In this episode of On Point with Korem, Jan Kestle, the CEO and Founder of Environics Analytics discusses how geography is the secret sauce inside of data and analytics as well as the overall importance of space in data. Jan noted that there are more solutions to leverage spatial analytics and that certainly the technology is available to more people. As such, Jan discussed promoting how universities need to take an interdisciplinary approach to educating students about the use of geospatial technology and also about how this new era of mobile movement data provides exception insights.
The Power of Data and Analytics
Geospatial Needed in Business School Curriculums
Spatial Analytics is a Specialized Discipline
The Complicated World of Mobility Data
Mobile Movement Data: From Insights to Activation
Bell Canada’s acquisition of Environics Analytics
The Power of Data and Analytics
Joe Francica: So Jan before we get started talking about Environics and maybe some of the partnership things that we do, with Korem. I gotta take you back to when we first met each other and in the Compusearch days. I mean, did you ever expect that there’d be so much emphasis, so much interest in geospatial data? It seems to be exploding and my observation is maybe there’s more emphasis on data these days than even on software. I just think it’s remarkable just to see this evolution.
Jan Kestle: Well, I think for sure I couldn’t say I expected it. I’ve always…you know I’m a mathematician by training and I think that you can explain the world in terms of statistics and math. And so, I came to that Compusearch world with that perspective, and you may recall that the early days of Compusearch they were using spatial analysis to understand who lived around a retail location or what customers were likely to have certain demographics, even before GIS was proliferated on the desktop. So back then, did we know what we know now? Which is my favorite mantra is; “Geography is the secret sauce inside of what we do in data analytic”! And there have been lots of slogans over the years; “Everything happens somewhere” and “If we know where you live, we know a lot about you”. So I don’t think that we really understood and frankly yes , I don’t think people still really understood the importance of space in data, as you well point out there’s so much more data available and also there’s so much more capability to do analytics available so the technologies available to more people. I sit on the boards of five or six major universities, and we advise a number of colleges in Canada around the power of data and analytics. And they sit in computer science departments, or math departments…on a very small number of occasions the geographers get involved, but geography is an absolutely key element and location is a key element to data and analytics and location and intelligence is kind of like the new exciting thing. So it’s kind of funny for me since we’ve been doing it for decades.
Geospatial Needed in Business School Curriculums
JF: Well, you just brought up several different things that I wanted to ask you about but the first one was, when you say you sat on the boards of universities. What I’ve been disappointed with, is that geospatial technology and location intelligence have not been introduced as part of a curriculum in business schools. When I went into business school, I think I met one professor in real estate finance who could actually spell the word GIS and I really thought by this time, there would be a greater understanding of what you just said, which is the importance of geography and geospatial in business and I just haven’t seen that! Maybe your perspective is different?
JK: Well I don’t want to come off as a raving nationalist, but I do think that Canada has been a leader in some of the geospatial analytics. Very early adopters of GIS pioneers and so for many decades Ryerson University was the place to go and then there’s the Cogs group in out in Eastern Canada. There’s been a lot of focus in Canada on geospatial and in geoscience. And so, for example, at Ryerson University, the Ted Rogers School of Business has the center for The Study of Commercial Activity which is led by PhD geographer, Tony Hernandez, who came over from the Ryerson applied geography school, so that’s the obvious example. But I think there are some other advances as well, like one of the things that I’m very passionate about, and I’m advocating for in many roles that I’m in, is that integration of geography in the business schools. But the general problem we’re trying to overcome is this compartmentalization in learning. We want to teach students critical thinking. We love in the business community those really geeky people whom you can at least converse with about the business objectives and we love those business people who aren’t “snowable” on technology. And the only way you get that, because we know people are left brain, right brain and we know some people are going to have that ability to cross but most people won’t, so what I’m really passionate about, and I do see some progress in the programs we work with in Canada at trying to get an interdisciplinary approach around data and analytics, and I’m always the voice at the table for bringing the geographer. So, with U of T, with Queens, with Western, with York, with those business schools sometimes at the center. Sometimes with the I.T department. We’re seeing more cross-fertilization than say ten years ago.
