It happens only once every five years:  On February 8th, Statistics Canada will release the first results from the 2011 Census showing the population and household counts for all geographic areas across the country.  So what can we expect from the latest count?

The short answer is that the 2011 Census  is expected to show a continuation of some population growth trends but also some turnaround stories. 

To be specific, the results are expected to show that Canada has a  population of about 33.5 million, up nearly 6% from the last census in 2006. This would represent the highest growth rate among G-8 countries and the highest growth rate for Canada since the 1986-1991 period. Today, much of the story about population growth can be traced to immigration since nearly two thirds of Canada’s growth is accounted for by international migration.

The Census also is expected to show a continuation of some long term trends. For many decades, Canada’s population has gradually shifted to westward regions, particularly Alberta and British Columbia. For example, 50 years ago the 1961 Census showed these two provinces accounting for just over 16% of Canada’s population. The new census is expected to find that their share of the population has increased to about 24%. At the same time, Atlantic Canada’s share is expected to account for just about 7% of Canada’s population—down from over 10% in 1961. Quebec will also likely show a continuation of the decline in its share of population to around 23.5%—down from over 29% in 1961.

Not surprisingly, Alberta will continue to be Canada’s fasted growing province followed by British Columbia. However the turnaround story of the Census will be Saskatchewan. The last Census in 2006 found that Saskatchewan had lost just over 1% of its population since 2001. This time, Saskatchewan is expected to show above-average growth and rank as Canada’s third fastest growing province. Other provinces likely to have undergone faster growth compared to 2001-06 are Prince Edward Island and Manitoba since both have attracted an increased number of immigrants in recent years. On the other hand, growth in Ontario is expected to be lower for 2006-2011, and likely below the national average, since Ontario’s share of immigration in recent years has declined. In fact, Ontario may show a slight decline in its total share of the Canadian population from 38.5% in the 2006 Census.

Across metropolitan areas, population growth will continue to be highest in the largest areas that attract immigrants. The western CMAs of Vancouver, Calgary, Edmonton, Saskatoon and Regina are expected to be the five fastest growing CMAs, with their population growing about 10% or higher. In fact, Saskatoon may rival Calgary as the fasted growing CMA. As a result of immigration, the Greater Toronto Area will also likely report  close to 10% growth. As in the past, the suburban areas of Peel, York and Halton are likely to lead the GTA in growth.  At the other extreme, the effects of the recession   are likely to be reflected in very low, if not negative,  growth in industrial centres like Windsor, Sudbury, Thunder Bay, and St. Catherines-Niagara  in Ontario and the Saguenay area in Quebec. Another major turnaround story is likely to be Barrie, Ontario, which was by far the fastest growing CMA for the decade 1996-2006. The area’s 50% population increase had been due in part to economic and technological opportunities. But in the latest census, Barrie is expected to show average growth of about 5-10 % for the 2006-2011 period.

At a more detailed geographic level, the census will also show the population churn in urban neighbourhoods. In some locations, new high rise condominiums have sprung up and attracted young singles and couples. In some of the older suburban areas developed in the 1970s and 1980s, we expect to see smaller households as older children finally flee the nest. On the other hand, homes formerly occupied by empty nest couples or widows may now be welcoming young families. Such changes in households and attendant lifestages will have clear implications for marketers.

Indeed, businesses can use the census statistics to better understand their markets and improve their profitability. By calculating their sales according to the new CMA and provincial counts, they can determine their market shares for each area and recalibrate sales targets. With access to new population and household data, retailers can better decide which locations are best for new or expanded store operations. For businesses, government agencies and not-for-profits, the latest demographic data contained in the Census will give them better insight into their customers and trade areas and ultimately lead to improved marketing, merchandising and decision-making.

Please join Doug Norris for a webinar highlighting the new Census release on population and household statistics, on Tuesday, February 14th, at 2:00 PM. Register Here

# # #

One of Canada’s leading experts on the Census, Doug Norris, Ph.D., is a Senior Vice President and Chief Demographer at Environics Analytics. He joined EA in 2006 after nearly 30 years with Statistics Canada, where he earned the nickname of “Mr. Census” in his role as Director General of Social and Demographic Statistics. Currently, he assists companies, government agencies and not-for-profit organizations in using census and other statistical information for planning and marketing projects.

