August 23, 2017
Consider this: According to Pew Research Center, 79 percent of Americans now shop online--51 percent of whom do so on a mobile phone.
Once unconventional forms of shopping suddenly aren't so unconventional at all.
That in mind, weather should theoretically have more influence than ever on our impulse to drain a bank account for some momentary bliss. Is the checkout cart on Amazon really the best way to cure the rainy-day blues? Curious, we reached out to Xueming Luo, Chair Professor of Marketing at Temple University's Fox School of Business, for some insight. Luo recently partnered with a large company to conduct an experimental field study that observed purchasing habits of mobile customers during varying weather patterns.
Do we spend more money when weather is gloomy? Rainy, cloudy, whatever the case may be.
So, the answer is during a sunny sky--compared with a cloudy sky--people will spend more. With the rainy sky, people spend less. And this is significant because we think people, during a sunny day, they'll be in a better mood and when they're in a better mood it triggers all kinds of purchasing decisions.
What's really interesting is how [our test subject] communicated with their customers in different weather. On a sunny day, it matches up with a positive kind of messaging--one that's promotion-focused. It encourages the customer to think of buying now. On rainy days, people are more negative in their mood, so they match with a kind of negativity in tone and message. That's the more surprising part: You have to consider the type of communications strategy under different weather.
How did you go about testing this?
So basically what we did in research is we randomized [six] million people, mobile users. We know each user's local weather. And then we decide the communication message and content. We used the data to find [our results].
What were some deeper findings of the study?
Purchase responses to promotions are higher and faster in sunny weather, but lower and slower in rainy weather, relative to cloudy weather. In terms of odds ratios, sunshine leads to about 1.21 times more response to mobile promotions, and rainfall leads to about 0.9 times less response than cloudy weather.
Survival models at the hourly level suggest the hazard rate of purchase responses in sunny weather is 73 percent faster than that in cloudy weather. By contrast, the hazard rate of purchasing in rainy weather is 59 percent slower than that in a cloudy sky.
Better-than-yesterday weather and better-than-forecast weather engender more purchase responses and vice versa. A good deviation from the expected rainy or cloudy weather with relatively rare events of sunny days will substantially boost purchase responses to mobile promotions.
The ad copy of mobile promotions interacts with sunshine and rainfall. Compared with a neutral frame ad copy, the prevention frame hurts the initial promotion boost by sunshine, but improves the initial promotion drop due to rainfall.
How can the knowledge from this be applied in the long-term?
[Our partner] has API with the local weather of each user on social media, like Twitter or Snapchat. So, say in Philly today it's sunny now; it can give you different display ads or personalized ads in-feed. And this is done by other platforms--even Google. They all have these local-weather-targeted messages. And even, for example, people in Boston during winter, it' s very snowy and cold so the trigger from the ad would [be different than] Florida where there's sunny weather.
I think, also, a lot of people will be able to do the content, the personalization for the content, using certain words that depend on weather. You can boost ROI depending on weather.
It's funny. I would expect people to spend more money when it's rainy to counteract the bummed-outness from the weather.
Your intuition is right. On average, rainy weather does worse; fewer people buy compared to sunny. However, this negative effect can be countered a little bit if you design your communications a bit differently. That's our achievement. For example, compared to cloudy days there's about a 30-percent-lower purchase probability. People tend to purchase less. However, if you use a prevention-focused method [that boost purchases during rain] ... then you can counter this negative effect a lot.
Some of this depends on geography and perception, no? In Florida or Texas, where it's often sunny all the time or hot, you may look at a cloud or rain and feel relief.
You're definitely right. What we accounted for were changes in weather, rather than day to day. As you're saying, if it's sunny all the time people expect it, right? So, what we find is that if it's really, really rainy and all of a sudden it's sunny, that could have a much stronger effect. Vice versa. Compared to sunny and all of a sudden you have rainy--it could be a certain temperature or humidity that controls for that effect and so you find there's an even stronger prediction. And also, if you deviate from your normal weather--it's expected to be cloudy but you end up with especially nice weather--the effect gets even stronger.
... [To control for correlation versus causality] what we did was we were careful to control for time of day, the day of the week, rainfall, air pressure, humidity, etc. If it's sunny but too hot, there won't be any effect. If it's 110 degrees in Texas or Arizona, no one likes sunny weather anymore. We also had purchase timing as a model which is more robust evidence ...
Anything to add?
I think where this study has some interesting material for advertisers and brands and managers is you just have to work with the weather. What's interesting for maybe company brands, they have TV, internet, mobile promotions regardless of weather, to promote their brand. But what our realization is, is if you can forecast weather in each location in the U.S., then 'OK we know the focus of weather' and use that information to create a different kind of messaging to their customer ...
"Don't miss this opportunity in the advertisement." That's our recommendation. Which is easy to do--you don't invest much. It's kind of framing, a change in content. Personalize the message to the local weather and it may bring higher ROI and bang for your buck. Weather is everywhere.