Case Study: Fido

CLIENT:

Fido is a wireless phone service provider owned by Rogers Wireless.

CHALLENGE:

For some time Fido, which features an assortment of dogs in its marketing materials, has sold mobile phones through three main channels: retail stores made up of dealers, corporate stores and resellers; the Internet; and direct-to-consumer phone orders through FidoDirect. The company wanted to steer more customers to FidoDirect but was unsure which advertising channel to use. Fido sought EA’s help in developing a targeted marketing campaign and answering two important questions: Can unaddressed mail—an inexpensive alternative to address-specific mailings—work better than other media, such as freestanding inserts or billboards? And what is the best approach to selecting postal walks with high concentrations of potential cell phone subscribers?

SOLUTION:

To better understand Fido’s subscribers, EA first employed its PRIZM segmentation system to determine which lifestyle types have a higher propensity to be FidoDirect subscribers and to respond to previous direct mail campaigns. The analysis showed that FidoDirect’s best subscribers tended to live in areas that were urban, ethnic and working-class. With its own predictive model, FidoDirect scored all of Canada’s six-digit postal codes within the Fido footprint based on the demographics and purchasing patterns of its current subscribers. Then, using FidoDirect’s best performing postal codes from top deciles, EA scored letter carrier routes (each consisting of 250 to 400 households) within the company’s footprint to determine where to deliver FidoDirect’s unaddressed mail promotions.

RESULTS:

The response rate for FidoDirect’s targeted group handily beat the company’s control group, but EA believed it could improve the results by incorporating segmentation to the selection methodology. Focusing on Fido’s top-scoring postal codes, EA used a proprietary technique to weight and score the letter carrier routes that also contained above-average levels of Fido-friendly PRIZM segments. Then it dropped unaddressed mailers to this new list of high-scoring carrier routes derived by combining Fido’s internal model with the PRIZM analysis. The result: the combined targeting model yielded a 41 percent lift over Fido’s internal model alone. And by further tracking response rates, EA and FidoDirect learned that the combined targeting resulted in an increase in calls from potential clients about a week after the unaddressed mail drop. As a result, FidoDirect began to stagger its mailings to avoid overwhelming its call centres. The company also found that its new direct channel customers tended to download data—like ringtones and pictures—at higher rates than other customers, providing insight to support new offers.

“It’s been shown in data mining that multiple models are more accurate and robust than a single one,” says Krzysztof Dzieciolowski, Business Manager for Database Marketing at Fido. “In our acquisition campaigns for high-potential prospects, we successfully combined our predictive models with highly informative PRIZM clusters to create a more innovative solution.”