JK: But it’s still a challenge, you’re correct, it’s a challenge, but it really needs to happen because if we don’t embed location in databases about customers and citizens, because that’s really what it’s all about (for me), it’s … analytics to understand how to make Canada or how to make the world a better place. So how do we do better healthcare delivery? How do we look after education? And also, from a consumer marketing point of view, people really do like to get messages about things that are of interest to them, and they really do like to be able to go into their local store and find the products and services that are of interest. So, I always like to talk about data for good, in the context of kind of what’s generally perceived to be good in civil society, but also what makes people’s lives easier in a very busy time when people are juggling kids and jobs and all kinds of other challenges…never mind the pandemic. The location intelligence, geodemographics, as we used to call it, all of these things actually can make people’s lives better, but we need the support from the education community to teach people how to think across the silos.
Spatial Analytics is a Specialized Discipline
JF: Yeah, I would agree and I have had conversations with Tony and have invited him or others at Ryerson because I agree with you. I mean I’ve known Ryerson’s excellence in that area for many years, so I hope to get him on (a podcast). But you also raise another interesting question.
JK: There’s also some pretty interesting stuff going on at U of T (Toronto), there’s a group there, Eric Miller, who used to do a ton of…has always been known for doing a lot of transportation planning and he’s got quite a core group of geographers and I hate to mention one, because I’m going to think of the five that I didn’t mention, but you know there are lots of, I would say, it’s maybe our time has come.
JF: Right, well and I used to be involved in a conference called Applied Geography, which I think we had Toronto, we had Ryerson, we had North Texas University, North Carolina, those schools seem to be more proactive. But to your point then about trying to do this cross-disciplinary work in education, it brings up the issue of educating a new crop of users. In particular, those not as familiar with geospatial technology. And I’m thinking primarily about the BI users. The people who now are asked to work with Tableau and Cognos and Information (Builders), that is of a spatial nature but they don’t have the background necessarily, think spatially. Because, as you know, I used to work for another Canadian company that had that tag line of “thinking spatially” and I don’t know whether it’s there yet. I don’t know whether they appreciate. I think they know how to make pretty pictures with Tableau. I’m not certain they know what they’re doing! So that’s my opinionated observation.
JK: Well, it’s a big challenge because what we do in spatial analytics, which is what I really like to call it, it’s a very specialized discipline. It makes use of geography; it makes use of operations researchers. I think that helping people understand the business problems they’re trying to solve, making sure that data are organized and are clean and are used effectively, is part of what the sort of BI people are coming to terms with. But I’m more of the point of view that from the business perspective, we have to have the right partnerships. Because everybody can’t be a specialist in some of these very detailed things. So in our company for example, we produce…we have a platform called Envision, which makes use of a whole lot of different kinds of tools, including ArcGIS and Alteryx and various other spatial algorithms, but we make it usable for the program manager or the marketer and not the data scientists! And we do the work of understanding what kind of dashboarding we want to use and what kind of visualization, and our geographers bring the spatial analytics into that bundled solution. Now that’s our choice. I’m not saying that’s the only way to do it but my experience is, for the people who are really experts in visualization, which is what’s so important in that last mile. We do great analysis but if non-technical people can’t go and do something then it’s very difficult, so I think that teaching some element of the importance of geography. like for me. what should be taught to the BI space is a lot more about what geospatial variables should be, and how to keep them clean, and how to keep them organized. But the element of the algorithms or the workflows, where you’re really going to say “if I have this many people who I