The Weather Network’s next forecast: your shopping habits

The Weather Network built its reputation on predicting blustery days. Now it aims to predict something much more profitable

 
Pierre Morrissette, CEO of Weather Network parent Pelmorex Media
Pierre Morrissette, CEO of Weather Network parent Pelmorex Media. (Portrait by Hamin Lee)

The forecast is calling for rain, and Pierre Morrissette forgot his umbrella. “I must admit, I didn’t totally focus on what’s happening today,” he says. This admission should be cause for mild embarrassment: Morrissette is a man who has devoted his life to weather. He is the founder and CEO of Pelmorex Media, which owns the Weather Network. This grey November day is going to bring not only rain, according to the company’s forecast, but also severe wind gusts of up to 90 kilometres an hour.

But if there is any consolation, it’s that today is also Storm Watch day at the Weather Network. Regular television programming is interrupted for periodic live updates on the storm—or, in company parlance, “active weather.” Storm Watch days are good for business. “I hate to say it, but the more active weather we get, the more it impacts people connecting to all of our platforms,” Morrissette says from his office in Oakville, Ont. “Numbers go up, and revenues follow.”

Morrissette says this with a practiced air. He has, after all, been in this business for more than two decades. He took over the fledgling network in 1993 and remains its controlling shareholder. The Weather Network commands 80% of the Canadian market but, in some ways, it feels outdated. The thought of tuning in to a television channel to check the forecast seems downright cumbersome when you can easily get the weather on any smartphone—many of which come preloaded with a rival weather app. The embedded iOS software pulls data from the Weather Channel in the U.S. Yet plenty of Canadians take the extra step of downloading the Weather Network’s app. In 2013, it was among Apple Canada’s 25 most popular apps. As of December, it had been downloaded more than 7.5 million times.

The challenge the company faces is similar to the one any large, entrenched incumbent encounters: How does it keep growing? One answer is to look outside Canada. In the next few years, the company aims to become a sizable international player, and it has expansion plans for major European markets (including Germany, France and Italy) and even the U.S. The company is also figuring out novel ways to boost advertising revenue. It has undertaken an ambitious project to target ads on its digital properties based on weather conditions. That means more than predicting shovel sales will spike when the mercury dips. For the Weather Network, it means hunting for counter­intuitive ways that humidity, pollen count, UV index and more affect consumer behaviour. The company claims it can optimize ads to ensure they appear when the weather conditions are just right, and charge a premium to do so. “We believe the weather is the single biggest factor that influences user experience that you’re not thinking about,” says Tony Patel, Pelmorex’s vice-president of monetization and customer analytics.

Ultimately, the company is trying to forecast a lot more than the weather. It’s trying to forecast how we behave.

This new advertising initiative is a far cry from where the Weather Network began in 1988. The creation of Montreal engineering firm Lavalin, Weather Now (as it was called) and French-language offshoot MétéoMédia were part of a diversification strategy. The strategy was, of course, met with skepticism, and Lavalin found itself under pressure to divest the assets during a rough patch.

Morrissette, meanwhile, had just founded Pelmorex Media. A Montreal native with an MBA from the Ivey School of Business, he owned a few small-town Ontario radio stations through Pelmorex and was eager to expand. He was impressed by the system Lavalin had developed, which allowed local forecasts to be dropped into a national feed. Lavalin lacked an advertising business, though, and depended on subscriber revenue. Morrissette thought he could turn it into a proper media company and convinced his board to purchase it in 1993.

Today, Pelmorex generates around $100 million in revenue. The company has a history of making wily business decisions. In 2010, for example, it built a nationwide emergency alert system that can issue warnings (weather-related or otherwise) to televisions and radios when requested by police and government. The rights to operate it were granted to Pelmorex by the Canadian Radio-television and Telecommunications Commission. In return, Pelmorex allocated must-carry status on basic cable, a highly sought-after distinction because it guarantees revenue.

Through it all, Morrissette has been at the helm. Well, almost. He actually stepped down in 2013 and appointed Ron Close, a former Bell Media executive, as his successor. Close resigned just over a year later for personal reasons. Morrissette, now 67, resumed his old role.