The recent release of geographic products from the 2011 Census marks the first new census data to become available from Statistics Canada. The arrival of boundary files at the lowest levels of geography—dissemination areas and dissemination blocks—means that users will soon be able to visualize updated demographic characteristics at the neighbourhood level. That’s good news for anyone looking to refine a marketing program or analyze changes across locations and over time.

The dissemination area (DA) is the smallest standard geographic area for which all census data are distributed. DAs are composed of a relatively small population, usually between 400 and 700 people. According to the 2006 Census, Canada consisted of 54,626 DAs. In addition, DAs are at the heart of Environics Analytics’ PRIZM segmentation system, with each DA assigned to one of EA’s 66 unique PRIZM segment.

For the 2011 Census, one of Statistics Canada’s objectives was to keep DAs as stable as possible while not allowing DAs to become too large in terms of population. For this reason, and because DAs must respect higher geographic levels—such as census subdivisions (typically municipalities) which may change boundaries—some DAs are modified from one census to the next. According the 2011 Census, there are now 56,204 DAs inCanada, an increase of 1,578 since 2006.

Of the 54,626 DAs identified in 2006, 51,785 (95%) of these geographic units remained the same for 2011. The remainder were modified in various ways. A total of 583 were split into two or more DAs: 313 were split into two DAs and 270 were split into three or more DAs. The biggest split occurred in Milton, Ontario, one of the fastest growing municipalities in the Greater Golden Horseshoe, where one 2006 DA was split into 21 DAs for 2011. An additional 2224 DAs were subject to more complicated changes where parts of one DA were assigned to one 2011 DA and other parts were assigned to others. These were often the result of boundary changes at higher levels of geography. Finally, there were 34 2006 DAs (17 pairs) that were collapsed into 17 DAs for 2011.

Approximately one quarter of the 2006 DAs that were changed in some way are located in small town and rural areas. Many of these shifts were likely the result of municipal boundary changes. Most of the remaining DA changes resulted from increased population in rapidly growing suburbs around large urban areas. One exception was the urban area ofCapeBretonwhere virtually all the DAs were modified in a special redesign of the DA boundaries in order to better match local areas of interest.

We can also look at the DAs that were changed in terms of their 2006 PRIZM segment assignment. Not surprisingly, nearly one quarter of all child-filled Pets and PCs DAs and 17% of Mini Van & Vin Rouge DAs have modified boundaries for 2011. These segments are dominated by new suburban areas with young families. Other segments that had considerable change were New Homesteaders (rural midscale couples and families), Big Sky Families (middle-aged, midscale Prairie farmers), and Grads and Pads (young, lower-middle-class urban singles). When the new version of PRIZM based on the 2011 Census and National Household Survey data is released in early 2014, some DAs might change from one segment to another, reflecting demographic changes in the last five years. In addition, new segments might be introduced and others retired from use.

The 2011 Census counts for the dissemination areas will be released on February 8, 2012. Since 95% of DAs have not changed since 2006, this means that it will be possible to directly compare the 2006 and 2011 populations for most the DAs in Canada.

# # #

One of Canada’s leading experts on the Census, Doug Norris, Ph.D., is a Senior Vice President and Chief Demographer at Environics Analytics. He joined EA in 2006 after nearly 30 years with Statistics Canada, where he earned the nickname of “Mr. Census” in his role as Director General of Social and Demographic Statistics. Currently, he assists companies, government agencies and not-for-profit organizations in using census and other statistical information for planning and marketing projects.

To be completely honest, I forgot that they made a 5 o’clock in the morning. Had I not double-checked the two alarms I set before going to sleep the night before, I would have sworn that I had been asleep for five minutes tops. But it was the day of the Environics Analytics User Conference and there was a lot to do before the official 8:30 start. In addition to waking up at the crack of dawn, I had been “voluntold” a day earlier to write a blog entry documenting my first user conference, which only added to the quiet nervousness I already felt about the day ahead.

You have to understand where I am coming from here. At 23 years old, I have spent three-quarters of my life in school. Since saying my goodbyes to academia less than a month ago, I have been working full-time for Environics Analytics (EA) with a great team of people involved in everything from research to marketing. Everyone has been welcoming and encouraging, and I truly could not ask for anything more. But this would be the first time I would experience the culmination of months of hard work and take part in this celebration of minds.

