Personalization as a Service

Market: brick-and-mortar retail, ecommerce

Problem Space: customer value experience

Problem: retailers commonly offer sales and coupons to incentivize customers to purchase various products. Offers like these are generally provided to customers based on customer data that suggests what offers a group of customers may respond positively. Offers are not generally designed specifically for a particular customer, however. As a result, the offers may not incentivize each individual customer.

Assuming a retailer has the capability to identify targeted offers for each customer, they would face numerous technical challenges in actually implementing digital advertising campaigns that utilize these targeted offers. For example, identifying personalized offers for an entire customer base, which may include millions of customers, and disbursing such personalized offers may be impracticable due to scaling challenges. The timeliness of an offer may also affect the efficacy of the offer. A targeted offer for a customer may be based on outdated customer data, so that by the time the customer data is analyzed, a targeted offer is generated, and the targeted offer is delivered to the customer, the targeted offer may no longer be relevant or motivating to the customer. Providing personalized offers on a large scale may require increased time for data transfer and data analysis, yet the increased time may cause the personalized offers to be less effective.

Solution: a PaaS offering that provides omni-channel data syndication, generalized to support the personalization of any type of content.

Beginning with the customer value experience problem space and the personalization of offers and offer-related content for each customer at scale, a set of base data requirements was established to determine what data transformations would be required to support personalization on a mass scale. The dataset in its most actionable state became a collection of unique offer identifiers per customer, along with other meta data to help determine and associate related offer content and placement, as well as support further sorting and filtering of the offers. Once received, the actionable data set is made available to internal and third-party applications via a private API endpoint.

Thanks to the PaaS offering’s utilization of cloud infrastructure and a robust HA/DR strategy, any application with the proper security permissions can receive the data required for personalizing offers on demand, at the moment of customer interaction with the content, by calling the API utilizing certain parameters for sorting and filtering. Because the PaaS solution is data agnostic (the suite of APIs providing high-volume pass-throughs for collections of things for each customer), the solution can be employed for any number of enterprise use cases that call for personalized content at scale.