What a 100–200 Page Content Library Costs to Build

"What does a 200-page site cost?" usually gets answered with a page rate multiplied by 200, which quietly prices 1 of 5 stages and none of the engine. The honest budget has five line items, two of which are one-time, three of which scale with page count — and on our own 3 production builds, the infrastructure line is so small it barely deserves a row [our data]. This page works the model at 100, 150, and 200 pages so the arithmetic is checkable.

What are the line items?

Two one-time, three scaling, one optional:

The order matters more than the amounts: research that can still say "no" belongs before the engine, and the contract belongs before page one — on our first build, audit corrections concentrated in pages written before the contract existed, while contract-governed pages audited clean [our data].

What does the production line cost?

Take a per-page lifetime labor figure and multiply — ours is the illustrative 7 hours per fact-audited page (research 1.5, draft 2, audit 2, publish QA 0.5, refresh 1 annualized) worked through in what a fact-audited page costs. The model at three library sizes:

Audit plus refresh is 3 of the 7 hours — about 43% of lifetime labor, or 450 of the 1,050 hours at the 150-page size — and that share is exactly what discount library quotes omit. A "$150 per page, 200 pages, $30,000" proposal is not the table's $70,000 line at a bargain; it is a different product with the audit and refresh stages deleted. Refresh earns its row on the evidence, too: content under 3 months old was about 3x more likely to be cited in ChatGPT answers in Kevin Indig's 2026 analysis (Growth Memo, 2026) — a correlation, not a law, but the wrong side of it is a decaying asset.

What does the platform actually cost?

Almost nothing recurring, if the architecture is honest. Our fleet runs static site engines — MDX files in git, statically generated pages, no CMS subscription, no database — and the hosting line across 3 production brands is $0 a month on the canonical stack, plus one build paying $20 a month for a server-feature variant it uses for AI site search [our data].

The real platform cost is the one-time engine build: the frontmatter contract that drives every page surface, the schema templates, and the build gates that check citations, links, and freshness on every deploy. That work is days, not months, when the engine already exists — and it is what makes page 200 obey the same rules as page 2 [our data]. Budgets that show a large recurring infrastructure line are usually paying for architecture a content library does not need; Google's guidance cares that pages are crawlable, helpful, and reliable (Google, helpful content guidance), none of which a CMS subscription buys.

How fast can a 100–200 page map actually ship?

Days, on the git-dated record — which matters to the budget because calendar time is where agencies pad and where in-house plans stall. Our greenfield build completed its five-cluster map within 2 days of first commit, with 167 pages counted at audit time; our second build shipped 67 reviewed pages in 2 days on a live-traffic domain and reached 298 content files in 14 days; our third holds 106 files against a 102-page architecture [our data]. The full field report, including what broke at that speed, is the 100-pages-in-a-week scenario.

The budget lesson is that velocity came from the one-time lines — the contract and the gates — not from extra production spend. The same hours get worked either way; what the engine changes is how many of them are spent re-deciding things the contract already decided. A slow build costs more without producing more.

Where do library budgets go wrong?

Four patterns account for most of the waste we see. Paying per page while skipping the engine, which ships 200 pages that disagree with each other. Design and tooling before content — custom CMS work and paid dashboards purchased before a single page exists to manage or measure. Skipping the research line because it feels like delay, then spending the entire production budget discovering the niche was unwinnable — the most expensive way to learn the cheapest fact. And volume past the audit budget: if funding 200 audited pages is not possible, the right move is 100 audited pages, not 200 unaudited ones.

That last trade is the one we argue against our own interest: we sell builds, and smaller is often the correct answer. Whether your niche justifies 100 pages, 200, or none is a research question with a cheap first pass — the fit check — and the standards those pages would be held to are documented across the operator's guide. Budget from the table above, and make every quote you receive name which rows it includes.