By Stephen Shaw
As a nation, we often describe ourselves in terms of what makes us different. English versus French. East versus West. Affluent versus poor. Urban versus rural. Elites versus working class. Progressives versus conservatives. But those differences are never quite as stark or binary as they appear. We are more of a cultural mosaic — a community of communities — defined as much by our lifestyles and values as our demographics.
A lifestyle is how we choose to live. It is the totality of our habits, interests, attitudes. It is how we think — what we believe — how we spend our time — what we cherish most. And those lifestyle characteristics are remarkably correlated with where we live, for the simple reason that we like to live where we feel most at home. Where we live, in short, says something about who we are. Our next door neighbours may look different, may be slightly older or younger, but chances are they probably share many of the same values and beliefs. They probably watch the same shows — buy the same products — vote for the same political party.
That simple calculus is known as geodemographic segmentation. It is based on the premise that “birds of a feather flock together”. Marketers have been using lifestyle systems for more than half a century to benchmark their customers against the population, find look-a-like prospects in the market, select media channels, craft tailored brand messages, pick the best retail site location and much more.
The concept of geodemographic segmentation was initially developed by a social scientist named Jerome Robbins in the early 1970s. He took the first computer tapes of the U.S. Census in 1970, classified the demographic variables into five domains (social class and affluence, family life cycle, mobility, ethnicity, housing style and degree of urbanization), and found the key factors that accounted for most of the variance between neighbourhoods at the ZIP code level. He then grouped the zip codes into distinct, homogeneous clusters. It was a eureka moment. By knowing the ZIP code someone lived in, you could reliably predict their lifestyle (and by extension, their media and product preferences).
That revelation led Robbins to start up the company Claritas in 1974 for purposes of commercializing his cluster segmentation model which identified 40 distinct lifestyle segments. Later the company launched its PRIZM system (Potential Rating by Zip Markets) which became an instant hit with marketers who had begun to recognize that America had become a highly fragmented society. As niche marketing grew in importance, the demand for geodemographic tools soared.
In Canada the counterpart to Claritas was a company called Compusearch, also founded in the 1970s, whose geosegmentation model became very popular with marketers. In 1993 the company lured Jan Kestle away from her senior role as head of the Ontario government’s Statistical Centre and soon after appointed her President. After Compusearch was sold, Jan left the industry, only to return in 2003 to form her own geodemographic company, Environics Analytics, in partnership with the Environics group of companies. Today the company is owned by Bell Canada and is the leading supplier of geodemographic products and tools in Canada. Jan has become the doyen of marketing analytics in Canada, presiding over a dynamic team of 200 data scientists, software developers and marketing specialists.
Shaw: What initially drew you to the field of geodemographic segmentation?
Kestle: I was a mathematician by training. When I graduated, the first job I got was with the Ontario Statistical Center where I had to spend a lot of time understanding the concepts of the census, which was a great foundation for what came later. My last job there was leading a team called The Focal Point. I had to understand what kind of data 27 Ontario government ministries needed for a wide variety of programs. And then I had to go to Ottawa and sit at the table in the federal-provincial negotiation for the National Statistical System that was being developed at the time.
When Compusearch asked me to join, their business up until then had been primarily focused on helping retailers understand who lived in their trade areas. But they had expanded into automotive, and quite smartly they wanted to go after consumer package goods, financial services, and government. So, they asked me to help them sell their data back to the government sector.
Shaw: How big was Compusearch at the time?
Kestle: About 50 people when I joined. It had been founded by Bill Goldstein, and Mike Williams had just taken over as VP and was going to become President. Mike had a similar background to me: he’d worked in government statistics. He was responsible for taking what Compusearch had built and promoting it to other industries. So, he recruited me, and I joined in sales, even though I’d never really been involved in sales. But I never thought of selling data as really sales — I always loved the fact that you can solve business problems and society’s problems with statistics. But that’s when I first got exposed to geodemography.
Shaw: Did Bill Goldstein get his cue from Jonathan Robbins at Claritas? Or did he independently come up with the idea of leveraging census data for marketing purposes?
