Workshop notes,
logged.

AI answer engines read your pages to decide whether to cite you, but the "GEO" and "AEO" industry mostly sells markup with no mechanism behind it. What actually helps: a clean entity graph, honest schema that matches the page, copy a retrieval system can lift, and letting the right bots crawl. Plus the popular advice I deliberately ignored.

A perfect Lighthouse score is a standing proxy for conversion rate, paid-media efficiency and organic ranking, not a vanity badge. A generalised guide to getting Performance, Accessibility, Best Practices and SEO all to 100 and keeping them there, the measure-fix-re-measure loop in Chrome DevTools that does it, and the regressions a fast first paint can introduce.

Google's Conversational Analytics (Preview) lets a team ask a BigQuery table questions in plain English. The risk isn't that it can't write SQL; it's that it writes confident, wrong SQL. How to stand one up that's read-only by IAM, accurate on the trap questions, and reproducible from code.

When 70-80% of revenue arrives via phone call days after the click, the default ad-platform stack optimises against the wrong signal. A five-tier funnel, three-layer conversion taxonomy, the non-negotiables, and the four phases that get you there without polluting bidding data.

Drop your Google or Meta ad, pick three rival angles, and a simulated three-person customer panel scores all four. Iterate using the panel's feedback until your copy wins, then export the winning angle as a Google PMax asset group.

Not reporting infrastructure. The foundation that makes credible attribution, MMM, and audience automation possible. How we build one for a multi-brand operator running approximately fifteen DTC stores across Europe.

Bayesian MMM with Google Meridian as the engine. Priors that incorporate domain knowledge, posteriors with honest uncertainty, and multi-scenario budget planning as the deliverable rather than another static report.

Enhanced Conversions and CAPI feed the networks the signal they need to bid well. MMM tells the business where to spend next quarter. Most teams run only one of these layers well, and it shows in the results.

Server-side conversion APIs recover meaningful ground against consent denial, blockers and cross-device loss. A realistic read on what Google Enhanced Conversions and Meta CAPI fix, what they don't, and where the actual ceiling sits.

IM review, management interview prompts, systems access questions, attribution red flags, and the three numbers that tell you whether a target's growth is real or model-driven. From multi-million pound deals advised since 2023.

Catch the class of Consent Mode v2 failure that silently kills tracking for weeks. Seven 60-second staging checks before any production deployment that touches measurement.