Grace—and Wool Overcoats—Under Pressure

My assignment for the Environics Analytics 2011 User Conference was to work coat check. Some of you may laugh, but I’d argue that manning the coat check at a conference, particularly as a newly hired employee, is one of the most important responsibilities. Others may say that it is a rubber-room of sorts, a place to collect all the low people on the totem pole. What you may not realize is that I, along with the others tasked to work coat check, made up a select group of EA employees who got to interact directly with almost every single conference guest. It was really great to see how diverse the group was and to recognize the common thread of excitement; everyone was eager to hear from a range of speakers and industry leaders.

While I was only able to sit in on one presentation, that of Dr. Brenda Jones, Chief Social Scientist at Parks Canada, many conference attendees were particularly vocal about how much they enjoyed the variety of topics and discussions on the agenda. A number of them also noted that this was the best EA conference they had attended; they said they were looking forward to seeing the new product offerings like BigMaps Canada and learning about how social media is being incorporated into segmentation-based marketing.

Cloak Room Confessions

In addition to meeting the attendees, I got to know my coworkers on a different level. Even though I have been in and around the office for the past ten months as a part-time staff member, it can be difficult to strike up conversations with people during the workday. But in the coat check room, a few strategically placed chairs proved to be an irresistible draw for my fellow employees looking to take a quick break during the event. Among the overcoats, my coworkers told me how the iconic EA Geek came to be (legend has it he wandered into the office and began creating complex algorithms for the fun of it), how he has evolved (two years ago he was given a girlfriend; this past year the Geek couple unveiled their daughter), and the marketing role he plays (the go-to guy for any company’s data challenges). I heard about a couple of coworkers’ educational backgrounds (sociology, environmental studies), what they do outside of work (travel abroad as often as possible), and some of their aspirations (graphic design, marketing). And they were all quick to talk about all the fun they have had working at EA.

Perhaps the most fitting conversation ended my day. The very last guest to pick up his coat gave me some simple advice for my tenure at EA. He told me how important it is to continue to learn from my peers, and I can tell you with utmost confidence that I will have no trouble finding a wealth of information and creativity within a stone’s throw from my desk. After all, we’ll always have the coat check room to remember.

In the end, I came. I saw. I hung up coats and I conquered — and I hope to do it all again next year, by which time I am sure I will conveniently have forgotten all about that 5 o’clock-in-the-morning thing.

# # #

A Client Advocate, Uschi Erne helps transform EA research and analysis into successful marketing programs for customers in the finance, insurance, telecommunications and travel industries. She is expert at uncovering critical consumer patterns and profiling target groups to help clients communicate with their customers and reach untapped markets.

This blog post is long overdue. And I don’t mean that in an “I’ve been meaning to write this for a long time” sort of way. I mean it in the way that I was told to have the post written by August 2nd. That’s because my topic, the Alteryx Grand Prix, occurred last March. My only defence is that some events can be so life-altering that they require time and distance to put them in proper perspective. That, and I just kept putting it off.   

Some of you probably know that the Alteryx Grand Prix is a fixture at Alteryx’s annual conference in Colorado, where representatives from ten companies compete to solve complex business problems using Alteryx’s geographic business intelligence platform. The whole event is way over the top. It’s hard to do it justice without pictures, but imagine a hotel ballroom decorated with racetrack-like grandstands, stacks of tires and fencing, and an emcee doing breathless play-by-play over a PA system for the conference attendees. Giant screens even show footage of what the contestants are doing at their computers—like a helmet cam shot from inside a NASCAR racer. The programmer who can solve a series of problems—called “laps”—in the fastest time is declared the winner.

Really, though, at the end of the day the Alteryx Grand Prix is just a programming competition—albeit a geeky contest with a very impressive reward. The grand prize consists of the following:

1)      A trip for two to Las Vegas
2)      Three nights at a fancy hotel
3)      A rental car for the weekend
4)      A day at the Mario Andretti Speedway racing “Indy-style” cars
5)      $500 spending money
6)      An enormous trophy
7)      The right to celebrate with said trophy by drinking out of it at the bar with Alteryx employees
8)      The right to feel like a minor celebrity for the remainder of the Alteryx conference

Obviously, this is the kind of glory any programmer would want. So in 2010, I competed in my first Alteryx Grand Prix, finishing a respectable second. Figuring I had a better chance to win this year, I developed a methodology that I believed would increase my odds.

