Tracing the elusive digital customer without 3rd party cookies

A watershed moment for marketers when Google made the controversial declaration on 3rd party cookies. Before which, our Customer Analytics portfolios were flush with ‘segment of one’ and ‘hyper personalization’ and ‘audience mapping’ offerings. Now what? How do we deal with this unplanned glitch?

Not to ignore the demands from General Data Protection Regulation (GDPR), California Consumer Privacy Act (CCPA) and Safari Intelligent Tracking prevention, that pose tough challenges and steep non-compliance costs for marketers.

Over the past 10 months, I’ve worked across the spectrum with customers, internal stakeholders, and experts who have contributed to the Unified 2.0 standards.

A multitude of strategies such as independent but compliant data collection green rooms, setting up walled gardens, Customer Data Platform investments are available for scale roll outs. Yet for a business line like mine, most requirements were ad hoc and too small in scale — though not in pain — to justify big investments for my customers.

It’s not a single business problem either.

1. You may have an existing CRM, but want to link existing customer details with anonymized social feed and/or other customer inputs.
2. Or you simply want to reconcile duplicate customer identities within your burgeoning database.
3. Or you operate in a B2B2C medium and want to understand customers better, but don’t really have any end consumer data (Acquirer Banks and Card Associations in the C&P industry for example)

Yet from a solution standpoint the key is “persistent hashing”. Not technically a challenge per se, but a ton of hurdles like “establishing true persistence” or privacy constraints or inadequacy of relevant data ( For ex if you have SSN or Driving License nos of all your customers, you have a robust persistent ID, but generally these kinds of data points are not comprehensively available)

Multiple solutions explored during the past year, including the customized use of the Unified ID 2.0 (highly recommended).

1. Use of dedupe, blocking keys and matching algorithms to reconcile known and unknown customer information at an individual level for 2 clients with a large number of duplicated end customer data

2. Micro-segment profiling to supplement Digital VOC, and hyper-profiled recommendations for an Ad based channel

3. Session or Device level profiling for assisted shopping for an Ecommerce portal

All experiments have generated acceptable, but not perfect results. Market offerings are getting more and more innovative in balancing marketing needs with privacy and confidentiality concerns.

My team and I developed our Adaptive Marketing Analytics — Dynamic Identity Graph accelerator last year using some of these combined approaches. We were able to demonstrate the actionability of a hybrid approach to a large MNC successfully.

Methods are still evolving and sooner than later I expect we will have less experimental and more accurate solutions too.