I write about martech, technical SEO, and the seam between marketing and engineering — the broken pipe in a stack, and why it matters to the people who have to explain it upward.
Some of these pieces also run on LinkedIn, and Hackernoon — this site is the canonical, permanent version of everything.
Posts
-
The hreflang tags were correct. Every one of them redirected first.
Two hreflang failures, years apart and worlds apart in resources, both missed for the same reason: nobody checked whether the tags' targets loaded directly or bounced through a redirect first. -
AI agents expose a gap between SEO and dev work. Marketing should own it.
SEO has AEO, GEO, ASO, and ADO, but none of them cover what happens after an AI agent lands on your site and has to actually use it. That gap belongs to marketing, not engineering. -
A practical AI governance framework (and what agent files change about it)
Combining and expanding four earlier posts into one working framework — registry, risk tiers, review process — plus what AGENTS.md-style files mean for the whole thing. -
Some web pages are now written for your AI agent, not you
A new kind of document is showing up online, install guides written for AI agents to execute rather than for people to read. A real newsletter is the case study. -
From readable to actionable
What happens when an AI system can stop reading your site and start doing things on it — closing out a six-post series on agentic readiness. -
llms.txt is worth watching, but not overselling
Broad LLM crawlers mostly ignore this protocol; specific agentic tools are actively using it. It's cheap to create — but it's not a strategy. -
Making YMCA information machine-readable
Sitemaps get you found, schema ensures you're understood correctly once you are — pulling apart two things marketers lump together as 'technical SEO stuff.' -
Preparing for agents starts with better user experience
None of this requires new technology — the accessibility work that's been on every audit checklist for a decade is exactly what makes a site usable by AI agents. -
Ranking, being cited, and being usable are different
A page can rank #1, get cited by ChatGPT, and still fail once an AI agent tries to use it — SEO, AI visibility, and agentic readiness are three separate things a site can pass or fail. -
AI search may only be the transitional stage
Search engines helped people find a page. Agents might soon finish the task the page existed for — the first post in a series on what that means for organizations like the Y. -
Marketing brings the use case. IT checks the data path. Legal checks the exposure.
The final piece of the AI governance framework — a review process that only pulls in legal and IT when the risk tier calls for it. -
Not every AI use case carries the same risk
Sorting AI use cases into 3 risk tiers makes the review process fast instead of exhausting — what belongs in each tier and who signs off. -
Don't start your AI policy with 40 pages nobody will read
Start with a 5-column spreadsheet instead — what a working AI tool registry actually needs to track. -
A marketing coordinator finds an AI writing tool on a tuesday
What AI governance failure actually looks like — a normal person doing normal work inside a policy vacuum — and the framework built to get ahead of it. -
I wrote ruby on rails for years before i ever touched a content calendar
A clean data foundation is invisible until it breaks — on bad inputs, AI hallucinations, and who gets left out when data skews toward the easy-to-measure. -
I once pulled a report where our ad platforms claimed more sales than the company actually made
Why every ad platform overcounts its own contribution, and why one source of truth plus a holdout beats stacking dashboards. -
20 minutes disappeared into going around the room naming the first concert we ever attended
On mandatory fun, companies that call themselves family, and why the volunteer day at the food bank actually works. -
I spent a wednesday afternoon in may at the city club watching a room of executives read a Kafka story
On Kafka's Poseidon, martech plumbing that's finally ready to run itself, and why the desk still feels safer than the ocean. -
A tool wrote 30 pieces of content for a team i know last month
Why activity metrics like 'posts generated' and impressions are decoration, and the test every number on a dashboard should have to pass. -
The riskiest thing you can do to a website that already works is redesign it
Why redirect maps are the unglamorous deliverable that actually determines whether a site migration keeps its rankings. -
Nearly 400 people showed up to hear that AI has moved from experimentation to execution
What it actually takes to govern AI-drafted content for a multi-branch nonprofit — and why writing the rules took longer than the drafting. -
Where are your AI superpowers actually hiding?
Don Shin's six places where AI agents create disproportionate value, and why the answer to vendor lock-in is owning the one thing that matters. -
I used to think AI vendor lock-in was a tomorrow problem
Don Shin on the pattern behind AI vendor lock-in, and why the trap never feels like a trap when you're moving fast. -
Picture it: 2018, MFS Supply. I built myself into a corner.
How building a dashboard nobody else could maintain taught me to build the system first and the dashboard second. -
AI content is highway hypnosis
Why AI-generated content leaves readers blank, and what that means for SEO as the meeting point of marketing technology and human desire. -
I used to think data should come in a nice tidy package
What building dashboards for a chemistry lab taught me about zen, dirty data, and asking the right questions before you build anything. -
This might get me fired
Watching a carefully built UTM attribution structure collapse in Daxko, and what fermenting pasta sauce has to do with directional data. -
Gene Autry would not approve of your Hubspot workflows
How agile culture and vibe-coding turned 'move fast and break things' into documentation debt with a deadline — plus a fermented pasta sauce recipe. -
Recycling other people's ideas is often called innovation
Evaluating server-side Google Tag Manager for a mid-sized nonprofit, with a detour through fermented hot sauce. -
I like alcohol (or: What making mead taught me about marketing)
What fermenting honey wine actually has to do with nurture sequences and data hygiene.
subscribe via RSS