Know your tools: I work with Alteryx more than I like to admit—not because I’m ashamed of it, but because most people outside of my office have a difficult time understanding what Alteryx is, which makes explaining what I do every day a little awkward. I probably know more about the roster of Alteryx tools than I do about the Toronto Maple Leafs starting line-up—knowledge that definitely helps in the competition but hurts in dressing room conversation when I’m suiting up to play on my beer league hockey team.

Example:

Teammate: “Did you see that big save by Reimer last night?”

Jason: “No…did you see my save last night? Saved 20 seconds on a 35-second wizard by pointing the spatial match tool directly to a yxdb.”

Teammate: (Leaves dressing room midway through my sentence.)

Study: I printed out most of the Alteryx help document and studied in my free time during the week prior to the event. The excess studying was probably a result of a conversation I had with Jan Kestle, the Founder and President of Environics Analytics, which went like this:

Jan: “Are you nervous about the Grand Prix?”

Jason: “Not really. I know that no matter how well I do, you guys will be proud of me.”

Jan: “(Long pause.) Yeah, I guess so—just make sure you come in the top three.”

Cram: At the conference, any spare moment I had prompted me to disappear into my hotel room, read over materials and explore tools that I wasn’t quite as familiar with. I realized that I might be taking things a bit too seriously when I had a conversation with another competitor, and he pretty much laughed at me for this last-minute cramming. He declared that he hadn’t been studying, nor had any other competitors. Go ahead and laugh, my friend, I thought. We’ll see who has the last laugh. (Cue maniacal cackle.)

Get as much help as you can: Your pit crew and the referees at the Grand Prix are there to help you. Listen to what they say because the information from a thirty second-conversation with them could save you five minutes in the future.

Get lucky: I realized very quickly during the actual competition that all of my studying was more or less a waste of time when most of the questions initially left me scratching my head. I ended up settling on the simplest solution I could come up with, and then I hoped for the best. As it turned out, I did not have the best answers among the competitors, but I did have consistently slightly-better-than-mediocre answers. Meanwhile, the other competitors who had good answers for some questions had terrible answers for others, so that at the end of the race, my slightly-better-than-mediocre answers tallied up to the best time. Not the most glamorous way to win, but I’ll take it!

Celebrate: The winner’s trophy is by far the largest trophy I have ever won and features a cup big enough to fit 2.5 beers (believe me, we measured). Drinking out of a trophy cup has always been a dream of mine, and while, no, it’s not the Stanley Cup, that doesn’t mean you can’t pretend.

Repeat Performance?
Now that I’ve won once, I’m occasionally asked if I will be competing in this event next year. Frankly, I’m still undecided. On the one hand, it would be nice to go out on top and give the other great Alteryx users at our company a chance to compete. On the other hand…VEGAS BABY!!!!

Unfortunately, my racing gear made me look less like a cool racing star than a toddler in a jumpsuit.

“CA-NA-DA!! CA-NA-DA!!”

# # #

Jason Dunkel is Product Manager of ENVISION, EA’s web-based platform for consumer insights and market analysis, and is responsible for developing applications and reports for customer solutions. As a programmer, he assisted in the development of ENVISION and continues to provide custom solutions to clients’ marketing challenges. Nine months after his triumphant win, he still takes occasional victory laps around the office.

These days, most companies have some sort of strategy for marketing to newcomers and immigrants to Canada. The marketing landscape in Canada is an ever-changing tapestry of young and old, newcomers and longtime residents of the country.  And as marketers develop their understanding of how to communicate to the population, the shifting trends in immigration have the potential to change the rules. Today, organizations in all sectors need to understand the changing cultural diversity of Canada and the challenges it presents in delivering products and services to Canadians.

On November 1, the Minister of Citizenship, Immigration and Multiculturalism presented the Annual Report to Parliament on Immigration ( http://www.cic.gc.ca/english/resources/publications/annual-report-2011/index.asp).  A key part of the report is the announcement of the immigration levels planned for 2012. The report indicated that, for 2012, the overall admissions range for permanent residents will be 240,000 to 265,000. This is the same range as for the last 6 years. So one thing we can be sure of is that there will be about a quarter of a million new consumers in the Canadian marketplace during the course of 2012.

