How retailers can transform their marketing by leveraging predictive analytics techniques to solve business problems…
Retailers and their marketing teams could be forgiven for feeling a little punch drunk. While they were wrestling with seismic changes in the retail industry, along came a global pandemic to accelerate that change, and bring a whole new set of headaches.
It’s in this tricky environment that we’d like to talk a little but about predictive analytics. Far from being a mere fixation for the techheads in the data department, predictive analytics can be a lifeline for marketing departments in times of turmoil.
Why? Because the deployment of simple and not-so-simple predictive analytics techniques can answer your actual questions. Questions you need to answer to move forward in your marketing planning and drive growth in your sales.
Here are six examples of questions you could answer with a predictive analytics project…
A huge number of factors impact a store’s success: the number of competitors nearby, proximity to a carpark, rental costs, experience level of the store manager, the list goes on...
Modelling profitability at a store level can help assess which factors have the greatest impact on revenue, allowing you to look for locations with similar attributes to those stores which are most successful.
The answer to this question can guide anything from how much stock to produce to when to start marketing to a customer to encourage their next purchase. Survival analysis delivers insight into when a major event, such as a customer making their next purchase, or ending their subscription, is likely to occur.
This insight can be leveraged to hone and personalise marketing activity, ensure that next purchase happens and ultimately increase loyalty and customer retention. What marketing team wouldn’t benefit from that?
Retailers can use econometric modelling to forecast future revenue. Econometric models gather as much data as possible on factors which impact sales across a given time period, commonly a year. For example, ATL and BTL marketing spend, major marketing campaigns, weather trends, major cultural events and product launches.
Plotting this data at a daily resolution against revenue can reveal what impacts revenue most. You may end up surprised – and if you’re surprised, you can change tack knowing you’ll get incremental gains.
The relatively new kid on the block, in-store video analytics provides vast customer insight. Customer demographics, time spent at specific points in the store and reactions to products and in-store experiences are all key data points that modern in-store video analytics can capture.
Cameras can also be used to tailor in-store experiences to individual customers. For example, Californian eatery Caliburger uses cameras to identify customers on their loyalty scheme and display their favourite meals as they approach the till. Cameras also anonymously capture the demographics of those who show interest in a new product launch in-store, helping inform marketing strategy.
We’ve all experienced the frustration of receiving a recommendation for an irrelevant product. At best, customers shrug it off as inevitable, but at worst the recommendation could drive down a customer’s opinion of that retailer.
Either way, it represents a missed opportunity. Collaborative filtering is a technique used by online retailers to inform product recommendations at the individual customer level. It uses a customer’s URL history to recommend the product they’ve browsed the most, or items which customers with similar histories have bought.
Or it takes a customer’s purchase history, and recommends products bought by other customer’s with similar purchase histories. Recommendations account for 35% of Amazon’s revenue, the company that popularised collaborative filtering. It’s worth investing in getting it right.
Marketing spend budget is rarely unlimited. So, knowledge of which customers will generate the most return on marketing investment is invaluable.
A customer’s Lifetime Value (LTV) helps assign marketing spend at the customer level. LTV is a predictive analytics technique that combines purchasing frequency and value, ideally with churn and survival models, to estimate how valuable a customer will be throughout their time as your customer.
Retailers can target those customers with a high LTV with messaging designed to boost relationships and loyalty, while those with a lower LTV are targeted with promotional offers to increase spend and loyalty.
Answering burning questions like these bring predictive analytics out of the data department and into the wider business. Predictive analytics offers a diverse array of techniques to retailers and helps you remove the element of assumption from decision making, instead using data to influence future success both online and in-store.