to any business, any industry, or any service. From that level, that 232 element level, we then engage with clients and get their and your specific customer data. An understanding of what's important to you and your business and that's-- Who's buying, and what do they look like, and how are they engaging with your company. And so we'll take as many attributes as are important to your business and the primary drivers for your business.

And so that will likely include whether or not someone's a customer, as a primary one, but what products and services are they buying from you, how long have they been a customer, how likely are they to stay with you. All these things, all this great information that most companies already have on their existing customer bases in their data warehouses and their marketing analytics programs, we'll pull into the modeling process and group those 232 distinct elements into approximately 50 tactical segments that are customized for your respective business and industry. And now I've worked in marketing for a long time-- 50 tactical segments is a lot for any company to manage or deal with. And so what we'll also do is roll those 50 tactical segments, based on input from our clients as to what is the primary thing they would like to segment on, into approximately 10 strategic groups.

And this is what you think of when you think of the high level segmentation schemes and segments that I can actually action against and develop marketing programs against. Typically, it's at the strategic view. Any above the line marketing that you're doing, advertising, you're looking at the strategic group level. And at that strategic level, that's where you start to prioritize those segments that are important for your business, that are more likely to be customers, that are exhibiting certain value metrics to your business that you want to go out and spend money on marketing against. So that is the high-level overview of the methodology in which we use to build custom segments for our clients.

And before I get too much further and we start diving into the ElementOne platform a little bit, I did want to just kind of set the stage and lay out some common terms that we use at TARGUSinfo*. We already talked about elements. Elements are how all those US households are grouped together based on the demographics and geographical and behavioral information we talked about. Those are rolled up into segments. The segments are really just those elements grouped together on client-specific behaviors and attitudes, as well as underlying demographics that go into that. We certainly don't want to group elements that are very far apart demographically into segments because those become very difficult to both find and action against.

And then finally, we roll those up into strategic groups that are your segments rolled up into higher level groups, based on those common or very critical business drivers for you. And then once we have that segmentation scheme developed, we create profiles. We have 14,000 profiles pre-built into our segmentation platform. But for clients that has this particular customer data set, you're going to want-- and a customer behavior, maybe a customer's buying a particular product or service-- building custom profiles to look at your customer base again. And then finally, just refresher for those on an index-- I know we have some statisticians that are in the audience here-- but it's a value that represents the propensity of those segments to exhibit an attribute or behavior that's captured in that profile.

So we use the index as a relative measure to show segments that are over-indexing or exhibiting a higher likelihood towards a particular behavior, or those that are under indexing and are less likely to exhibit that behavior. So kind of enough with the background. I'm going to switch over to the platform and just give you a quick overview. ElementOne is, as I mentioned our analytics platform, it's a web-based platform. Which is great because for those of you that have worked with IT, those can be very challenging implementations for server-based applications that needed to be installed on the customer premise. So ElementOne is an online platform that allows you to upload and manage your data and analyze it all within the platform. And so you can analyze your customers and prospects in the platform and understanding kind of what they're doing. This is built against the segmentation scheme that we just walked through. So for our clients to better understand their customers and prospects, we've built a segmentation scheme that's loaded into this platform. And that's what you're analyzing your customers against.

You can look at where they are in the market through some of our mapping capabilities. And then you can also build custom groups based on your tactical segments. And so maybe there is a particular behavior that you would like to analyze further or group your customers on. You're able to do that within the platform as well. So I think more importantly than this background is getting to why we're here today, which is talking about making real time analytics valuable to your business and taking those insights and applying them in real time to your marketing activities. And so that starts with analyzing and understanding your customer base. And so for the purpose of this demonstration, let's say we've gone through the process of building a segmentation scheme as a base foundation.

And for this, I went to the most granular level possible, which is the 172 segment level. It's far more than the 50 we typically see in the client engagement, but for walking through and showing the power of this platform, I felt it was important to go down to a fairly granular level. And so what we're looking at here is the 172 segment level. This is a dummy data set that I put together of about 50,000 customers that have bought a product or service. And what I've done is created a profile around that group of customers and loaded it into the platform. And what we're looking at is the distribution or the index distribution of those segments relative to customers. And so what we're looking at on this index here is there's segments exhibiting a very high propensity to buy this product or service relative to their size in the market. So we take all US households, which I mentioned we have data on all US households, and their size in the market or their composition, and compare that to the distribution and composition of the customers in that data set.

