Why the advertising stack was always going to move to cohorts, aggregates, and shorter memory, written well before the industry admitted it.
An early argument that the direction of travel in digital advertising was not a regulatory inconvenience to be waited out but the architecture itself changing: cohort-based targeting, aggregated and anonymous reporting, and analytics that stop depending on a durable record of a person.
It anticipated what the public clouds would do next, building anonymisation and privacy controls in as defaults, which in turn shortens retention and replaces the individual customer record with statistical and modelled approaches.
Read now, it is the blueprint under our marketing-operations practice: build the stack so that it works on aggregates and first-party signal by design, rather than retrofitting consent onto an architecture that assumed the opposite.
This is our summary, not the article. The full text stays with the original publisher. Read it there.