Researchers under fire

Technologically-driven changes - like data mining and online surveying - are profoundly changing Canada's $650-million market research industry, closing some doors and opening others.

Technologically-driven changes – like data mining and online surveying – are profoundly changing Canada’s $650-million market research industry, closing some doors and opening others.

On the one hand, technological advances have allowed traditional clients to take much of their data gathering in-house, mainly through loyalty programs that track their customers’ every purchase. This creates a very real threat to the livelihood of traditional data-gatherers. On the other hand, the technology has upped the flow of raw data, creating a huge demand for talented market researchers who can analyze that data and figure out what it all means.

Since the mid 1990s, computers and IT developments have nurtured the proliferation of data mining, the act of collecting and analyzing customer transactional information in company databases, largely procured through loyalty programs.

Because these are customer databases, data mining is largely an in-house function, and there’s the rub: traditional data gatherers could find themselves losing business as demand for their services declines.

Not surprisingly, this has left the Professional Marketing Research Society, which largely represents traditional market research companies, caught between a rock and a hard place. The society has a natural urge to protect its traditional members from losing business, but it also spies an opportunity to bolster its membership by welcoming data miners into the fold.

This trend is so worrying to the field that the PMRS struck a Futures Committee in July 2000 to look at the impact of such issues as data mining, Internet surveys and globalization on the industry and its membership.

Last April, the Futures Committee issued its first report, concluding that data mining ‘is increasingly replacing more traditional forms of research and also presents risks in terms of standards and best practices.’ The committee further concluded that one of the potential consequences of this trend are ‘lost revenues for custom research companies.’

But as much as Futures Committee member Bob Collins agrees there is some in-house activity ‘that would otherwise generate revenue for market research companies,’ he is not concerned.

‘It is not an either/or proposition,’ he stresses. ‘Each tool, [custom research and data mining] has its own strengths,’ and the two data gathering techniques can be used to complement one another. But then, as director of the customer insights group for LCBO/LLBO, Collins is speaking from the client side of the equation.

One of the strengths of data mining, says Collins, is its efficiency and cost effectiveness, particularly in relation to low incidence brands. Such brands can be expensive and profitable, but only purchased by one out of a hundred customers, and before data mining, gathering information on this elusive segment was costly.

Janet Hawkins, SVP, customer strategy and information, TD Bank Financial Group and TD Canada Trust, is also sold on the merits of data mining. ‘There’s only so far you can go on the marketing research side,’ she says. Once she receives all the information and charts, she says she sometimes finds herself saying: ‘So what?’

On the other hand, with data mining, she gets ‘very quantifiable’ data on a specific segment of the bank’s customer base that is ‘actionable.’

Rob Shields, VP customer relationship management and loyalty at The Bay, is another client with greater expectations: Gone are the days of passive observation, he says. ‘Market research needs to be active.’

For The Bay, which launched its new HBC Rewards loyalty program this year, data mining plays a key role in customer satisfaction. ‘It gets the customer to craft what the value proposition should be.’

But while the payback from data mining is significant, both Shields and Hawkins continue to do traditional market research, because data mining also has its limitations. As Hawkins notes, ‘It can’t tell you why,’ and ‘It can’t tell you about the competition.’

Marketers still need to understand the motivation behind purchases, why a particular product is not selling, or why customers chose this particular store, in order to customize their sales strategies. They also need to have a sense of the total customer universe, not just their particular segment, in order to devise strategies to increase their share of a customer’s wallet.

By drilling down through the database, Hawkins says she might notice a trend out of the norm or identify a big issue with a particular segment. That’s when she draws a sample of those customers and commissions an outside custom research study to find out why it is happening. Because of this, Hawkins says she probably still commissions as much custom research as she always has.

Data mining, and the loyalty programs that supply the data, are all about retaining an existing customer base, while market research can provide data on potential customers. This gets to the heart of the strengths and weaknesses of both tools. And both related marketing goals – retaining loyalty and growing market share – are crucial.

For example, with just a 5% increase in its customer retention rate, a company can reap a 25% to 125% increase in overall profit from those retained customers.

But focusing all efforts on loyalty at the expense of growth can be dangerous. ‘It’s like making the decision that one day I will be able to say I know who my last customer will be,’ says Stephen Popiel, senior VP research and development at Goldfarb Consultants of Toronto.

Besides, many traditional marketing firms have already developed an ‘if you can’t beat ‘em, join ‘em’ attitude. Several have already decided that as well as offering traditional market research services, they’ll also work with client databases, especially for those companies without data mining and statistical experts on staff. Popiel says Goldfarb specialists have already been ‘seconded to client IT departments to come up with solutions.’

Meanwhile, PMRS president Ivor Thompson maintains that he doesn’t see data mining as a ‘threat,’ but an opportunity. His main concern is that all the specialists jumping on the market research bandwagon are guided by the same standards and practices as traditional market research firms, for the good of the industry.

‘It gets tricky linking back survey information to purchasing patterns,’ agrees Tim Wingrove, senior partner at Toronto-based market research company Commins, Wingrove and Futures Committee member. ‘The information should only be used in aggregate. You shouldn’t go back to an individual respondent [armed with their personal information] to further annoy them.’

Collins, who is also the public relations chair for PMRS, notes that his company has a special system in place to protect privacy. The LCBO is a sponsor of the Toronto-based Loyalty Group’s Air Miles program.

While names and addresses are collected by Loyalty Group, individual Air Miles sponsoring companies only know their customer by a special collector number. Thus, companies like the LCBO can only ferret out aggregate trends, like the fact that people who buy scotch are also likely to buy imported beer. So for customer convenience (the LCBO is not allowed to promote the sale of alcoholic beverages, but data mine for customer service only), the LCBO has recently erected a new in-store display which offers both scotch and imported beers.

Adding data miners to the PMRS roster would mean they would be governed by the organization’s code of ethics, which would keep such firewalls in place, says Thompson. Welcoming the new members would also have the added bonus of cross-pollinating traditional researchers and new specialists to educate and inform each other’s disciplines.

The PMRS will decide whether to open its doors to new members at its annual general meeting on Nov. 5.