Predictive Analysis in Unusual Places

You never know when an analytics opportunity may appear. Recently, when I stopped for gas in rainy, grey Seattle, I saw a sign near the pump that read:

Cold Day? Try Our Hot Coffee At the Temperature!
A 12 oz. cup will cost you the same price as the temperature.
If it’s 40 degrees, the coffee is 40 cents! (Plus tax)
Temperature based on our thermometer. Ask our cashiers for details.

The sign got me thinking. Was it a good ad or a bad one? It certainly caught my attention. But did it compel me to buy the coffee?

The creative and copy weren’t very professional. The offer also seemed to flunk Economics 101. I mean, don’t people usually want coffee more in the winter, when it’s cold and dark outside? Yet, the offer gives you the lowest price for hot coffee when you want it the most. Supply and demand. Hmm…

I decided to score the ad on a simple, subjective scorecard:

  • Ad Goal
    Generate awareness, sales, and profit.
  • Score
    + For awareness. ++ It passed the social media test, as I’m now writing about it.
  • Brand Sentiment
    Positive sentiment, as I appreciated the humor.
  • Business Result
    Selling coffee at a smaller profit margin when people want it most — and are willing to pay the most — feels a little backwards.

The most important question may be, did the ad “drive” more traffic to the gas station and make them buy gas? In a city like Seattle, where buying coffee is an option everywhere you look, and the temperature rarely gets too hot or cold, I seriously doubt anyone will care if they save five cents if they “buy now.” (Were the ad run during wintertime in Chicago or Nome, AK, maybe the result would be different.)

I shared my coffee story with my friend Jason Burby while we were on a recent flight. He one-upped me with his story of a recent trip to the liquor store. As he walked the aisles looking for his favorite gin, he saw this sign:

Automatic Price Reduction — $3.79
Ninkasi 22 oz. Bottle
Reg. $3.79. Save 0%

The type was huge and bold, highlighted in yellow. He’d never heard of Ninkasi, but soon realized it was a micro-beer. Only after that did he realize the offer was for a whopping 0 percent discount.

  • Ad Goal
    Generate awareness and sales
  • Score
    + It worked. ++ It passed the social media test, as he told me and now I’m writing about it.
  • Brand Sentiment
    Positive sentiment, as we both appreciated the humor.
  • Business Result
    The product received better awareness than sitting on the shelf in parity with the rest of the 100 other similarly creatively named micro-brews. Ultimate result: beer bought.

Though both ads were designed for simple, everyday use, they also have broad applications. Examining them from analytics and predictive modeling standpoints offers insight into how to market your most important products.

Setting ad goals is very similar to predicting the future. You need to think about what you want to achieve and make sure the incentive ties to the business outcome you desire. Next, understand what the offer will cost. Understand the upside. Do the math. Be sure to include a contingency for unknown factors. Then analyze the performance and response rate to your ad and offer, and be sure you can clearly attribute the performance and results.

Lastly, don’t forget to optimize. Fine-tune the offer and ad to get the maximum performance. You may not be able to sell your product at a zero percent discount, but if you’re smart and clever enough, you may get close.

Related reading