We use cookies to personalise content, to provide social media features and to analyse our traffic. We also share information about your use of our site with our social media, advertising and analytics partners. For information on how to change your cookie settings, please see our Privacy policy. Otherwise, if you agree to our use of cookies, please continue to use our website.

Connected Recognition

Enabling a 360⁰ View of the Customer

Merkle’s Connected Recognition process manages and associates identifiers across channels and media to fuel the execution and optimization of integrated marketing and sales activities.

 

 

There is an astounding amount of available data about every person. From traditional census data to emotional and personality-driven insights, it is theoretically possible for a brand to paint a highly detailed, accurate picture of each of their customers. But this promised land of insight requires several often-siloed departments to share information, not to mention significant technical and process implements to collect, store, and interpret the data.

As a result, marketers resort to sub-optimal point solutions. For example, many brands use a small purpose-built data integration or feed email remarketing. This typically leverages an email identifier only. Similarly, for site optimization, marketers are beginning to leverage personalization, but typically ignore the rich CRM data that should inform decisioning. Each approach ignores valuable amounts of data related to the history of activity, depth of the interaction, and breath of insights available through more robust linking approaches. Connected Recognition (cR) overcomes this and more.

The Identity Map

The solution creates an identity map across all party or user identifiers that become tied to an individual. This map becomes your reference base by cross-referencing all identity attributes related to both “known” and “unknown” data. A master party (or customer) identifier is generated to leverage all possible information from both strong and weak identifiers.

  • Strong identifiers are those that can be linked to an offline user identity in a customer data file (such as name and address). These include transaction IDs such as online order, registration or lead IDs, as well as account login numbers, customer numbers, service tags, etc.
  • Weak IDs are the primary identifiers found in digital interactions. By their nature, they are anonymous and often very transient. Examples of weak IDs include IP addresses, cookies, device fingerprints and social login or handle.

The Connected Event Stream

Once the identity map is created, the Connected Recognition solution creates an event stream where all customer engagement is tied back to the master identifier. This event stream is created for all “known” parties, whether they are repeat customers or anonymous parties that visited the website. All consumer interactions can be tied to the event stream to support integrated, cross media and channel marketing decisions. This includes outbound marketing touches such as direct mail, display impressions, email, or mobile texts, and inbound interactions such as important events that occurred while a consumer browsed the website – downloading a white paper, registering for email, browsing a product or filling out an information request form. The final output is something that looks like an active timeline for each consumer with the brand.

Click here to view full size image

The recognition solution preps the data for the identity matching process by applying data standardization and hygiene procedures. The processes are unique and customized for each type of data (e.g. email, mobile, terrestrial address). Once the data is clean and standardized, identifiers are connected back to common parties. The process is a combination of the traditional offline customer data integration (CDI) and an additional processing to handle all of the anonymous linkages, a process we call digital data integration (DDI).

DDI supports the linking of high volumes of digital activity to classic CRM data; the most challenging component is the web data. Large online marketers may average upwards of two, three, or four million unique visitors a day – this extrapolates out to tens of millions of page views per day.

The Connected Recognition Process for DDI

The solution uses the latest technology to manage big data by organizing and aggregating it for sales and marketing activities.

  • Data loading:  At an international cookie level, cR loads batch or real-time web feeds containing customer activity, marketing impressions across media, and the full web log file. The product has a flexible, meta-driven extract, transform, and load (ETL) process to allow for different configuration sets for different countries. When changes are made to the website, the product is reconfigured within 24 hours to pull in new data elements, process new fields, or eliminate fields. Because 100% of the data from the web log is ingested, no coordination is needed to support changes.
  • Data filtering and QC: Web data is messy and often considered bad. cR utilizes standard filtering and QC reports to indicate how much data is outside of pre-defined tolerances to identify issues early in the process.
  • Linking: Merkle designed specialized linking algorithms that boost match rates with fewer identifiers. For one top online retailer, the solution increased identification rates from 10% to 25% of visitors. The process allows matching and processing on multiple identifiers and is also backward-looking, sifting efficiently through up to six months of anonymous history to find matches to newly “known” visitors. It handles ID conflicts, such as visitors logging in under conflicting IDs in the same session. Additionally, different linking rules can be configured for specific business needs and country-specific requirements to ensure compliance with country-specific privacy standards. The linking ties digital activity back to terrestrial information, which layers in the full range of CRM data, inclusive of demographics, customer purchase activity, service history, and more. An example is shown below.

An anonymous consumer is looking for auto insurance. The consumer’s experience is as follows:

Click here to view full size image

With Merkle’s Connected Recognition linking, it is possible to connect nearly all of these interactions across multiple devices back to a single master identifier. Note that by the end of the example, all of the devices used by the individual can be tied back to a strong identifier. The final associations after linking would be as follows:

Click here to view full size image

Effective linking is challenging because of situations such as a consumer using multiple devices. Merkle’s process is designed to run efficiently and manage across multiple devices and types of identifiers.

  • Data Transformation: Weblog data in particular is not designed to be easily used for customer-driven marketing programs. The solution captures and stores over 100 different web interaction types that occur for anonymous and known customers at a session level. The data model is designed to be especially efficient, with final data volumes nearly 20X smaller than the input feed. Standard aggregates are then created to support most common targeting, analytic and reporting uses.  

Connected Recognition Powering Connected CRM (cCRM)

Connected Recognition is the future of CRM. With the digitization of media and channel interactions, the ability to maintain a traceable 1:1 relationship across classic CRM data and online engagement is critical.

Merkle’s solution leverages the power of big data technology, but was collaboratively designed by marketing, technology, and analytics professionals to build a solution that provides a positive ROI for CRM programs. The solution takes massive amounts of unstructured data across devices and makes it actionable for email, direct mail, display, site recommendations, and more – in a global environment.

This integrated data allow us to not only build and execute better programs but to make sales and marketing activities more measureable and accountable. Here is a small sampling of results from Merkle’s Connected Recognition solution:

  • Stronger offline CRM modeling (double-digit lift in key segments)
  • More efficient integrated email programs (multi-million-dollar lift in revenue for web-based trigger programs)
  • Significant lift in display optimization and cost savings on DMP data integration
  • Huge gains in site personalization ($10MM+)
  • Improved site pathing based on integrated online and offline data and organization and analysis of online only data (a 3% increase in conversion and over $3MM in incremental revenue)

For more information about Connected Recognition, contact us.