The actual number of immigrants to Canada fluctuated between 221,000 and 262,000 for the   period 2001-2009. However, in 2010 total admissions of permanent residents to Canada were 280,681. This exceeded the planned range and was the highest level of immigration in 50 years.

For marketers to truly tap into these new markets, the shift in the type of immigrants coming into Canada is important to note.  It’s not enough to try to market to “new Canadians” as a whole; marketers today need to understand where they are coming from and how they are behaving in the marketplace as they become a part of Canadian society. While Canada receives its immigrant population from over 200 countries, in 2010, just over 50 percent of new immigrants were from 10 source countries. Also in 2010, the Philippines became the top country of origin for immigrants (36,578) surpassing both India (30,252) and China (30,197). Immigration from the Philippines has increased substantially over the decade from a low of 12,928 in 2001 to nearly triple the number today.

The big cities continue to be the largest magnets for immigration, but between them, the pattern of settlement changed substantially over the past decade..  In 2010, one third of all immigrants settled in the Toronto Census Metropolitan Area (CMA), 17% in the Montreal CMA and 13% in the Vancouver CMA. The big change was a drop in the proportion of immigrants going to the Toronto CMA, from 50% in 2001 to 33% in 2010. A further analysis suggests the drop was a result of lower immigration to the City of Toronto (down about 50%) in addition to increased immigration to the suburban 905 area. Most other parts of the country saw their share of new immigrants remain about the same or slightly increase. Being able to understand the needs of these growing suburban immigrant populations will separate those marketers who are successful from the rest of the pack in the next few years.

But it’s not only those who are coming to stay in Canada who represent an opportunity; there are also those who stay for a while, then leave. In addition to the trends in the number of permanent resident immigrants, over the last decade there also has been a substantial increase in the number of temporary workers and students in Canada. As of December 1, 2010, there were 282,771 temporary workers in Canada—up from 96,390 in 2001. In addition, there were 218,161 foreign students in Canada—an increase from 136,495 in 2001.  Being able to see the patterns where these workers are going and which schools these students are selecting will allow the astute marketer to find opportunities and provide services to these itinerant residents.

If CMOs and agencies are to continue to put out relevant and effective marketing to these new Canadians, they need to understand and develop strategies that appeal to the unique tastes and sensibilities of these markets.  The rules are changing all the time. And keeping a finger on the vibrant pulse of the Canadian marketplace will be more important than ever in 2012.

# # #

One of Canada’s leading experts on the Census, Doug Norris, Ph.D., is a Senior Vice President and Chief Demographer at Environics Analytics. He joined EA in 2006 after nearly 30 years with Statistics Canada, where he earned the nickname of “Mr. Census” in his role as Director General of Social and Demographic Statistics. Currently, he assists companies, government agencies and not-for-profit organizations in using census and other statistical information for planning and marketing projects.

This week, more than 150 people are participating in the 8th Annual New American Mainstream Business Summit in Miami, Florida.  The summit’s theme “Multicultural Goes Mainstream” reflects the booming ethnic populations in the U.S. By 2017, the population of the three largest ethnic groups—Hispanics, Asians and African Americans—will be nearly 121 million, or 34 percent of the population. Already, Hispanics are the largest minority group in the U.S., which explains why this year, the most popular name for newborns is José.  By 2040, ethnic groups will be the majority of all U.S. citizens and the country will be a “majority minority.”

The current geographic dispersion of Hispanics is bucking past trends. In 1990, most Hispanic immigrants settled in gateway states like California, Arizona, New Mexico and Texas. By 2012, however, majority concentrations will be found in 356 counties in states as far north as Minnesota, Montana and Washington. This booming population is driving greater consumption of a range of products favoured by Hispanic consumers: mobile phones, energy drinks, gum, low-sugar foods, casual dining restaurants and car and life insurance. But it’s also raising challenges for marketers trying to determine which language, message and offers to use to connect with Hispanics across the acculturation spectrum: from recent immigrants to the children of second-generation Hispanics who are more immersed in American culture than their grandparents. “We’re still at the beginning of multicultural marketing,” one speaker after another noted at the summit. “Analytics must take center stage to provide answers to these questions.”

http://www.geoscape.com/summit.asp

# # #

The Vice President of Marketing, Michael J. Weiss is responsible for increasing the awareness of Environics Analytics through media and marketing initiatives. An internationally known expert in geodemographics, he directed the naming and imaging of the PRIZM C2 and PRIZM CE clusters, and he has written three books, numerous articles, ads and data descriptions for EA products and services.

