Learn · track

Martech and data

The tools behind integrated marketing, how they fit together, and the customer data that connects them, collected with consent.

4 pieces28 min

  1. Building a martech stack: layers, choices and vendors

    The four layers of a marketing technology stack (data, engagement, content and measurement), what joins them, when to build, buy or compose, and a procedure for choosing vendors on use cases, interfaces, data, consent and exit terms.

    In short
    • Think in four layers (data, engagement, content, measurement) with consent running through all of them, and name one owner and one system of record for each.
    • Most stacks are under-used, not under-equipped: Gartner's 2025 survey put utilisation at 49%. Fix process and integration before adding a tool.
    • Choose vendors on three to five written use cases shown on your own data, and on how you leave: export, contract terms and who owns the data.
  2. First-party data and customer data platforms: when a CDP is worth it

    What first-party and zero-party data are and why they matter more as browsers block tracking, what a customer data platform does, packaged versus composable CDPs, how identity resolution works, and a test for whether you need one at all.

    In short
    • First-party data is what you collect in your own relationship with a customer; zero-party data is what they choose to tell you. Both still need a lawful basis and, for many uses, consent.
    • A CDP builds one persistent customer record and makes it available to other tools. Packaged CDPs keep their own copy; composable ones work on the company's data warehouse.
    • A CDP is worth it when several sources must be joined for named, valuable use cases and someone will own it. Without those, a CRM or the messaging tool you have is enough.
  3. Consent, identity and cookies: where things stand in 2026

    What consent rules cover, where Chrome, Safari and Firefox stand on third-party cookies as of October 2026 after Google kept them and retired most of the Privacy Sandbox, how consent mode and the TCF pass choices on, what server-side tagging changes and what it does not, and where clean rooms fit.

    In short
    • Third-party cookies did not disappear: as of October 2026 Chrome keeps them under user choice, while Safari and Firefox block cross-site tracking by default. Plan as if they are unreliable.
    • In the EU, consent is about access to the device, not only cookies: pixels and similar techniques are covered too. Server-side tagging moves processing, not the need for consent.
    • Make one consent record the reference, pass it to every tool (consent mode, the TCF), and measure with methods that hold up when part of the audience says no.
  4. AI in marketing: what the evidence supports, and how to run a pilot

    Where predictive and generative AI are used in marketing, what controlled studies show about productivity and quality, the risks to manage (accuracy, brand safety, intellectual property, bias and privacy), disclosure rules as of October 2026, a governance set-up and a procedure for a pilot that proves value or stops.

    In short
    • The best evidence so far shows faster work and better drafts on well-defined tasks, with the largest gains for less experienced people; it does not yet show effects on sales or brand.
    • Capability is uneven: tasks that look equally hard can fall on either side of what the model does well. Keep a human accountable for every claim, image and message that goes out.
    • Since August 2026 the EU requires deepfakes to be disclosed and AI chat to be identified. Run every pilot with a baseline, a control and a stop rule.

Other tracks