So in this example, we have segment 166 here that has about 100,000 households in the US, which is 0.11%. So very small, when we talk about these elements and these very tactical segments, we're talking about very micro segments. This particular data set has 136 households that fall into that segment, which is a 0.35% composition. Which is three times bigger than their composition in the market, or a 315 index. And so what that's saying is, all things being equal, we would see the same household distribution that's reflective in the US, the same in this customer base. But that's not what's happening. This particular segment, 166, is buying this product and service three times more often than their size would suggest. And so they're a very over-indexing or a high-propensity segment. I can sort all of these segments by that index, which I'll do right here. And you can see in the index bar chart below, that segment 166 is the second highest propense segment. Right next to them is segment 78. So these two segments are exhibiting the highest likelihood of buying this particular product or service. What's interesting is when we look at the demographics of these particular segments. And the segment 78 is approximately $60,000 average income, where segment 166 is $15,000 average income. Average age of segment 78 is 28, while segment 166 is close to 40. Segment 78 has children, segment 166 does not have children. Segment 78 never owned their own home, while segment 166 tends to or half of them own their own home. Segment 78 lives in very affordable cost of living areas, while segment 166 lives in the most expensive parts of the country. And both of them tend to live in the suburbs.

So I think the takeaway here, is what I wanted to show is for this particular client, these two segments are exhibiting the exact same propensity to be a customer for this product, but they exhibit very different demographic characteristics. And when you start to look beyond just the demographics and the behaviors of these segments, and understanding of what they do, the products and services they interact with, you'll find that they're also different, very different from a behavioral perspective as well. And so going back to the process, this is identifying who's buying this product or service.

And to take it a step further, how we consult with our clients and help them or prioritize, is we'll use things like their buying propensity for a product and service, and use that as a function for grouping segments together. So I just want to flip over here. I've taken that same data set and grouped high-indexing segments together into these higher level groups. And so I've broken those out into primary targets, secondary, above-average, average, below-average, and then non-targeted at all. And so what this is, is essentially a roll-up of all of those underlying segments that were exhibiting that likelihood to be a customer, into a larger kind of macro segment, who this company would use as their primary target group.

And so if you look across the distribution here, you can see for this primary segments or target group, they're showing a very high index, almost 2 and 1/2 times likelihood of buying the product or service for this particular company, and clearly a strategic target for them. This second group is really a secondary target. They're over-indexing for buying this product or service, but not nearly as much as the primary target. And then you kind of move on down the line. It's important also to know who to not go after with your targeting, and to not to prioritize when they come in the call center as a less-qualified need because they fall in this segment that's least likely to buy a product or service.

And so this non-target group is indexing at 37, which is 60% less likely to buy the product or service for this company, and clearly a segment not worth investing marketing dollars against. And so once you're starting to get a picture of who's buying and prioritizing who to target, you want to start to go out and look and understand where the opportunities exist. And so for that, you can go in and analyze markets against these segments or groups that have been built. So for this example, I want to just go down and grab segments 78, who we saw as the highest propense segment for this particular product. And I want to know where segments 78 are-- what parts of the country, and specifically, I operate, for example, in the Texas market. And so I want to look at where segments 78 are in Austin, Texas. And so I want to do that all the way down to the zip code level. And so the market analyzer within the ElementOne platform allows for you to take any segment or group and apply that to any market in the US, all the way down to the block group level.

And what we'll see is the number of households that are in that media market relative to this segment-- in this case, I'm looking at zip codes-- and then the number of households that fall within that zip code and that segment, the composition of that and the penetration. So it's preset to look at the deciles. So you wanted to look at a top decile of that segment, in terms of their index. For Austin, the Austin market, right here, geocode 6482, has 4,148 households in it, 53 of which fall into segment 78. Which is a 1.28% penetration rate, which is over the average by two times. And so it shows you where-- and this is, since the segment I used was so small, you were looking at very small household counts-- but I think it shows the granularity and it shows the pinpoint accuracy that you can get to at these segment levels within a particular geography. And for a lot of people, looking at this view on a map is very valuable. You're wanting to look at where are those hot spots within your market for a particular segment. In this case, we're looking at Austin and surrounding areas.

And we can dive a little bit deeper here for this geography and see that our segment is living kind of outside the city. The dark orange here represents the areas or parts of the city that are showing higher concentrations of the segment in which we're targeting. So if I zoom in a little bit more here-- and it looks like I'm getting in on San Marcus, let me go up a little bit closer to Austin-- we can see exactly where segment 78, which is our highest propense segment, tends to live within the Austin market. The darker the orange here on this slide shows the darker the concentration of that segment within those zip codes for the Austin market. So the application of this are really to help for market planning and identifying what markets are going to be most attractive to go after, segments that tend to buy from companies. In addition to that, you can look at forecasting and projecting a new product onto a given geography or footprint.

Many of the clients that we work with have very tight areas that they serve, and so you can customize these views based on geographies that are important to your business, and look at the segment distributions within those specifically. This can really help inform out-of-home communications, where are you going to put billboards or signs, drawing traffic or attention to your service. Or even event planning and understanding the pockets of a market that are going to be most attractive, or the segments that your company is attracting. And so this is just scratching the surface. And Paul mentioned at the forefront, we're just going to really just give you an overview of some of the capabilities here and how you can make your analytics more impactful.