It’s pretty hard to separate geography from maps—the two are inextricably linked. But just why are maps so darned important to geographers?  I would argue that a map—as the title of this blog suggests—allows you to tell a story.  There’s something very powerful about a map that you don’t quite get from other types of graphics.  Perhaps it’s because when you map a variable, you’re “tying it to the ground” which effectively “grounds” the data, making it more real to the end-user.  Or maybe it’s just hard to resist the visual appeal of a well-made map.  Oftentimes, clients will ask us to produce maps that they can insert into their PowerPoint presentations, and I don’t blame them because maps can be persuasive, even to the toughest crowd.

What’s the toughest crowd I can think of off the top of my head?  Teenagers.  Sarcastic, disinterested, know-it-all teenagers.  To all the parents of teenagers reading this blog, I’m about to do you a favour.

Let’s start by looking at a map of the Toronto census subdivision (CSD).  The variable being mapped here is Population 15 Years or Over by Educational Attainment; specifically, the percentage of the population, by dissemination area, with no certificate, diploma or degree.

Programming note: all maps in this article were produced by ENVISION, Environics Analytics’ powerful online micromarketing tool.

This is what’s called a choropleth map; essentially, the darker the area, the higher the concentration of whatever variable is being mapped.  In the above map, educational achievement—or rather the lack thereof—appear as the dark purple areas. The darkest areas, representing about one-third of the population, are made up of high school drop-outs.

Next, let’s look at the same geography. But this time we’ll map average household income.

For this map, the darker the area, the higher the average income of the households in that neighbourhood.

Now, these two maps are pretty impressive on their own, but put the two side-by-side and a more revealing story emerges.

This is my favour to all the parents of teens.  Next time little Sally or little Johnny whines about having too much homework, show them these two maps.  The relationship is clear as day; simply put, high school drop-outs earn less money—money for iPad apps and ice cream, smartphones and cake pops.  A long-winded lecture would never drive the point home quite so quickly or quite as starkly as these two maps.

Personally, one of my favourite parts of being a geographer—and possessing the skillset to create maps in particular—is having the ability to illustrate even the most complex relationships in a format that everyone can understand. Even teenagers.

# # #

As a Research Analyst with the Standard Research team, Melodi Moini helps clients solve their marketing challenges.  She has expertise in a number of sectors, including retail, not-for-profits and government agencies, and previously worked as a GIS technician at De Beers Canada Exploration.  She holds a bachelor’s degree in geographic analysis from Ryerson University.

WealthScapes is a balance sheet representation of neighbourhood wealth: an estimate of a neighbourhood’s collective liquid assets, real estate assets and outstanding debts. The database contains essentially two classes of variables: aggregate balance variables (i.e., aggregate balance of GICs, term deposits and outstanding credit card debt, as well as aggregate value of primary real estate) and incidence variables (i.e., number of households with GICs, term deposits and outstanding credit card debt, as well as the number of households that own their primary residence). Additionally, WealthScapes includes a household count for each neighbourhood as well as various income statistics.

Generally, most geodemographic research is conducted at the household level. Consider a marketer who wants to target households with high mutual fund balances. To convert the raw WealthScapes data into an actionable statistic, the marketer would need to turn the aggregate mutual fund balance of the neighbourhood into an average statistic. While the numerator of this average statistic is the aggregate balance of the target product (in this case mutual funds), there are choices with respect to the proper denominator. For example, the denominator could be the average mutual fund balance per household (meaning all households, whether they own the product or not), or it could be the average mutual fund balance per holding household (defined as a household that either owns the asset or is the holder of the debt). Similarly, it could be just as reasonable to calculate the denominator based on the average mutual fund balance per investment holding household (households that own one or more investment products). All three statistics have merit and it is the responsibility of the analyst to determine the most appropriate denominator for the research at hand. With that challenge in mind, this blog has been written to guide the analyst in choosing the proper denominator.

Average Balance Per Household

The average balance per household is the best statistic to use about 90% of the time. Usually, marketers want to target spatially defined groups of households (i.e., neighbourhoods, postal codes, trade areas or cities). In order to compare multiple groups of households, the simplest approach is to imagine what the “average” household looks like. By dividing the aggregate balance by the total number of households, the average balance per household captures two separate effects at once: the propensity of households to hold this asset or liability and the average balance of those households that hold this asset or liability.

