E-commerce SEO · Agentic commerce · Data

The feed is the storefront. The rest is how machines read it.

I'm Brandon Lazovic, an e-commerce SEO consultant and data engineer. I write about how AI is changing the way people find and buy products, and what merchants actually control as agents and answer engines take over.

Read the writing About me

how a machine reads this page

{
  "@type": "Person",
  "name": "Brandon Lazovic",
  "jobTitle": "E-commerce SEO + Data Engineer",
  "knowsAbout": [
    "agentic commerce",
    "product feeds",
    "structured data",
    "AI search visibility"
  ],
  "sameAs": [ "linkedin", "youtube" ]
}

This site practices what it documents: every page ships server-rendered schema, clean semantic HTML, and a real feed.

Three threads

The writing clusters around three arguments.

Each piece is a deep dive that restores the technical layer the 280-character format drops. They roll up into 4 pillars.

The Feed Is the Storefront

The structured product feed is no longer a Shopping input; it is the surface AI rebuilds commerce around. Control is completeness, accuracy, and representation.

14 deep dives 02

How Machines Read the Web

Agents and crawlers read structure, not pixels. Semantic HTML, server-rendered schema, and embeddable item data decide what gets understood and surfaced.

31 deep dives 03

The Model Is the Dependency

Every model you build on is a dependency you do not control. It carries its own behavior, spends your context, and runs on the credentials you hand it.

8 deep dives 04

Crawl, Index, Rank

Three stages, in that order: a crawler fetches the page, an index stores it, a ranking system picks it. Most ranking problems are really crawl or index problems, and both are measurable on your own site.

6 deep dives

Recent

Latest deep dives

Daily signal lives in The Pulse

09 / 15Search Console counts your AI impressions and cannot connect one of them to a visitGoogle's generative AI report shipped worldwide with impressions and no clicks. Three separate changes hit AI citations in the same two weeks, and the report cannot price any of them.How Machines Read the Web 09 / 14Most of your bot log was never a crawlerMost of what looks like crawler waste is probing for files you never published. In my own logs it was 80.1%, and a keyword denylist put the same traffic at 21%.Crawl, Index, Rank 09 / 11Anthropic's commerce blueprint hands you the ranking functionAnthropic's commerce agents blueprint ships no payment code, no feed format, and no ranking. The method deciding what a shopper finds is one you write yourself.The Feed Is the Storefront 09 / 11A redirect that works in your browser can still fail every crawlerA browser resolves a relative Location header and caches whatever it is told. A crawler reads the raw headers, and this site's own 301 is missing one of them.Crawl, Index, Rank 09 / 10Google treats 404 and 410 the same, and the code that changes its behavior is 429Google's own docs say all 4xx errors except 429 are treated the same, so a 410 does not remove a page faster than a 404. The code that changes crawling is 429.Crawl, Index, Rank

Standalone feature

DuckDB just got a landlord: AWS is acquiring DuckLabs

DuckLabs is joining AWS. The MIT license and foundation governance hold, and AWS says it wants DuckDB to run best on AWS. Those are different promises.

Read the essay