Programmatic Audience Strategy in a Cookieless World: A DV360 Playbook

The cookieless transition stopped being a future problem years ago — Safari and Firefox blocked third-party cookies long before Chrome's saga, and signal loss already shapes every campaign we run. The advertisers thriving in DV360 today are not the ones who found a cookie replacement; they are the ones who rebuilt audience strategy on layers that never depended on cookies.

Here is the architecture we deploy, in the order we deploy it.

Layer 1: First-party data — the foundation

Your own customer relationships are the one signal source no browser can deprecate.

Customer Match onboards hashed emails and phone numbers into DV360 for direct targeting and exclusion. The strategic uses, in rough order of ROI:

1. Suppression — stop paying to prospect at existing customers. Unglamorous, instantly measurable.
2. Lifecycle segments — lapsed buyers, high-LTV cohorts, category buyers due for replenishment.
3. Seed audiences — the raw material for lookalike expansion.

Floodlight-based site audiences capture on-site behaviour (visitors, cart abandoners, converters) under your first-party consent framework. They power remarketing that survives cookie deprecation because the relationship is direct.

The prerequisite nobody skips anymore: a consent management platform wired correctly, and a CRM whose segments mean what they claim. We spend more discovery time on data hygiene than on media strategy — deliberately.

Layer 2: Google signals — scale with intent

DV360 exposes Google's audience intelligence, built from logged-in, consented ecosystem data rather than third-party cookies:

• In-market audiences — actively researching a purchase category
• Custom intent — built from the search terms and URLs you specify
• Affinity and custom affinity — durable interest profiles for upper funnel
• Demographics and life events — moves, weddings, graduations

The craft is combination logic: in-market and affinity intersections beat either alone; custom intent seeded from your actual converting search terms beats generic categories every time we test it.

Layer 3: Contextual — the renaissance

Contextual targeting — placing ads based on what the page is about, not who the user is — has quietly become sophisticated: real-time page-level classification, sentiment awareness, and category granularity far beyond the keyword lists of 2015.

In DV360, contextual layers serve three jobs: reaching cookieless environments (Safari, Firefox, CTV), building privacy-safe prospecting reach, and acting as a brand-suitability filter on everything else. Our default: every prospecting line item carries a contextual layer, either as targeting or as suitability guardrail.

Layer 4: Modeled everything

Accept that measurement is now partly statistical. Enhanced conversions, consent-mode modeling and data-driven attribution fill consent gaps with modeled estimates. The discipline that matters: keep one measurement spine (Floodlight + GA4), document what is observed versus modeled, and validate models against holdout experiments a few times a year.

Putting it together: a reference architecture

For a typical full-funnel program:

Funnel stage	Primary audience	Layered with
Awareness	Affinity + contextual	Frequency-managed CTV/video
Consideration	In-market + custom intent	Contextual suitability
Action	Site audiences + Customer Match lookalikes	Recency windows
Retention	CRM lifecycle segments	Suppression of recent buyers

Every layer above works today, without a single third-party cookie, inside standard DV360. The advertisers who struggle are those still waiting for a like-for-like cookie replacement. It is not coming — and with this architecture, you will not miss it.

Want your current audience setup stress-tested against this playbook? Our managed and co-managed teams run exactly this audit in week one. Talk to us.