Here’s an example of a good case for using the average balance per household. A marketer wants to sell mutual funds to a neighbourhood and has determined the best prospects are those households that already hold mutual funds. The marketer’s goal is to sell the highest mutual fund balance (dollars). Assuming that the cost of targeting and response rate of a single household are fixed (constant), the marketer would target those neighbourhoods with the highest average balance per household first. Because the marketer cannot target those individual households with mutual funds, it is best to target neighbourhoods where the propensity to hold mutual funds is high, the typical average balance of mutual fund holders is high–or, preferably, both. Effectively, the average mutual fund balance per household has already made the requisite trade-off between high propensity and high balances in order to maximize expected revenues.

Average Balance Per Holding Household

The average balance per holding household is the best statistic to use about 10% of the time; typically when the marketer already has some information about the target audience (for example, a list of homeowners or clients who already hold investment products). This statistic is calculated by dividing the aggregate balance or value of the targeted product by the number of households which either hold this product or hold another appropriate product. This is a much more subtle statistic and generally takes one of two forms: the average balance per holding household of the targeted product or the average balance per holding household of another product. An example of the former is dividing the aggregate mutual fund balance by households who hold mutual funds, whereas an example of the later is dividing the aggregate mutual fund balance by households that hold one or more investment products (stocks, bonds and mutual funds). As the calculation implies, these statistics do not address the neighbourhood’s propensity to hold these products but rather only addresses the average balances of holders.

Continuing with the example developed previously, let’s say that the mutual fund marketer already has a collection of mutual fund holding customers in the target neighbourhood whose average mutual fund balance is, say, $40,000 per household. But the average balance per holding household in the neighbourhood is higher, say, $50,000. This means that either (a) the marketer’s clients likely have $10,000 on average in additional mutual fund holdings that can be targeted for conversion to the marketer’s company or (b) the marketer’s clients tend to have lower mutual fund balances than the average holding household in the neighbourhood. The truth is probably a mix of the two situations and it is the marketer’s role to determine (usually by corollary evidence) to what degree one explanation is more accurate than the other. If the marketer decides that (a) is the likely cause of the deviation, then the marketer can target her or his clients with a conversion campaign to consolidate all of the mutual funds with the marketer’s firm. If (b) is true, then the marketer can fine tune a marketing message to the remaining households in the neighbourhood—perhaps extolling the virtues of the marketer’s more premium offerings.

Average Balance Per Investment Holding Household

Alternatively, one might consider a denominator that is not the product’s holdership count. Empirical evidence suggests that households can be put into one of two general classes: debtor households and investor households. An investor household would be a household that holds any stocks, bonds or mutual funds (a basket of more long-term investment products). Debtor households (households with significant outstanding debt) tend not to hold many, if any, investment products. In addition, investor households tend not to hold significant debt (beyond short-term credit card debt). Therefore, another reasonable statistic would be average mutual fund balance per investment holding household.

Continuing with the example, let’s say that the mutual fund marketer already has a collection of customers (investors of every stripe) in the target neighbourhood whose average mutual fund balance is $30,000 per household. The average mutual fund balance per investment holder in the neighbourhood is, say, $35,000. This means that either (a) the marketer’s clients have on average $5,000 in additional mutual fund holdings that can be targeted for conversion to the marketer’s firm, or (b) the marketer’s clients tend to have lower mutual fund balances than average investor households in the neighbourhood. If corollary evidence suggests that (a) is most likely true, the marketer can compare the findings of the average balance per mutual fund holding household and the findings of the average balance per investor household to determine whether the marketer’s customers tend to share only a portion of their mutual fund business with the marketer’s firm or whether the customers are likely using other firms for their mutual fund portfolio management.

Perhaps the best example of using a denominator that does not correspond to the balance numerator would be outstanding mortgage debt per real estate holding household.  Remember that WealthScapes does not make a distinction between mortgages held on primary real estate versus mortgages held on other real estate. In many cases, the marketer will already have a collection of clients who own real estate and a household needs to hold real estate in order to have a mortgage. Therefore, assuming a constant response rate per household and a fixed cost of marketing to each household, the marketer for mortgage-related products would want to target those areas with high mortgage balances per real estate owner.