see on their cell phone through privacy compliant permissioned use of location data. If I’m going to try to use that to project demand for a certain kind of cereal in a grocery store. You better bring some geographers to the table, and you have to have that combination of geographic expert.” I’m not optimistic about having all the people in the BI space understand the power of spatial analytics but firms like ourselves, and Korem, and others, are bringing those worlds together. What’s really exciting for me is, it’s a partnership game! And especially in Canada, for us, we’ve continued through Compusearch and Environics Analytics to help customers understand what we do, what Korem does, what ESRI does, what another part of the whole space…what everybody does, and tried as much as possible to identify…very often people see us as competitors, those different kinds of firms that I mentioned, but very often if you look at the Venn diagram, the area of overlap is very small. And the area of the union, i.e. the potential that firms that do different things, what we can do together to help businesses and governments, is much larger than the areas in which we overlap. So you get into some of this “frenemies” or “coopetition” where the element of spatial analysis that many firms are providing, is different, and if people tend to think Compusearch and MapInfo are the same business, before the acquisition. Or, if Environics and ESRI are in competition with each other, it’s not really true! The devil’s in the details! And the beauty of it is, that if you bring the partners together you can get some really amazing outcomes. And you know that’s part of what we do with Korem. They have a strong presence in Quebec. They’re a tech integrator. We’re not that at all! We produce data and we produce micro marketing analysis for businesses to solve business problems and we produce platforms that allow us to do that. But if the company comes along and says well we’d like you to help us redesign our whole system and bring spatial analytics into that, in an enterprise-wide system, then we would want to partner with a Korem, because they’re the technology integrators. And on the other side of it, when customers come to them and want to do that more integrated marketing or consumer insights analysis, they actually offer our data inside our envision solution. So what I think is really important, and I’m sorry…but what I think is really important is that it’s too complicated to make the customer navigate and so there have to be partnerships and businesses have to be explicit to the customers. Because you have a big bank or a retailer, they got a purchasing department, they look up a bunch of companies that do these six things and then you’re going through an RFP and people are wasting time and money. So I’ve taken the view in building Environics Analytics and it’s our job to explain to the market what we do and what our 27 partners do and where it makes sense to bring us together and so when you deal with a big BI department in it, in a big advertiser, you have to help them understand, where spatial analytics fits into the whole picture.
The Complicated World of Mobility Data
JF: Yeah, and to that point one, of the new data types that seems to have gotten a lot of attention, you’re very aware of it, is mobility data. Whether it’s traffic mobility or footfall data. And that’s just seemed to be another data type that we need to explain, maybe more in detail. Everybody understands the concepts that we carry around these little mobile devices but to be able to utilize that data in a meaningful way. It just could look like a bunch of points or a bunch of lines, and where’s the answer in all of that data, right! They don’t want to see the pretty lines and the dots. They just want the answer.
JK: Yeah, I mean it’s a very exciting and very complicated world. We’ve had an R&D team evaluating the use of mobile movement data for consumer insights and location intelligence for over four years, PhDs and MA’s. And now, also speaking specifically about Canada, because we do it for Canada, and for the US, and we work with a number of partners in that GPS location data field. But one of the things that was very important to us, and we and we really went through a lot of learning, is that when people talk about “oh we’ve got 300 million or 10 million or we’ve got this many phones were tracking” if you really want to understands, if you really want to use the ping of a phone inside a pizza restaurant to understand how far people drive and what the nature of those customers are, you have to think about how frequently is