It doesn’t take technology for good salespeople to know to put shovels front and center when a blizzard’s coming. Heck, even NYC street vendors know to sell sunglasses in summer.
AOL believes weather factors into sales on a much deeper level — and it claims it has the technology that can deliver these insights to marketers and advertisers.
New York City-based AOL yesterday announced the addition of local weather as a variable to its attribution platform, Convertro, which the media giant acquired last year. Through the platform, marketers can tell how much an ad influenced a customer purchase decision, versus how much local weather might have.
In other words, was it the offer that led the customer to the purchase — or the fact that it was cloudy?
“By allowing marketers to understand the impact of weather at an individual consumer level, it enables them to more accurately measure their media strategy’s return on investment and cost-per-action,” said Amy Mitchell, head of Convertro at AOL Platforms. “We’ve updated the attribution model to consider local weather patterns on past purchase decisions.”
Hasn’t this technology been available before? Yes. But AOL officials said not at this granular level.
“Traditionally, this was done through media mix modeling, which only looks at how weather affects the impact of ads from a macro level,” Mitchell told CMSWire. “Ultimately, these solutions lack the user level understanding that our multi-touch attribution approach brings — e.g. people in NYC react to rainy days differently than those in Seattle.”
It’s all about, Mitchell said, giving “credit where credit is due.” The AOL platform drills down to the zip code and over multiple time periods and applies “a more data-driven approach.”
“This makes attribution,” Mitchell added, “more accurate and helps marketers understand the true impact of their media buys.”
How Does it Work?
Mitchell called the new tech a “data science project.” She said weather is now a control variable at the user level in AOL’s multi-touch attribution statistical model.
AOL officials source and crunch different types of datasets:
- Anonymized user-level geographical data
- Time attributes
- Daily and historical weather data (temperature, precipitation) from the National Oceanic and Atmospheric Administration (NOAA) for all US zip codes
“The depth provided by this approach to weather modeling is unmatched,” Mitchell said. “We consider weather type, time windows and measurement type: absolute versus relative. As a result, we can build 27 weather model variables for every person. This is an attribution industry first.”
There is no additional cost for weather attribution. It’s available now to existing Convertro users.
Mitchell said to expect more out of the new algorithm.
“The introduction of an ‘external’ factor like weather lays the foundation for AOL to introduce even more context-based variables to the algorithm,” she told CMSWire. “The local economy for a consumer and seasonality are just two of the possibilities. We can basically ingest any external variable that can be mapped back to a user’s location and time and evaluate its impact via our algorithm. This will make attribution clearer, smarter and more effective for marketers in understanding how their efforts influence's the customer journey."