And so I'm not going to go into too many more details on the market analyzer. We could probably spend an entire webinar around that. But did want to just highlight that section. And so that's really speaking to the strategic analysis. And as you can imagine, really understanding who's buying and why they're buying has a lot of applications. What's, I think, critically important is knowing that and being able to find those households. Whether it's through their physical address, their email address, or even their presence online, there's ways to reach the segments that we've identified for our clients. And so you can already start to envision in how that makes your marketing so much more impactful, when going in you're knowing, these are the segments that buy my product and service at three times the rates of average.

And I know what websites they're on. I can reach them through online marketing in real time. I know where they live, so I can send them direct mail communications. I know their phone number, and so when they come into my call center, I'm doing outbound telesales. I can interact with them on a whole different level because I know who they are and that they fall into my primary segment. And so from that, I think I wanted to get into a little bit of-- that's really the strategic side and the opportunity analysis-- I think we wanted to dive into a little bit on the customer acquisition side.

And so looking at the channel in which you acquire customers. And that through certain channels, you may have a very different segment that's coming in the door than others, and optimizing your marketing mix around that channel behavior from your customers. And so for that, I wanted to look back, once again, back at this 172 segment level. This view is looking at all customers, but I want to look at those customers that are coming in through the online channel specifically, and then compare those with those that are coming through my total channel. And what you're looking for here is differences and what's that delta.

Segments that are coming in through online that are not coming in through my traditional channels or might represent my traditional customer, are an opportunity on the online side. So I'm going to sort this same data set. Just so you know what I've done, I've projected the total customer profile, that's represented with the blue bars on the index chart, against the green, which is just my online customers in this example. And if I sort those by the online index, and so those more likely to come in through the online channel, I see very strong correlation here, as you expect for most businesses between those customers that are coming in through online and those that are coming in through traditional channels. But what I wanted to highlight, is you have different degrees of propensity and you can see these green bars over-extending beyond some of the blue.

And what that's showing you is you have segments that yes, they're more likely to be customers. They're also much more likely to come in through your online channel. And so that's the media you want to go after. For those segments that are their buying online, you want to reach them where they are online. What you see here is a few segments that are over-indexing for being customers, for in this example, but under-indexing on being customers and coming in through the online channel. And so what that tells you is these segments, you may want to group them differently, and reach them and find how are they coming on. Are they calling our call centers? Are they coming into our retail stores? How are we acquiring those customers? And adhere your communications towards those segments differently, than you would your online segments.

And so we've talked about optimizing based on channel, now I want to get into the media planning here quickly. Yeah, let me just double check if we have any questions here that I'm missing. And looks like we're all set so far. But please do, don't hesitate to type in a question. If I'm going too fast or something I say raises a question in your mind about the platform or the process here or what you can do with this data. The next thing I like to look at is how could you impact your media planning.

And we work with a lot of different media agencies directly. Many of our clients will give access to this platform to their media agency and say, OK, we've done the homework, we've identified these are the target segments we want to go after, how do we develop a media plan that's tailored or customized toward them and then execute those media buys. And so I had mentioned previously that we have about 14,000 different profiles in this platform. We have a tremendous number of media profiles as well. Just to give you a sense of some of the categories that we have, very deep on the demographics and lifestyles of the segments, product potential across a variety of categories, deep in the media planning area, channel planning, as well.

And then we get into some of attitudinal statements around messaging and creative, and what marketing really resonates with individual segments. And promotion pricing as well. But for media, I just wanted to take a print media example here. And you can see we have, for print media, about 400 specific profiles just geared up around print. And this is going to take a view of all segments here, of the 172 view that we're looking at, relative to those print media behaviors. What I want to highlight is that this is tailored to the segments that are customized for our particular clients and the group that those clients may put together.

And so for this example, we already looked at where segment 78 lived in Austin, Texas, but we want to understand what's the print media for 78. And as a contrast, we also want to look at segment 166, which was, if you remember, the two segments most likely to buy this product or service showing almost the exact same propensity to buy. But what I want to highlight here is within their media, within media planning and what they read, because we're looking at print here, they exhibit very different behaviors. And so I'm going to sort these based on the highest propensity. By the segment 78, which is these guys are reading Muscle and Fitness at a high index, which is also shared by segment 166. And so you start to think about, if we group these segments, which we did do, into that primary target group, this is a publication that's shared across these two segments.

But I did notice here that segment 78 reads Playboy, while segment 166 does not. So that may not be a magazine that our advertisers, that you as an advertiser are interested in buying media against. But the parenting group has a magazine here that both segments are over-indexing on, and actually one where there's fairly high penetration. And so that would be an example of where there is alignment between these two segments and one that would warrant consideration if they were grouped together. So this is a quick example and I'm already going along, but I just wanted to highlight the depth that you can go into in your media planning-- and the impact you can make when you're associating your specific target segments that you've done the analysis and shown are showing a greater propensity towards buying from your business, and what they're reading, what they're watching. And so a good example there of that. And now I want to go back to the presentation and talk a little bit about what does this mean.