that phone going to be there and the geofence…what we used to call the polygon…that you’re looking at to see whether they’re inside. We need to create a subset of those mobile phones to make sure that it’s reliable enough to make a predictor. So some of the early debates about how many phones are you covering and how many people do you have? It really comes down to a lot of methodological work to decide what’s a valid way of understanding a consumer’s path, or a citizen’s path, if you’re talking about transportation. So what we did is, we really wanted to make sure that what we were seeing…so number one; we have a product called MobileScapes and it’s 100% privacy compliant. So we buy data from the location data providers. We integrate a database for Canada that has sufficient sample size to make some reasonable assumptions. We actually weigh the data using some…I mean so exciting, you’d love to see it because you’re one of the people that understand it…but we weigh the data using both geodemographics and kind of proximity assumptions about what does this phone that I see inside this geofence, how many people can it actually represent^ So in old-fashioned geodemography with survey, we would use something like PRIZM or some demographics to say, “well if this person can be used as an example of their type, we can make a typological projection”, that’s the geographic part. Except if you’re doing it with this massive big data source, you have to look much more locally at how you’re doing that projection. And so we’re bringing these data down from a number of sources through really big data sources in the cloud. We’re integrating them. We’re bringing them down daily. And, we’re producing weekly results of consumers and citizens, and where they go, using a proprietary set of geo fences that we’ve built. And making estimates for gen pop because you’re not your cell phone and your cell phone…what gets captured is not representative. So long answer but the key with any big data source, whether it’s mobile movement data or other kinds of tracing data, it is what it is. Like the early debates around big data was “well now that we’ve got all this big data, we don’t need any of that other stuff, like survey data and government data”. My mantra is; reliable survey data, and administrative data, and census data makes big data usable, so we are mashing up big data from mobile phone sources and others. Moving the locations so that you’re not actually looking at rooftop because we want… we know that small anonymous aggregates protect the privacy and the trust of individuals and they’re also very reliable statistically, and then we put them into models that we benchmark and weight with data that we know is kind of verifiable in the real world and doing it with cell phone movement data is just the beginning. There’s all kinds of data sources where we can bring that discipline together but you know methodology and quality and data standards have to be a part of the conversation, not just the size of the data and technologies, and I’m always unpopular at conferences when I say, “I love the investment that we’re making in technology and I you have these brilliant people who can do all kinds of things but if they don’t understand the methodology”! Like I say to people, “Okay, I’m gonna give you our data, which has a lot of attributes about citizens and consumers. I give you 30 000 variables for every six digit postal code in Canada but if you’re going to attach that to eyeballs in the digital ad space for example, which is kind of part of the location intelligence. I say to people how are you going to make that match”? And they say, “Oh well, we have data scientists”, and I say, “well that’s good and what do the data sciences do?”…“well we have algorithms”. But the methodology has to be primary when you’re harnessing big data. And as with Korem and ourselves, we’ve had lots of conversations and lots of exchange with many other spatial data analysts, and with Statistics Canada as well, to try to get the conversation to focus on…you don’t take one kind of data and throw it out and lurch over into another. You have to have quality standards.
Mobile Movement Data: From Insights to Activation
JF: Yeah, so you may have already answered a question I wanted to ask you about mobile trace. Because of the spatial resolution, because of how detailed you can get with some of that data and can develop audience profiles, does that data render what we would call our traditional psychographic information obsolete? Or, does it enhance something like a PRIZM or is it just augment…a geodemographic segmentation system?