Kestle: There were a number of companies that pioneered geodemography at the time. In the UK, there was CACI and CCN Marketing, which eventually became Experian. The person behind that was Richard Webber1. Around the same time Jonathan Robbins started Claritas in the U.S. and Bill started Compusearch in Canada. But the initial concept of geodemography came out of the University of Chicago2, suggesting that if we know where you live, we can make a reasonable assumption about who you’re likely to be.
Bill initially went to Statistics Canada and pioneered an arrangement whereby they agreed to make their data commercially available to the business community. He started off by writing code — he was a programmer as well — where he could put a dot on a map and draw a circle around it, adding up the enumeration areas3. So, looking at the demographics, you could see who lived in a trade area. And then he brought in Tony Lea4 and others to develop the cluster algorithms that led to the Lifestyles segmentation.
Shaw: Did Tony develop on his own unique statistical approach to creating the lifestyle segments?
Kestle: When you build a segmentation system based on cluster analysis5, a lot of art goes into it — knowing which variables to select, how to ensure you’re not using co-directional variables like income and age. Tony pioneered the idea of using a relatively small number of weighted variables. But his work was a collaborative exchange with the small handful of geodemographic companies that existed at the time in Canada, the U.S., and the UK.
Shaw: What made Compusearch so successful in the 1990s?
Kestle: We were able to take survey data and project it to the ground, creating a whole new perspective. For example, if you have the survey results from 10,000 people and you’re making an estimate of what that’s likely to be for 750,000 postal codes, you need some pretty cool statistical methodologies. And so I’m very proud of the fact that through the ‘90s, Compusearch extended way beyond just the segmentation system.
We’d also expanded into the U.S. but by that time Compusearch was owned by Polk. And they were a great partner because the relationship between demographics and what kind of cars people buy is really exciting if you’re a statistician and a modeler. But Polk decided to divest its consumer marketing assets to focus on its automotive work. And that left Compusearch a little bit of an orphan. So one of the last things that happened before I left is I arranged the sale of Compusearch to MapInfo. And the reason why that was an appropriate home for Compusearch at the time was their kind of desktop mapping really needed data.
Shaw: In 2003, you decided to start Environics Analytics. What led you to that decision?
Kestle: I really felt like there was an opportunity and there was a core group of former clients and employees who agreed with me. We decided to put together a business case. I was looking for $1 million. That doesn’t seem like a lot of money, but I have tell you, trying to raise $1 million in 2002 to restart a business like that, well, it wasn’t easy. So, it was really great when Michael Adams, who owned Environics Research, stepped up. That was after more than a year of having banks and information businesses take us all the way to the altar, but never quite get there.
We were going to base the business around a segmentation system, but what was really exciting about doing it with Environics Research was that they had the psychographics — the social values data. It was an opportunity to combine a segmentation system, which is primarily demographic, with attitudinal data. So Michael invested in us and I recruited a founding team. Then we just rolled up our sleeves to build a segmentation system that was going to take us to the next level.
Shaw: How did you go about it?
Kestle: When you create a segmentation system, a cluster system, there’s hundreds of potentially good solutions to consider. But we had the technology to run a solution overnight against a thousand different variables and see empirically whether the system was giving you greater lift. Because when you build a system like that, you’re looking for the magnitude of small differences. You’re looking for high highs and low lows. You don’t say in advance you want this number of segments and they need to be this big — you actually allow the data to speak.
We decided to license the brand name PRIZM from Claritas. It was the first system used in North America and it was well recognized. We would work on which segments are similar in Canada and the United States because people always want a North American or a global system. At Compusearch we always used to say, “Let’s build a North American system!”. But you’re getting rid of your differentiation by doing that. So we built our Canadian system independently, and then we actually looked at the behaviors of all the segments in Canada and the U.S. and we twinned them. And so, we were able in the launch of the first product to say this segment exists in both countries. It took us a year to build the first PRIZM system. You not only have to do all that math, but you have to create the cluster profiles and names.
Shaw: Critics argue that geodemographic segmentation is old school. How do you respond to them?
Kestle: It’s the age-old question. Most marketers like to use a tool like PRIZM for acquisition. But it only starts there. You can also use it to understand your share of market, to understand sponsorship opportunities. It connects to location, so you can think about it for merchandising, and because it connects to values, you can use it for creative and message targeting. Even if a brand builds custom segments around lifetime value, you can still link them to the outside world through PRIZM.
