Client: CAA South Central Ontario
Agency: KnowledgeBase Marketing
Challenge: To raise awareness and increase usage of the CAA club’s seven travel agencies by CAA members
Despite their highly visible storefront locations all over North America, the travel agencies of the Canadian Automobile Association and the American Automobile Association don’t push their travel services to just anyone.
Sure, they’ll provide travel counselling to the odd non-member who walks in off the street, but they’re not really interested in pursuing that line of business. That’s because their mandate is simply to underwrite the non-profit lobbying and roadside assistance activities of their parent organizations, which operate to serve members only.
Having to work within an arbitrarily narrow customer base means the caa and aaa can’t amortize their costs over a broad customer base. It’s a situation that makes accurate targeting all the more important.
In the case of CAA South Central Ontario, which covers Hamilton, Ont., and the rest of the Golden Horseshoe area, the challenge was to try to pump up usage by the club’s 164,000 member households from a measly 12% to a slightly more respectable – and achievable – 15%.
Rather than just tap their database for members who had used the club’s travel agencies in the past and fire off a mailing, executives first handed a brief to the data mining specialists at KnowledgeBase Marketing of Toronto: Help us identify members ripe for an overseas travel offer.
‘We asked KnowledgeBase to take a look at people we defined as high-value,’ says Elaine Hignett, database marketing coordinator for CAA South Central Ontario. ‘They came back with a model that predicted which of our members not using the travel agency could be high-value travelers.’
The weapon of choice was predictive modeling, a kind of statistical crystal ball that identifies which of a company’s customers are engaged in certain kinds of buying behavior, then finds other customers that ‘look’ and might therefore act – given the proper incentives – like them.
Starting with South Central’s member list, KnowledgeBase began to build a ‘propensity model’ by pulling a group of customers from the database who had purchased travel services in the previous 12 months and comparing them to a second group that had not.
Having identified basics like age-range and address, the next step was to run each member file through overlay data from Statistics Canada and Revenue Canada, taking into consideration average employment income, average social assistance income, average rrsp contribution, and demographic variables like age and marital status.
Factoring nearly 300 variables into the model, KnowledgeBase established a score (called a decile) from one to 10 that predicted each member’s likelihood of accepting a caa travel offer. Decile one meant a low propensity, decile 10 a certified deer-in-the-headlights.
So what does a high-propensity customer look like? In caa’s case, a Decile 10 prospect would probably remind you of a well-to-do, if not extravagantly rich, uncle.
A caa member for about 283 months (23.58 years), he uses non-travel services like emergency roadside assistance and has an associate member (probably Mrs. Decile 10) attached to his membership. He’s an investor, lives within 3.5 km of a caa travel agency and likes to play it safe, purchasing extra personal accident insurance when he travels.
Not surprisingly, members in the lower deciles tended not to share these characteristics, or shared them to a much lesser extent.
Having established this model, caa and KnowledgeBase settled on a ‘pretty run-of-the-mill’ offer, says Hignett, travel agency standards like package tours and escorted bus excursions to Italy, Greece and Spain, but price-pointed at about $2,700 per person to generate big commissions. A $100-off coupon and a free upgrade for a car rental rounded out the deal.
‘The creative was very nicely done, but the offer wasn’t anything you couldn’t negotiate yourself,’ she adds. ‘Any customer could get $100 off. You just ask for it and you get it.’
A drop of 25,000 was established favoring members in the upper deciles, though a significant number were mailed to a ‘control cell’ of members in the lowest deciles to measure how well the model actually predicted a propensity to purchase.
The breakdown was 7,000 pieces to decile 10, 7,000 to decile nine, 6,800 to decile eight, and 600 pieces each to a deciles one through seven.
The overall response rate was nearly 7%, with 1,694 trips booked through CAA South Central’s seven travel agencies.
‘We looked at the data and let it dictate what kind of customers buy travel,’ says Lori D’Alessandro, manager of strategic communications for KnowledgeBase.
‘Let’s say you’re a bank marketer pursuing credit card acquisitions – you want more people to sign up for a Gold Card. So you look into your database of six million customers and decide, ‘ok, people likely to buy a Gold Card probably have a line of credit or a mortgage with us.’ You think of things in your head that might reasonably be assumed to be criteria for people who own a credit card. Then you go to your ims people and say ‘Here, pull me a group of these people.’
‘The difference between that and what we did is essentially we let the data tell us what that customer looks like instead of using what we call judgmental criteria to select people.’
Sidebar: Sunoco taps into CAA
Even without a predictive model to boost impact, marketers in other auto-related organizations say caa’s member database is a robust targeting resource.
Gas retailer Sunoco recently tapped the auto club’s Ontario database for a neighborhood marketing initiative targeting caa members living within three kilometres of Sunoco gas stations.
‘We did a geo select and identified caa members who were basically our next-door neighbors but weren’t coming in to us,’ says Haydn Northey, marketing manager for Sunoco. ‘We sent out a letter and some incentives like a car wash or gas coupon and got a phenomenal response. We had 50% of those people come in and 43% of them have now come in three times or more.
Trading on an existing partnership that allows caa members to earn 2% of the value of any purchase made at a Sunoco station and apply it toward next year’s membership fee or other caa products or services, Northey says Sunoco is counting on access to caa’s 1.7 million Ontario members as a way to boost fuel sales through the millennium.
‘Obviously our core customer base comes from the immediate areas around our stores, so we’re really going to focus on neighborhood marketing. That’s going to be our push over the next two years.’ MD