JK: Yeah, we use both together. I think one thing I’m proud of is you kind of say, “well maybe geodemography is a bit of old school, like we don’t need that postal code stuff anymore because we’ve got real data”. For me, bringing the two worlds together has really driven our growth and our business, even through the pandemic, has grown tremendously in the past year. And the single most thing I can attribute it to is combining things like mobile movement data with things like PRIZM. Because here’s the deal; Number one, if you’re in the ad space, you want to think about your customers in terms of the omni channel. Number two, you want to go from insights. Understanding who my customers are and different subgroups of the population. You want to go from insights to activation. So you want to be able to get the right message, to the right people, at the right time, in something like a PRIZM or the other commercial off-the-shelf segmentation tools can be the glue between understanding from my own CRM, my customer database, some sub-segments of the population that are customized because I’ve used a PRIZM to assemble and organize my own data, and then because the other thing we do is we use PRIZM and our other third-party databases to link into surveys and media measurement. So now I’ve got my own customer database. I’ve got all of the important media and marketing databases in Canada and the behavior databases and the psychographic databases that I can express in PRIZM terms. Now I can add PRIZM profile of people who were inside my geofence based on the tracing and the spatial algorithms we embed there. So, I can design a program, the right message, the right medium, across multi-channels. I can see who actually went inside a store by PRIZM group, so it’s still privacy compliant, but I’m doing attribution if I targeted my campaign to these groups that were based on my CRM. I used information about media preferences to know how to reach them. I can actually use the mobile movement data to see who visited the store and other kinds of big data sources to see who clicked. I can do all that at the detailed level of the segments and some people would say, “well yeah, it would be so much better if we could do it one-to-one” but when you start doing that “one-to-one” you’re dealing with the conditions under which the consumer gave that data. So my view for the future is the base PRIZM kind of stuff, combined with all these big data, is going to show the whole universe or the denominator of the equation. But the data that brands and organizations have, the one-to-one data that they have, is kind of the numerator. So, we’re going to use both. We’re not going to say real data “so-called” is better than model data because the real data might be biased. It might not give you market share. So the exciting thing, and the reason I think why our business is really flourishing, is because we’re combining what we learned from the old world but still making it usable in this era where we have so much one-to-one data. Because remember that one-to-one data that a consumer gave, or the permission they turned on in their cell phone, there’s permissions, there’s restrictions, and so we have to use all of that within the context of the trust with the consumer or the citizen, about what data you’re going to use and I believe people are willing to have us use their data (brands and governments and organizations), have us use their data in a responsible way that protects their privacy but still makes their life better. And combining those two things is really the exciting way that we get to do that.
Bell Canada’s acquisition of Environics Analytics
JF: Yeah, I know that’s really interesting. One last question before I let you go. I want to just ask you about the Bell Canada acquisition and maybe what your vision, or what Bell’s vision was for the acquisition? And particularly in the data space. It just seems to be a really interesting merger.
JK: Well I think we have a shared vision and that was the basis for… it’s hard to call a merger when you combine 250 people with a 52 000 person organization…so Bell bought the business and we’re running Environics Analytics as a separate entity that still is really pursuing the same mandate which is to provide data and analytics to the whole Canadian marketplace. So we deal with Bell and we deal with their competitive customers in the media and in the telecom space, completely on an equal footing. But Bell’s investing in Environics Analytics to provide data that deal with some of the complexities of the Canadian marketplace. That understand the Canadian privacy laws, which have their own challenges, and investing in EA’s ability to provide that universe denominator of a lot of data about what happens in Canada. And the investments that they’re making in our business and the things that we’re doing together are enabling us to do the kinds of combinations of big data and traditional data. They’re enabling us to build new platforms there and one of the other things that is happening because we’re accessing data sources from Bell and from others, again making sure that we’re following all the rules about privacy and permissions, but we’re accessing anonymous aggregated data from telecoms to drive some of these products and we’re able to put data out in more in real time. So the division from Bell was to support us, to grow the data business, to leverage the assets that they have, and even in the first year we’ve come up with some exciting new enhancements that are enabling, especially during the pandemic, enabling EA to move from producing data annually or quarterly, to producing weekly and monthly results about consumer and citizen behavior. That people can actually go and put into action and make a difference to help make people’s lives better. And I still would maintain at the end that geography is the secret sauce inside all of that. I mean I’m sorry, I know you got to go but we have 250 employees and I think more than 100 of them have some degree in geography.
JF: Oh wow, that’s awesome.
JK: We got data scientists, we got programmers, but it’s the geographers that are at the core! And it’s not necessarily a geography problem or report at the other end, but it’s happening under the covers all the time.
JF: Well I’m not going to disagree. My degree was not in geography, it was geology but same space.
JK: Yeah me too! Mine was in math so here’s me talking about geography.
JF: Yeah, I was just totally self-taught, what can I say! Well Jan, thanks very much. Really appreciate your time and certainly we appreciate the partnership with Environics.
JK: Great thanks so much for having me.
JF: Thanks Jan, take care.