Data Analytics FlockTAG's Poisson Regression Delivers Better Pizza Deals to Hungry College Students
Today's deal-hungry college students are constantly bombarded by one-size-fits-all digital advertisements and deal coupons delivered randomly through multiple communications channels. While these tech-savvy consumers prefer interacting via their mobile devices, they are sensitive to the type and frequency of commercial messaging they receive. When the volume or content of spam ads becomes too annoying or irrelevant, many mid-Millennials simply choose to opt out or disengage altogether. Businesses, in turn, miss out on increased customer satisfaction, engagement and revenue.
At FlockTAG, we have developed a sophisticated marketing model, based on data analytics, that enables us to cut through the background noise and digitally deliver targeted, high-value deals to students while driving more traffic to our clients' businesses. We know it works because we've seen the results.
Using our method, we have achieved redemption rates for digital coupons that range from 15% to 20%, and up to 45%. By industry standards, that is practically off the charts. Our levels are significantly above the average redemption rates for untargeted, low-value deal coupons, which languish in the 1% to 2% range. This means that students get the smart, personalized deals they value and are most likely to use at the campus restaurants and shops they frequent. Vendors, in turn, see an increase in the number of students who come into their places of businesses regularly, as well as an uptick in the dollar amount of transactions per visit and throughout the school year. Everybody wins.
So how did we get here? It wasn't easy. Over the past two years, our FlockTAG team has tracked 334,000 college-age FlockTAG users who were sent 724,000 deals, and we have compiled a data pool based on 4.4 million transactions. Working with our data-advisory board, we have crunched these numbers using data-analytics modeling, notably linear regression and Poisson regression, and come up with a complex, but highly effective, statistical method for estimating what students want and how and when to deliver it. Through our data model, we are able to segment FlockTAG users into different customer categories based on their individual preferences and past buying behavior. Then we work closely with restaurateurs and merchants to create the optimal deals that match the needs and wants of customers in each category. Using GPS technology, we can time the delivery of these deals via sms text message, push notification or email to cardholders based on their geographic proximity to a vendor's place of business. Finally, we cycle through the deal list over time to avoid deal fatigue, which results from too much repetition.
Sounds complicated, right? In practice, however, this so-called machine thinking works quite well. For instance, we can text a deal coupon for a discount on a large size coffee to a student who visits a particular coffee bistro only occasionally, and notify her of the deal when she approaches the shop. This targeted approach is designed to encourage her to stop in for coffee and eventually become a more regular customer. In another case, we can email deal discounts on food items and frequent-dining reward coupons to a student who eats on weekends at a campus restaurant, in order to motivate him to dine more often during the week and increase his spending at each meal.