Lessons, post-mortems and the things I got wrong, from two decades of measuring search. No theory I have not lived.
The single most expensive mistake I see is a site redesign or replatform shipped without a redirect map and a measurement plan. The new site looks great and the traffic is gone. The lesson I relearn every few years: treat a migration as a search project first and a design project second. Map every URL, watch the logs, and do not celebrate until the curves recover.
Teams fall in love with the clever launch page and neglect the boring guide that answers an actual question. Search rewards the boring guide. Write for the question someone is really asking, in the words they really use, and you will outperform the prettier page almost every time.
Every major update gets blamed for traffic drops that were building for months. When I dig into the logs, the decline usually started long before the update, the update just made it visible. Measure continuously and the surprises mostly disappear.
The same fundamentals that win on Google, clear structure, real expertise, being genuinely citable, are exactly what get you cited by an answer engine. GEO is not a different religion. It is SEO with the volume turned up and the margin for fluff turned down.
Big keyword volume looks great in a deck and converts like wet cardboard. The pages that paid the bills for my clients were almost always lower-volume, higher-intent queries that the volume chasers ignored. Intent beats volume. It is not close. This is the whole reason I rebuilt how I teach keyword research.
I have watched teams burn quarters trying to out-publish a competitor who was also guessing. Stop. The benchmark is what already ranks and what already gets cited, not what the other guy is doing. Read the results, not the rumor mill.
Given a choice between a brilliant content campaign and fixing the reason half your site cannot be crawled or rendered, fix the crawl first. I have recovered more traffic with a redirect map and a render fix than with any single piece of content. Foundations are unsexy and they win.
More field notes go out in the newsletter before they ever land here.
Get the newsletter