Shaw: Are values becoming a key driver today of people’s decisions — that is, another way to segment them?
Kestle: Well, I think that values have always been important. To what extent are values changing? The latest work shows that yes, there is more polarization, but there’s also a continuation of more traditional values in Canada. So, the jury’s out a little bit. I think the next year or so will be interesting.
Shaw: How can marketers use values to drive more effective personalization?
Kestle: The opportunity to send different messages to different people is not being leveraged nearly to the extent that it could be. Of late we’ve done a lot more work with agencies than we used to do in the past, but mainly it’s been on the activation side where we’ve allowed our data to sit in DSPs and in DMPs6 so that our customers can actually activate against the insights that we give them. But the next frontier for me is to go back to our friends on the creative side of the agencies and say, “Let’s really leverage the values.” If you’ve got three segments, surely you can write three different messages and three different creatives, and you can have that consistent across the customer journey. So I think the time has really come. By bringing the values into the segmentation we can help them understand what resonates more with their customers.
Shaw: Geodemography came of age in a terrestrial world. But today almost all marketing interactions are digital. How do you stay relevant?
Kestle: We’ve adapted to the modern world. Physical address may not be the way by which you target — it may be programmatic, it can be your IP address in your set-top box, it can be a whole lot of other things in a digital world, but there are ways with a high degree of accuracy and reliability to link our data into those digital ecosystems. So if you have an omnichannel campaign with different types of people, and you can bring the values in, you’re going to get a better result. So, that’s why I’m excited about the future. We’ve taken this old methodology and adapted it. I think it’s still tough because the software and the platforms are still a little bit fragmented, but the more we can bring that together we can go from insights to activation with really targeted messaging.
Shaw: For most of your clients, are you leading as opposed to following? I mean, surely it’s got be a very, very tiny group of companies who see the world the way you’re describing it.
Kestle: There’s an understanding that data analytics can make a difference. But it usually isn’t put in the strategic context of what business problems they’re trying to solve. So the CEO and the CFO have to insist that organizations use data more effectively. That means breaking down silos; it means a common understanding of how they’re going to act on the knowledge. So, yes, in answer to your question, the numbers of advanced users is still small.
We’re going to be bringing in a strategic data and analytics consulting team to help our clients: three people — someone, very experienced, from our team; a very experienced person from data and analytics at a big agency; and a very experienced person from consumer-packaged goods. Their job will be to understand the main business challenges and objectives; bring the stakeholders together; help identify the roadmaps and the blockages. And then we will do the things that can help accelerate that agenda. Because in many companies there’s just aren’t enough resources to do it all.
Shaw: Is the public concern about data privacy an obstacle to growth?
Kestle: The climate around privacy is causing some headwinds. I don’t think there’s any reason why we can’t be data-driven in Canada and honor the trust that consumers have when they give their data to government, to information companies, to brands. But the data and analytics community has to recognize that people are demanding, “Tell me what you’re going to do to keep my data safe, to make sure I can’t be identified.” We need to follow the principles of consent and transparency. There’s way too much of a false narrative out there that says businesses using data are all bad actors. The banks and the insurance companies, the telecoms and the retailers, and most large businesses in Canada have spent millions and millions of dollars in the past 20 years complying with PIPEDA, complying with CASL, and making sure that the work they do is consistent with the laws of the land. Companies like us, we only use data that’s consented and permissioned, and we are already retooling our systems so that we can be sure that we don’t use any PII7 and that the data that we get are anonymized if required.
Shaw: What new moves are the government considering to strengthen privacy?
Kestle: I’m excited that we’re going to finally have legislation8 that will bring us GDPR equivalency. But we need the government to make all the regulations and issues around de-identification and anonymization really clear. We have to make sure that statisticians and methodologists are in charge of how we manipulate data. For example, we’re working on increasing our capacity to use mobile movement data — everybody’s antsy about that — taking the signals that are passively collected from phones and using those data in 100% privacy-compliant ways to help with things like transportation planning, healthcare delivery, social service delivery — but also help with getting the right products on the shelves, with getting stores in the right location, with understanding the mix between online and offline.