Those on the analytics side are probably loving this. Those on the advertising side of marketing or communications side are like, I don't know what to do with 50 tactical segments, and what do I give my agencies. How do I brief my agencies on these segments in a meaningful way? And so like the example I showed you where I rolled up those micro-segments into more actionable groups of hyperpropensed segments and identification of a primary group, once you've done that, you can develop detailed profile against those groups and really start to understand the who of the persona that's buying your product or service.

And visuals really help with this. And so we helped put together kind of profiles of those target groups. And I've just put together an example of one here, which is kind of a generic one. It's this power players group which represents 25% of all households. And these are young professionals, out on their own, getting going with their career. They have higher incomes as a result of their profession. And they're spending money on whatever they really want. They are very social and enjoy hanging out and going out with their friends. You look at their demographic profile, about 27 years old, almost $100,000 in income. Very few of them have children. A lot of them are buying their own homes or condos. They're found in urban or suburban areas.

And as I mentioned, a lot of these guys are professionals, and so which explains their higher incomes. And if we think about their lifestyles and hobbies, all this was pulled directly out of the platform once we put together those groups. So we're not going out and doing primary research to better understand what these groups look like. All this is data that exists within the platform. And so understanding what their major life events are, graduated from college, they're getting married, even changing jobs. We know attitudinally they're very driven and want to get to the top of that career ladder. They're going out to bars and nightclubs, hanging out with friends, playing pool. A very physically fit segment, playing tennis, jogging, going to the gym. And then, OK, we looked at the media profile in the platform. We're pulling this out, what does it look like? OK, this is a segment that's most likely to be listening to the radio and also listening to music. Music is very important to them.

You can imagine, as you dig into their technology profile, all of them have iPods and iPhones. Their mp3s are important to them. This is a segment that's watching sports and very big into football and basketball. And heavy online users, as you'd imagine just given the demographic profile. They're using mobile broadband home and on-the-go, and using a lot of bandwidth when they're doing that. So this is just a representative example of something that we might put together for one of our clients.

But to bring it home and show when you're implementing segmentation, this way of conveying a profile and understanding at that strategic group level is critically important, particularly when you're briefing an agency on who you're targeting and developing creative. And along those lines, there is a role for primary research, even with our segmentation. And that's to address questions that can't be found in all of the profiles that exists within E1 platform. If you really want to know why someone is doing something, the motivations for it, the equity that your brand may have, it's important to go out and survey and talk to, through primary research, those customers.

The beauty of the ElementOne platform and what we can do is once you understand that segment, it can be applied in the platform to the total US population. And so you have a real good sense of the attitudinal and need to mention of a segment, but you know how to find them, where they live, and what they're doing. And so I think the combination of those two things makes for very powerful marketing communications and enhancing your overall marketing plan. And I think on topic, we're talking about real time.

You take that primary research and all that you know about a particular segment, and then through the TARGUSinfo* platform, the ability to score a lead, a prospect coming into the call center, someone that's online, through a display ad or engagement through some content online, or mailing someone who's just called the call center communication. Based on their segment and what you know about that segment, is a tremendous way to increase the efficiency of your marketing and the relevancy of your communications. And so with that, I think I'm towards the end of my time here. And covered the majority of the overview and hopefully you found this valuable. It's really just kind of a taste of what we're going to start to deliver in these more targeted webinar series and just a good starting point to get going here with this. Is there any questions that you guys have from what we've been going through today? Feel free to type those in and I'll try to address those for you. OK. Well, thank you so much for all your time. I did want to hand it back over to Paul to wrap up here.

Thanks, Jason. And it's the topics that we talk a lot about with our clients asking, how dot they go beyond the score. So this is one way we've started looking at helping our clients get more out of their data to make better decisions in their marketing. And more strategic decisions help to drive all that they do to improve customer acquisition. This is the first of our webinar series. We would love your questions, comments, feedback to help us to continually improve to do these and to focus on our next series. We'll do these every six to eight weeks. The future one, today was really a baseline, will be more around industry and issue focused webinars.

Our next one will be Wednesday, February 1 at 2 o'clock-- "Putting Customer Intelligence into Action." So examples around how our clients have taken their customer data to improve both their lead purchase process, their lead process, and the decisions they're making in marketing. And with that, thank you. And our contact information here, again Jason and I would love to get your feedback. Let us know your thoughts on other webinars we could be doing, other topics you're interested in, and we will absolutely respond to you. Thanks again and have a great day.

*TARGUSinfo was acquired by Neustar in Nov. 2011.

Related Resources