Winning Agency of the Year is getting to be a habit for Touché! – this is the media agency’s fifth consecutive medal. When trying to unravel whether the shop’s secret is its people, processes or data science, our conclusion is that it’s all of the above.

To maintain its momentum, CEO Karine Courtemanche says the agency has established an audience planning process that is “obsessed” with finding and reading consumer signals gathered from sources like social. But Courtemanche says social listening isn’t always enough; sometimes you need to trigger the conversation.

Take Can-Am’s “Women of On-Road” campaign, for example. The goal was to get more women – who make up a very small portion of the client’s audience – to try the company’s three-wheel vehicles. One of the problems, Courtemanche explains is that the companies set up social media pages to target people who have shown they’re likely to purchase from the brand. In Cam-Am’s case, it was men, which created a bias against women. To change that, the agency got creative with data collection.

Touché created a private Facebook group specifically for women to talk about some of the perceived stigmas when it comes to riding. Members of the group revealed that they felt they were being cautioned against the dangers of motorcycling more than their male peers.

Another topic was the misconception that women did not have the physical strength to ride a motorcycle.

The insights led Can-Am to partner with Rolling Stone’s “Women Shaping the Future” event, which showcased the problems women face entering the category, as well as creating a mentorship program for women in partnership with riding schools. The campaign reached three million future women riders, sales to women increased 15%, and women now represent 36% of Can-Am owners versus 19% among other manufacturers.

Courtemanche says that consumer signals are where science meets art. The art is finding the right insight via online reviews, products left in digital shopping carts, or search keywords. These signals are not always obvious and may require building strong data partnerships.