Adaptive Identity Resolution on Databricks with Hightouch

How do you deliver meaningful, long-term engagement with your customers when you can’t even recognize them between interactions? Consider the scenario where a customer discovers your company on Instagram, visits one of your physical stores, purchases your hottest-selling item, and then returns home to set up a loyalty account so that they might receive information about new product offerings. You have a new, high-value customer, but in your back office, this individual shows up as three different people (Figure 1).

Failing to connect the dots on this customer and the hundreds of thousands of others like them in your systems has costs. When retailers can’t connect customer touchpoints, they send discount codes for products that customers have already purchased, fail to recognize their best customers in-store, and waste advertising spend on loyal, long-term customers instead of strengthening their relationships with those who are less engaged. More important than wasted spend, your customers do not feel recognized, and without a meaningful connection with your brand, they are more susceptible to poaching by your competition. In an easily accessible and hyperpersonalized marketplace, brands can’t afford to ignore this problem.

Understanding identity resolution for retail

Identity resolution provides organizations with the ability to recognize customers across various digital and physical touchpoints. Stitching together identities around overlapping personal identifiers, i.e., email addresses, phone numbers, etc., with fuzzy tolerances for misspellings, abbreviations, and the like, identity resolution enables organizations to recognize that an individual who browses online is the same individual who shops in-store with a quantifiable degree of certainty.

Assembling and maintaining this information is an ongoing challenge. New customers and new customer details are constantly being added to the mix. Key identifiers such as addresses and even names shift as customers present themselves in an evolving manner.

Third-party solutions have long existed to manage this complexity, providing organizations the ability to map records to an identifier based on limited personally identifiable information. But the cost of scaling these capabilities across the millions of customers managed by many retail outlets, along with growing concerns around data privacy that limit both the accuracy and the appeal of these solutions, means that an increasing number of organizations are seeking to move this capability in-house.

Introducing Hightouch’s Adaptive Identity Resolution

Given these challenges, Hightouch is proud to announce Adaptive Identity Resolution. Adaptive Identity Resolution (AIR) delivers in-house identity resolution capabilities built on your first-party data. With AIR, organizations can stitch together customer identifiers to recognize individuals across a variety of touchpoints, developing a clearer view of the customer without overspending on third-party data, working with black-box solutions with limited control, or violating customer privacy in the process.

What sets AIR apart from other identity resolution capabilities is how it allows you to toggle between degrees of certainty in establishing customer identities. In some scenarios, such as when you are delivering potentially expensive promotional offers to your best customers or are dealing with sensitive information, you may elect to employ a high-confidence approach to establishing customer identity. In other scenarios, such as when you are engaging customers early in their journey or are broadly advertising a promotional event, you may elect to be a little looser in your requirements for recognizing individuals. With AIR, you are in control of the identity resolution process, allowing you to deliver the degree of precision needed by your organization at that point in time. This is a unique capability unmatched by other identity resolution solutions (Figure 2).

To simplify this adaptive approach for end users, Hightouch’s AIR presents identity information within a series of zones. When your business needs precision, such as in billing communications, customer service interactions, and compliance-critical engagements, you may elect to employ identity information from the high-confidence zone, which uses exact, deterministic matching on customer attributes. When you can be more flexible, you may employ identity information from the high-reach zone, which uses AI-based probabilistic matching that accounts for variations to deliver scale and reach, rather than perfect precision.

The flexibility is key. The same customer data powers both approaches, but you choose the confidence level across use cases to maximize accuracy and growth potential.

AIR enables business teams to manage identity resolution directly, without relying on engineering resources. Marketing and operations teams can adjust matching rules as needs change, eliminating technical bottlenecks and accelerating results.

In-house in Databricks

The core technical challenge in any identity resolution solution is scale. Comparing millions of identifier combinations to each other to determine which likely refer to the same individuals on an ongoing basis is computationally intensive. By building Adaptive Identity Resolution on Databricks, Hightouch is able to take advantage of the data intelligence platforms’ unique scaling capabilities. This means that organizations can generate results on large volumes of data quickly and at the lowest cost possible.

In addition, by keeping the data in your Databricks environment, Hightouch’s AIR preserves data privacy and enables organizations to work within the geographic restrictions many governments are now imposing on the movement of consumer data.

And when the data processing is done, organizations are left with access to the Golden Record data set, which resolves data quality challenges like choosing an email address when a customer has provided multiple. It fits natively into the Hightouch Composable CDP ecosystem and is equally accessible across your bespoke enterprise solutions. As you expand your AI and analytics capabilities on Databricks, you benefit from unified customer profiles built in Databricks. You can easily leverage them with your machine learning models, predictive analytics, and other innovations.

Building business value from your unified profiles

Armed with flexible, reliable customer identity information in your Databricks environment, you can more fully tap into the potential of the information gathered from customer interactions. While Hightouch is widely recognized as a leader in the Customer Data Platform space, identity resolution is essential to a range of enterprise functions.

Cross-channel personalization

A customer who frequently purchases organic produce in-store using their loyalty card automatically begins receiving personalized mobile app notifications about new organic arrivals when they're near your locations. Their in-store purchase history powers their digital experience, creating the kind of seamless recognition that drives repeat purchases and builds lasting loyalty. This cross-channel personalization typically increases customer lifetime value by connecting previously siloed touchpoints.

Marketing attribution

Trace the complete customer journey: see how an Instagram advertisement leads to a store visit, which results in an online purchase, followed by loyalty program signup and repeat purchases. This complete attribution picture enables precise marketing spend allocation and channel optimization. Instead of guessing which campaigns work, you know exactly where your revenue-driving customers come from.

Fraud prevention

Identify customers attempting to exploit first-time buyer discounts by creating multiple accounts with different email addresses but similar personal information. The unified view reveals these attempts to game promotional systems, protecting profit margins while maintaining positive experiences for legitimate customers. You stop revenue loss without creating barriers for genuine customers.

Inventory optimization

Connect online browsing behavior to in-store purchasing patterns to optimize inventory and merchandising decisions. When you discover that customers who research winter coats online but don't purchase often buy accessories in physical stores, you can adjust both your e-commerce merchandising and in-store product placement. This intelligence directly improves conversion rates and inventory turnover.

In a market where customer experience increasingly determines winners and losers, you can't afford to let data silos undermine your ability to recognize and serve your customers effectively.

Hightouch's Adaptive Identity Resolution on Databricks provides the foundation for customer recognition that adapts to your business needs, leverages your existing infrastructure investment, and scales with your growth ambitions.

Ready to transform how your organization recognizes and engages customers? Contact your Databricks team or book a demo with Hightouch to see how unified customer profiles can unlock your competitive advantage.