Let’s build a new example of a marketer selling life insurance. For now, let’s assume the demographics of the population are homogenous (so there is no higher premium for older household maintainers, for instance) and further assume that the policy coverage is directly proportional to the outstanding mortgage debt, the premium is directly proportional to the policy coverage and the marginal profit is directly proportional to the premium. Because the marketer has the ability to directly target the homeowner subset of the population, the marketer would want to target outstanding debt via the homeowner and should target areas with high outstanding mortgage debt per real estate holding household. Obviously, if the marketer could identify the households that had mortgage debt beforehand (for example, when targeting a new “starter home” subdivision where mortgage propensities are quite high), it would be more appropriate to target areas with high outstanding mortgage debt per mortgage holding household.

The Right Denominator for the Right Job

The aforementioned scenarios are by no means exhaustive. Great care has been taken in WealthScapes to produce robust denominators for all the levels of hierarchy of wealth. For example, in WealthScapes we provide aggregate GIC and term deposit balances despite offering the same aggregate balance (indirectly) as the sum of those GIC and term deposits in registered accounts and in non-registered accounts. However, the incidence for GIC and term deposit holdership cannot be reconstituted from the incidence of the registered and non-registered subcomponents available separately. Some households will own GIC and term deposits both within registered and non-registered accounts and would be counted once in the registered incidence, once in the non-registered incidence and only once in the total GIC and term deposit incidence. If one were to sum up all the incidences of holdership of the “end node” products (demand deposits, GIC and term deposits in registered accounts, GIC and term deposits in non-registered accounts, credit card debt, primary real estate, etc.) the sum would be well in excess of the total number of households in the neighbourhood due to the double counting. As a result, should the situation arise where one is interested in households that hold any stock or mutual fund, it is possible to derive the total aggregate balance of stock and mutual funds using the balances provided.

Figure 1: The primary variable hierarchy in WealthScapes.

Figure 2: The alternative variable hierarchy in WealthScapes.

Once again, there are numerous ways to interpret any data and there are numerous scenarios, each requiring a different degree of finesse. Should you have any questions about which denominator is appropriate in your analysis, contact your Environics Analytics client advocate as we would be happy to be of assistance.

# # #

Peter Miron is a Senior Research Associate, specializing in data mining, customer profiling, demographic forecasting, financial modelling, behavioural analysis and target marketing. The principal analyst for developing WealthScapes, he possesses a multidisciplinary background in quantitative economics, database development, spatial modelling and advanced statistical analysis. Peter also has broad industry experience, having conducted analytics in the automotive, banking, insurance, retail and travel sectors.

EA Products: HouseholdSpend, Demographic Estimates and Projections, Mapping, Delivery Area Analysis

Client: Swiss Chalet is one of five restaurant chains owned by Cara, the largest operator of full-service restaurants in Canada and the leading caterer to the travel industry. With brands like Harvey’s, Milestones, Montana’s, Kelsey’s and Swiss Chalet, Cara’s restaurants and establishments span the country with a presence in every province.

Challenge: To evaluate the potential for Swiss Chalet delivery and takeout for any market in the country, the chain needed a quick way to estimate food expenditures away from home, which would help determine how many new restaurants could be added to a market. The company also needed a way to communicate the delivery areas to the Call Centre staff so that customer orders could be efficiently routed and guaranteed delivery times could be met.

Solution: Using HouseholdSpend data, EA analysts provided a template for mapping existing and proposed delivery areas around restaurants that showed the key demographics and current Swiss Chalet delivery sales versus total expenditures for food away from home. This comparison also provided the current market share of restaurant expenditures and, with Demographic Estimates and Projections data, allowed Swiss Chalet to determine which areas were growing fast enough to either expand current kitchen capacity or to add another restaurant within the local market.

Results: Swiss Chalet used data-rich maps to modify the boundaries of its delivery areas to improve delivery times to its current customers and to add new restaurants in locations that were being under-served. EA also built custom applications allowing the Swiss Chalet Call Centre to identify which restaurant should deliver take-out food to new customers based on the redefined service area of each restaurant.  The work proved critical to Swiss Chalet’s nationwide rollout of its delivery services. In the words of Arnando Viegas, Director of the Cara Call Centre, “Cara’s requirement for effective and innovative analytical products and services from EA is critical to supporting our plans for aggressive and intelligent growth across Canada.