Until fairly recently, the company had taken only small steps toward directing digital advertising to weather conditions. “If an advertiser wanted to show a bunch of winter tire ads when there was more than five centimetres of snow on the ground, we could do that,” says Patel. “But the main issue was you had to know up front whether it was going to work.” In other words, would that ad be any more effective with five centimetres of snow than it would with zero?

Patel arrived at Pelmorex from Yahoo Canada in 2013 and undertook a six-month project to overhaul the Weather Network’s digital ad business. The company switched to a programmatic model, which, for simplicity’s sake, can be thought of as automated, data-driven ad placement. If you’re checking the weather in Las Vegas from your home in Calgary, you’re probably taking a trip there soon, and the Weather Network will serve up Vegas-related ads. That’s standard, though. The company promises it can do much cooler things with its data.

Pullquote: With the relative humidity between 40 and 60, a car ad was twice as effective.

When you visit the Weather Network’s website, all kinds of information—including your location and the temperature, wind speed, humidity and pollen count—is sent to its server to determine which ad to show you. The company can also track the exact conditions when a user clicks on an ad. Pelmorex hired a European firm to transform that data into graphs showing the number of ad impressions for each weather variable, allowing the company to see which factors might be driving behaviour. With this level of detail, the company might tell a coffee purveyor the exact temperature range in which an online ad is most effective.

The process has uncovered some surprising findings. An ad for an automaker yielded two to three times more click-throughs when the relative humidity was between 40 and 60. Patel and his team tested the finding rigorously, and it held up.

Patel says he won’t always be able to explain the connection between the weather and an ad’s performance, which could be a roadblock for some clients. “This is new, so a lot of the media buyers were sort of confused,” he says. “What we’re trying to do is build a relationship to teach and learn together.” Retailers have long tinkered with environmental conditions to boost sales, examining lighting conditions, music and scents to see how consumers behave. Should weather be any different? “As mammals, we have a sort of sixth sense, some sort of connection to nature, and the weather influences how we feel and how we think,” Patel says. “That’s the core of what we’re trying to get at.”

It might be a tough sell. Mark Sherman, founder and CEO of media buying firm Media Experts, wonders how useful the data really is. “The Weather Network has no real insight into how their ads are performing because click-through rates are largely unimportant,” he says. “What’s important is whether people buy stuff.” Patel counters that some companies may simply be looking to boost brand awareness through a campaign.

It’s possible, however, to better tie the weather to purchasing behaviour. The Weather Channel in the U.S. is already doing it. In 2013, it formed WeatherFX, essentially an in-house ad agency that unearths these connections for clients. The company initially went to some of its larger partners, including Walmart, and asked for sales data for every product in every store across the country. The process was “painful,” according to WeatherFX leader Vikram Somaya, since retailers are naturally protective of their data. But with enough wheedling, the team got its way. “We found that weather is an accurate and fine-grained predictor of how people will behave,” says Somaya. In some cases, it’s more useful than the usual demographic characteristics marketers apply. “A 54-year-old Hispanic mother of three and a 26-year-old white college graduate might do exactly the same thing in response to a particular set of weather variables,” he says.

As Somaya and his team dug in, they discovered connections that made no sense on the surface. In Dallas, for example, sales of insect repellent spiked at a certain dew points. “We wondered, Why the hell would that be?” he recalls. “We actually figured it out. It was because insects were hatching at these dew points.” Eventually, they stopped worrying about explanations. “Our data scientists were telling us, ‘It’s not relevant why it’s happening. What’s relevant is every time these conditions happen, people start buying these products.’” As such, it’s something of a mystery why after several days of rain in Seattle followed by four hours of sunshine, people indulge in fruit cups.

Pullquote: In Dallas, for example, ales of bug spray spiked at certain dew points

WeatherFX has this kind of data for a wide range of consumer goods. Lotion sales, for instance, are influenced by cloud coverage, wind speed and temperature. In Phoenix in the fall, above-average temperatures will lead to a bump in sales. In the spring, though, abnormally high wind speed does the trick. The data is also useful in telling marketers when not to excessively promote. During the summer in Atlanta, for instance, people don’t buy any more beer than usual, no matter what happens with the temperature.

The work has led to some interesting campaigns. The Weather Channel partnered with Pantene to promote hair products based on the humidity and with Michaels to advertise craft supplies hinged on precipitation. Tying the data to actual sales is a better approach than looking at ad performance. “When we can prove it’s impacting sales, that’s the holy grail,” Somaya says.

