Predicting Profits With Models

Understanding which customers will buy which product or service is at the heart of personalization, a booming and still evolving industry. In addition to better leveraging existing data and analytics, numerous new and rich sources of information are available to support predictive models that target the right consumers with the right products and offers.

Getting Started
The first place to investigate is still your customer data and interaction history. While this may not seem new to most marketers, it’s still amazing how underleveraged this information is.

The biggest issue is a lack of coordination between transactional, CRM and Web data repositories. In addition, most companies are still trying to mine this information for fairly static decisions focused on finding the right distribution channels or the right product bundles.

Personalization algorithms offer a significantly larger opportunity, but also require a more robust view of the data generation strategy to include situational information, such as time of day, location, customer data, purchase history, etc. Furthermore, how most companies leverage their Web properties to make better marketing decisions remains a huge opportunity for matching up customers to the right products or services.

Assuming companies have online ordering and fulfillment capabilities, information, such as historical purchases, individual market baskets, common market baskets and ratings/reviews can all lead to better alignment of offers to prospects. For companies that use their sites primarily to support the customer’s information-gathering needs, activities such as product or service viewed, newsletter sign-up and content sharing are also vital information for remarketing and personalized offer development.

In all cases, advanced personalization or targeting algorithms require significant effort around classification and categorization of products, services and content. If this seems like a significant challenge due to breadth of products or services, segment-level personalization engines can be used to present some level of customization.

Getting Help
Leading CRM solution providers already see where the possibilities are headed and are actively building or acquiring technology to support intelligent, data-driven and personalized product services offers. For example, Salesforce.com’s recent $400 million acquisition of RelateIQ is designed to leverage CRM data to help focus sales efforts to the right clients at the right time.

In addition to internal customer data, marketers need to also leverage outside data sources to better associate customers to buying opportunities. Historically, social data and social analytics have been viewed as public relations tools to better understand sentiment and communicate brand propositions and initiate conversations. However, social sites such as Yelp and Foursquare and marketplaces such as eBay are learning quickly that the power of their platforms is driven by the size and ingenuity of their developer networks. As a result, they are making their data widely (and many times freely) available through APIs.

This open access presents opportunities to explore the relationships between brands, products, networks and sentiment. It also opens the window to integrate social data with customer data. Open source and SaaS CRM platforms alike are now building capabilities to integrate a customer’s social data with relationship data (mostly permission-based). DataInformed provides more information.

Adding Data
In addition to social data, search data can inform relationships between products and brands. Targeted search, of course, has been offering a dynamic and simple way to target the best prospects. In addition, organic search data sources, such as Google Trends, offer deep insights into brand and product interest by time of day, geography and related searches. Although this information can be tricky to work with and has limited drill-down capabilities, it can provide great direction and confirm information gathered from other sources.

Finally, third-party data is becoming a powerfully rich source of targeting information. Long gone are the days where marketers would target consumers based on age, income and a handful of modeled segmentation variables. The aforementioned opening of social network data, combined with the prevalence of cookie/device ID tracking and the various customer data selling practices, have enabled third-party data to become a critical bridge in omnichannel marketing.

Government Oversight
The data aggregation process is so insightful, in fact, that it has drawn the attention of government regulators. In a May 2014 FTC report, the activities of nine representative data brokers, Acxiom, CoreLogic, Datalogix, eBureau, ID Analytics, Intelius, PeekYou, Rapleaf and Recorded Future, were examined.

The study revealed thousands of online and offline data points available for sale at the consumer level and collaboration between data providers to create even more powerful consumer-level data sets. The infographic shows the taxonomy of sources and activities currently available for sales by data vendors. Of interesting note is the contribution by government to the data brokerage industry.

An October 2014 article in Target Marketing focused on the efforts of the FTC to provide oversight of third-party data providers. The FTC to-date has primarily focused on making data security policy clearer and easier for consumers to understand. However, even with clearer data security policies, it is unlikely most consumers are going to take notice of how their data is shared.