How strong is the halo effect when it comes to selling candy and gum? This was among the many questions students set out to answer during the third annual Wharton Analytics Accelerator Challenge, presented by Wharton Customer Analytics (WCA).
Over the course of six weeks, five Wharton and Penn student teams collaborated with corporate executives to solve data and analytics problems. The students cleaned and analyzed real datasets to recommend workable solutions to their business partners. The challenge culminated at a summit on Friday, November 8 when the students presented their final recommendations to company leaders.
Out of the dozens that requested to partner with WCA, four companies of different industries and sizes were chosen. The Ferrero Group, one of the leading chocolate and confectionery companies in the world, presented their student team with an interesting challenge.
“Our team’s challenge was to look at Ferrero’s TV product advertising across all markets and see if there were better ways for them to spend their advertising budget, and maybe reduce their spending on certain products in certain markets,” said Nikki Bowser, W’21, one of the team’s MBA engagement leads.
Ferrero makes a variety of beloved treats, namely Nestle chocolates, Nutella, Ferrero Rocher, and Kellogg cookie brands like Keebler and Famous Amos cookies. But Ferrero executives tasked Bowser’s team to focus on Tic Tac and Kinder chocolate products, which are sold throughout the US, UK, and Europe.
“We wanted to see if advertising for one product would benefit other products in the same portfolio, looking to see if there was a halo effect,” Bowser said.
The team found mixed results depending on the product.
For example, Bowser said the team used marketing models to see if the traditional Tic Tac mints saw a boost in sales when Ferrero advertised one of their newer products, Tic Tac Gum.
“The data shows that when the gum was advertised, the mints benefitted,” Bowser said.
Kinder was another story.
The team discovered that in the UK market, when commercials for a chocolate egg called Kinder Surprise ran, sales for other Kinder products declined – a result likely due to the fact that Kinder Surprise is more popular around Easter and the winter holidays in Europe.
Overall, the team encouraged Ferrero to think more holistically about the mix of products they’re advertising during certain weeks throughout the calendar year to ensure they’re getting the most out of their ad spend. Using the findings, the students were able to make detailed recommendations based on the data.
The team’s corporate partner, Erin Breland, praised them for their insights.
“What I was most pleased to see from the outcome was that despite the different sizes of our portfolios and the heritage in each of the different markets, there was commonality to the models that could be run on a central level,” said Breland, the head of global business intelligence and analytics at Ferrero’s headquarters in Luxembourg.
Essity is a global leader in supplying napkin products, as well as personal care and hygiene items. The challenge for the team was to come up with a pricing model that would predict how the type of distributor would influence the price they pay for Essity’s products, based on data about distributors’ customers, location, and product type. Using a lasso model, the Essity team highlighted almost 2,000 distributors the company is under-charging and identified specific attributes with the greatest predictive power that could inform setting the best price.
Students on the team with TE Connectivity, a telecommunication product manufacturer, sought to create a model to predict sales and revenues of the company’s six largest international business units. The team applied data analytics and machine learning to identify up to $200 million of missed opportunities in fiscal year 2018 and forecasted nearly five quarters ahead with 90-95 percent accuracy.
Cubic Mission Solutions
Cubic Mission Solutions, a defense contractor that helps clients gather and disseminate information securely, had challenges split into a “customer” student team and a “product” team. The Cubic customer team used a machine learning algorithm called “k-means clustering” to group Cubic’s clients based on the price and frequency of the products they bought. The team recommended Cubic use this information to inform the way they sell their products. By analyzing data regarding products, Cubic’s product team suggested they recommend items to clients based on past purchasing history and consider bundling products.
The beauty of the Analytics Accelerator Challenge is that students must work through the same challenges business professionals face in the real world.
Across all five participating student groups, students said the biggest lesson they learned was the importance of “cleaning” the data. Students in every group said they spent much of their time re-organizing datasets to account for missing information, or merging multiple datasets, before they were able to review the data.
The Ferrero team, for example, worked with datasets in different languages and the Cubic teams worked through datasets that came from multiple, disparate sources. Because of this, students crafted innovative ways to extract insights.
Wharton Customer Analytics Executive Director Mary Purk said she was excited by the students’ work.
“It’s amazing how the students come together with a wide variety of skill sets and in a matter of six weeks are able to come up with these types of findings,” Purk said, “I applaud them for all the work they’ve done.”
— Emily O’Donnell
Posted: November 27, 2019