Retailers are constantly looking for the right opportunities to combine, cross-sell or upsell products or services to consumers so as to maximize revenue lift.
However, the probability for success rises dramatically when retail leaders are able to target those combinations of products that consumers are most likely to purchase based on a number of variables: when consumers have certain items in their shopping carts, when a particular offer is most likely to lead to conversion, etc.
Here, we take a look at a fictitious example of a grocery store chain with hundreds of locations and numerous product categories to better understand how the use of predictive analytics and statistical scoring within market basket analysis and affinity analysis can be used to maximize the effectiveness of a cross-sell campaign.
In this case, we’re focused on cross-selling gift cards. Gift cards are high margin items for retailers, often because consumers forget to use all monies contained within.
The use of analytics and data discovery can be used to help retailers identify the categories of items that are most often purchased along with gift cards. A customer who has those items in his shopping cart but hasn’t yet picked up a gift card has a high propensity to buy a gift card and would likely require just a small incentive to do so.
Instead of manually sorting through each of these categories to make these types of determinations, Spotfire in tandem with TIBCO Enterprise Runtime for R’s (TERR) high performance statistical engine can be used to automatically rank each product category carried by a consumer to the likelihood of a gift card purchase.
Not surprisingly, greeting cards as a product category rose to the top. This makes sense since a customer who purchases a greeting card is fairly likely to insert a gift card.
TERR enables the discovery of products that are least likely to be purchased along with a gift card (known as negative predictors). These products include gasoline since gas is more likely to be purchased outside of the store.
One of the benefits of using a statistical analysis engine is that it can strip out those categories that don’t have any significant statistical bearing on gift card purchases, thus only emphasizing the positive and negative predictors.
Predictive analytics can also be used to identify revenue lift opportunities if the grocer were to run cross-sell offers on gift cards. TERR is used to calculate each of the stores and rank them in terms of their respective revenue lifts, and can also be used to help identify the optimal placement of gift cards on the floor to optimize sales.
In addition, TERR can calculate the cost of an offer (e.g., the percentage of customers who are most likely to accept the offer, which includes receiving a free coffee at the front of the store that costs the retailer 75 cents in each instance).
The use of data visualization tools can also help decision-makers for the grocer visualize the cumulative uplift from the offer as more and more stores are added.
This includes depicting on a map which stores are generating the greatest revenue contribution from the offer. Such insights can help retail leaders identify those stores where the offer should be extended that will generate a positive revenue lift, striking out those stores where the offer would result in a loss.
The retailer can also conduct what-if analyses to explore how the revenue lift would be impacted if the variables were to change (e.g., changing the duration of the offer and whether this would result in a positive or negative revenue lift).
Using a predictive analytics suite that can calculate and display a full range of predicted outcomes can help retailers not only get the biggest bang from their cross-sell initiatives but also enable cross-organizational understanding of the resultant lift models and position more than just the statistical team to benefit from applying models.
- Please join us on Thursday, August 14, 2014 at 11 a.m. EDT, for our complimentary webcast: “Analyze Your Customers’ Actions: Market Basket and Customer Analysis with Spotfire.” In this webcast you will learn how Market Basket Analysis can help you leverage POS data to understand sales patterns, customer preferences, and buying patterns to create targeted and profitable promotions.
- Subscribe to our blog to stay up to date on the latest insights and trends in big data and big data analytics.