We have to help Canadians understand why data makes their lives better. And we haven’t done a good job of that, you know. We have to really explain that most people in our business are using data responsibly. And why data are so important to our future. When the census was canceled9 and then reinstated, it was amazing, I had people talking to me about it. And it made me realize that if we explained the issues properly and really helped people understand, we can counter the negative narrative which is out there. We have to be the advocates for data making life better and being used responsibly.
Shaw: How are you planning to help digital advertisers adapt to a “cookieless” future?
Kestle: We need a Canadian ecosystem where brands can share data — where the data from an advertiser and a publisher can be shared in a privacy-compliant way to understand what the best activation paths are. So we’re working on that. We’re building a clean room10 where first-party data can be blended in a secure environment for a limited period of time, consistent with the consents that exist, so that organizations can understand their co-marketing opportunities, their sponsorship opportunities. It allows us to develop audiences with partners and then measure the result. So we’re investing with a ton of partners in an ecosystem that will allow data to be blended so that we can assure our advertiser customers that there’s a Canadian solution that works for them in ad tech. We work closely with Bell Media11, but we also work with Corus and Rogers. And TELUS is a customer. And as far as our partnership with Bell goes, they’re not restraining us at all. In fact, they’re encouraging us to build a solution that works for the whole marketplace. And so, I’m very excited about that.
Shaw: Who would you put in the pantheon of advanced data users in Canada?
Kestle: I could name four significant retailers. Certainly a bank or two, and telecom. But we also see credit unions who are trying to compete with the big banks. We’ve done great work in the energy sector. We’ve actually had some federal government departments and municipal departments think about citizens and residents the same way brands think about consumers. So, the shining lights are organizations that really want to be data-driven. We have 1,000 customers — of those, I’d say 50 are on the right path. Maybe there are 20 who are amazing. They believe in data, and they’re really open to solving business problems differently.
Shaw: What’s your vision for Environics Analytics going forward?
Kestle: Well, we see our role as an accelerator and an enabler. Our plan is to continue to build more data faster, that is, weekly and monthly, instead of annually and quarterly, that’s number one. Number two, to have platforms that connect to the ecosystems so it’s easy for marketers to do things quickly. And then to invest in new technologies like a clean room, starting with the business problems people are trying to solve. We can give people the tools and they can do it themselves; or we can do complex projects for them, and everything in between.
We’re investing in more data production; we’re investing in better platforms. We have a great system called ENVISION, which is used by over 2,500 analysts and users and it’s great, but we’re rewriting it — moving it to the cloud, so that it’s modular, so that it connects to everything else. And we’re also investing in people who can help organizations implement the data strategy. And I think the next frontier is for us is to help people actually measure and know what works and what doesn’t. So, we are on a bit of a mission.
Stephen Shaw is the Chief Strategy Officer of Kenna, a marketing solutions provider specializing in delivering a more unified customer experience. Stephen can be reached via e-mail at firstname.lastname@example.org
Professor Richard James Webber was the creator of the geodemographic systems Acorn and Mosaic.
The study of geodemographics is often associated with the Chicago School of Sociology which supported the notion of classifying neighbourhoods based on shared characteristics.
An enumeration area is the distance canvassed by one census representative. An EA is composed of one or more adjacent blocks.
Tony Lea is a Senior Vice President and the Chief Methodologist at Environics Analytics.
Cluster Analysis is a statistical technique for classifying subjects into groups, or clusters, on the basis of how closely they resemble each other.
A demand-side platform (DSP) is an adtech solution that allows an advertiser to buy digital ads; a data management platform (DMP) is used to control cookie IDs and segment audiences in order to target them with specific ads.
Personal Identifiable Information.
The Consumer Privacy Protection Act will replace the federal Personal Information Protection and Electronic Documents Act (PIPEDA), which has regulated the collection, use and disclosure of personal information since 2001.
The Harper Conservative government cancelled the mandatory long-form census in 2010 and replaced it with the voluntary National Household Survey. Five years later the Liberal government reinstated the long-form census questionnaire.
A clean room is an encrypted, secure location where first-party data is anonymized and matched with aggregated 3rd party data.
Bell Media acquired Environics Analytics in 2020.