Case studies
Four engagements, covered in depth: the diagnosis, the constraints, what was built, how it was measured, and the limits of what the result proves.
Four rather than eight. Several results in my record have figures whose period or basis is not yet confirmed, and a case study written around an unverified number is worth less than no case study at all. Where a figure is pending, the page says so.
Retail electronics. Google Ads restructuring, product data quality, lifecycle email and CRM integration — and why the feed mattered more than the bidding.
Multi-category retail. A marketing department built from nothing, and the measurement baseline that made later claims defensible.
Moving from national to international selling: credibility, demand generation and qualification for long-cycle technical buyers.
Defining what return on marketing investment means when the business has few, large, long-cycle deals.
Every case follows the same structure: context, business problem, constraints, strategy, measurement setup, actions, result, limitations, and what I took from it.
The limitations section is not a disclaimer. In every one of these engagements several things changed at once, which is what real work looks like and what makes clean attribution impossible. Saying so is the difference between a case study and an advertisement.
Numbers are published only with their period, their basis and their method. Several figures in my record are currently held back for exactly that reason, and each page names which.
The most useful conversation usually starts with which constraint you think you have — and whether the evidence supports it.