With a comprehensive data strategy, a retail media network is able to:
- Capture information in a structured way at every engagement point. Even retailers who have invested in a customer data platform (CDP) often find that some data is not easily accessible or queryable. Loyalty data might be in one location, point-of-sale data in another, and online shopping data in a third. A retail media network isn’t effective unless first-party data can be integrated with other repositories, such as catalog data and the category information management system.
- Disseminate information to the customer at every interaction point. This could take the form of an offer or discount, or product details such as nutritional information.
- Capture clickstream data on each product. Most shoppers look at multiple products before choosing one. Capturing this information allows you to create a model based on category affinity.
Building out such a data strategy is a significant undertaking. In general, retailers proceed through four stages of maturity. As they progress, they gain the ability to personalize more effectively, show value to advertisers and even enable their in-store sales associates.
Stage one: A modernized data platform
In this stage, the retailer has succeeded in building a structured data platform that holds a vast repository of customer data, and can access it quickly and seamlessly.
To be successful at this stage, it’s critical for teams to understand the location and cadence of their data. Some data might arrive in real time, while transaction data might show up overnight. Other data may be on a weekly schedule. How do you access all this data to get a real-time understanding of what is happening in your store and on your digital platforms?
Stage two: Data integration
In this stage, retailers move beyond customer data to include data from other silos and platforms. This expansion helps uncover relationships between customer, product and sales data. Now the models can find patterns at a much more granular level: those that are distinct to a holiday weekend, for example, or a particular weather or geography.
To be successful at this stage, a clean data room can be built with integrations that allow relevant data to be ingested—along with the proper relationships and attributes—to be used in analytical models.
Stage three: Models and predictive intelligence
The next step is to build and implement models on top of the data platform that will analyze the shopping and buying patterns of each customer. Those models, in turn, give brands a view into how they can best address each potential buyer.
To be successful at this stage, the customer should be the central focus of all models. The attributes are all related to the customer: product and category purchase preferences, buying patterns and category relationships (showing which products customers tend to buy together), the average basket size of a regular shopping visit, and more. The insights from these models become critical in attracting advertising from manufacturers.
Stage four: Generative AI and large language models (LLMs)
At this stage, retailers are already curating information for their brand partners. By using generative AI and LLMs, that information can become useful to store associates as well. If a customer approaches an associate and asks where a product is located or if it is in stock, the associate can quickly do a natural language query to find out. As the system evolves, it can answer more complicated questions: Perhaps the customer wants a particular food, but they need the version without nuts to accommodate an allergy, and it needs to be delivered by Friday. By integrating information across the enterprise and allowing LLM access, retailers can improve customer service significantly.
Retailers have an important asset in their first-party data, but without the proper infrastructure to structure, access and make the best use of it, it remains under-leveraged and under-monetized. By building a modern, data-driven engine, retailers gain the ability to put that data to use through retail media networks—allowing them to engage in more effective collaboration with brands and more relevant messaging with customers.
Bruce Steadman and Rakesh Thaploo are Kyndryl Consult partners.
