Chapter 33 — Key Takeaways

Programmatic and Enterprise SEO at Scale — the one-page card.

The core claims

• Programmatic SEO is a neutral technique. Template + structured dataset = many pages, one per record. The same machine builds the web's most valuable properties and its worst doorway farms. The difference is entirely whether each page carries real value.
• Method is invisible to Google; value is everything. Google crawls and judges the finished page. There is no penalty for templating (or for AI assistance) — only for thin, near-duplicate, valueless results at scale ("scaled content abuse," doorway pages).
• A doorway page is a named spam violation. Pages built mainly to rank for a query/location while offering no value of their own, funneling users onward. The line between a location page and a doorway is real local content, not spun text. You cannot spin your way across it.
• Crawl budget is only a real concern at genuine scale. Most sites never need to think about it. It matters for sites with hundreds of thousands to millions of URLs, or large fast-changing ones.
• Enterprise SEO is a coordination problem, not a knowledge problem. At scale you recommend; others implement. Tickets, prioritization, and governance are the job.
• At scale, leverage cuts both ways. One template change improves — or breaks — every page built from it, all at once.

The quality bar — the four honest questions (apply to every generated page)

1. Would this page be useful if it were the only page on the site?
2. Does removing it make the web meaningfully worse for some real searcher?
3. Does it carry information gain — something specific to this slice a generic page lacks?
4. Could a competitor produce a byte-for-byte equal by running the same find-and-replace? (If yes → doorway.)

Doorway spam vs. legitimate location page

Crawl budget — the two forces (Google's own model)

Common crawl waste → fix: parameter/facet URLs → canonical + noindex then robots.txt; redirect chains → flatten to one hop; 404s/soft 404s → honest status codes; duplicates → rel="canonical"; deep pages → stronger internal links.

Log-file analysis — the ground truth of crawling

Reveals what Search Console's summary can't: exactly which URLs Googlebot fetches, how often, the status codes it meets, crawl latency of new pages, and where crawl is wasted. Verify Googlebot (reverse/forward DNS) first. Worth the effort at large scale only; small sites use Search Console → Crawl stats instead.

Scaling safely — the rules

• Subtract first. Build only matrix cells with real demand and a real local story. Fewer real pages beat more empty ones.
• Two layers per page. Shared service layer (templated, honest) + unique local layer (real, per-place).
• Double hub-and-spoke. Every leaf links up to its service hub and its city hub, and across to relevant siblings → no orphans, authority flows both ways.
• Never over-optimize anchors. Thousands of identical exact-match internal anchors is an engineered spam signal, not a boost.
• Treat template changes as site-wide deploys. Test on a sample; check rendered HTML, canonical, robots, and schema; stage and version; monitor by page-type, not page.

Key terms

programmatic SEO · template page · doorway page · location page at scale · log-file analysis · crawl capacity limit · crawl demand · index management · enterprise SEO

What you could do on your own site tomorrow

If you serve multiple services or areas, sketch your service×area matrix and run Rule 0: for each cell, can you name five true, specific local facts a find-and-replace could never produce? If yes, it's a location page worth building. If no, it's a doorway — don't build it. Then open Search Console → Crawl stats and check whether Googlebot is meeting mostly 200s or a pile of redirects and 404s.