Patel says Pelmorex is exploring something similar, but advertisers are a little uneasy about sharing data. Still, he’s optimistic the company can make headway in the meantime by optimizing ads for click-throughs. “We do see meaningful uplift and really good results that are well above average campaign expectations.”

While Patel and his team are diving deep into the company’s data to forecast consumer behaviour, Pierre Morrissette is looking for growth elsewhere: new markets. A few years ago, he realized that the company had to expand outside Canada, since it already had a dominant share of the market. Expanding also entails forecasting of a different element: which countries will be as fond of the weather as Canada is. A study from Montreal media firm Influence Communication found that in 2014, Canadian media outlets devoted 229% more coverage to the weather than the 160 other countries in the study. Part of the reason is likely that Canada is a big country with diverse weather patterns that change often, so we need to know how to prepare on a daily basis.

Pelmorex zeroed in on Europe, North America and Latin America, partly because of the high rates of digital media consumption. The first opportunity arose in Spain. ElTiempo.es, the country’s leading weather-related website, happened to be for sale. It was only a few years old and, in some ways, mirrored the state of Lavalin’s Weather Now when Morrissette bought it: The basic weather-reporting infrastructure was in place, but the company wasn’t generating much advertising revenue. Pelmorex hired Susana Rodriguez Urgel, a media and marketing executive who worked at Spain’s biggest telecom provider, to become the managing director of ElTiempo.es in Madrid. Urgel hired 16 local sales, marketing and technical staff, and oversaw a website redesign to improve the user experience. Urgel also recruited José Antonio Maldonado, a longtime radio and TV weatherman, to serve as ElTiempo’s meteorological director. Having started his on-screen career in 1986, Maldonado is a familiar face to Spaniards.

Since Urgel came aboard, the number of average monthly unique users in Spain has increased from just under three million to around 10 million. (She points out that this number is spread across the web, mobile and apps, so users could be counted more than once.) But Spain is not Canada; the country doesn’t have a weather culture. “We are the sunny country here in Europe, and when it’s sunny, people are not worrying about the weather,” Urgel says. The one major exception is rain. “When it’s going to rain, we want to know, because we never have an umbrella at hand,” she says.

In other countries, Pelmorex must adapt to the existing weather culture. In 2013, it established a presence in the U.K., where there is a healthy appetite for weather info, particularly anything related to precipitation. As a result, the company emphasizes its up-to-the-minute start and stop times for rainfall. The BBC provides forecasts, too, but “they’re not weather specialists,” says Taylor Emerson, senior vice-president of international development for Pelmorex in London. “They’re doing everything for everyone all the time.”

Most ambitious are Morrissette’s plans for the United States, where the Weather Channel, based in Virginia and founded in 1982, controls 50% of the U.S. market. AccuWeather is a distant second, with roughly 14%, according to research firm IBISWorld. The Weather Network already has a U.S.-centric website and mobile apps stocked with local forecasts, but Morrissette says he wants to learn from Pelmorex’s experiences in other countries, and to see a certain level of user engagement before investing resources in local staff. He’s also cagey about how the Weather Network might differentiate itself from its U.S. rival. “I’m not sure I want to give you a recipe,” he says.

The company may not have to differentiate itself at all. The Weather Channel holds about 10% of the Canadian market. “That’s without doing much marketing or anything here,” Morrissette says. Part of the reason could be brand confusion or because the Weather Channel owns a more intuitive domain name with weather.com. It’s possible the Weather Network could benefit from the same kind of brand confusion (or apathy) south of the border, though Morrissette says its content has to set it apart. “Really, it comes down to quality, accurate data that covers many, many, many different locations,” he says, pointing out that they can forecast within a one-kilometre radius of a user’s location, a smaller surface area—and therefore more useful—than the one-mile radius standard in the U.S. The U.S. is also a huge market, and a 10% share would double the size of the Weather Network’s digital audience. “I’m very confident we can pick up a meaningful market share in the U.S.,” he says.

It’s another forecast, in a way, and like each one the company issues, it puts its credibility at stake. Morrissette doesn’t appear worried, though. He adds with a sly grin, “Probably the Weather Channel won’t even notice.”

MORE ABOUT BIG DATA:

 

Comments are closed.