Playbook
WordPress vs. Static Site for an Insurance Agency Website
We measured 1,080 insurance agency websites and fingerprinted every site that reached the top speed tier. The pattern was too lopsided to ignore — and still leaves honest room for WordPress.
In our August 2026 study of 1,080 insurance agency homepages, the median mobile PageSpeed score was 61 — and of the 12 substantive sites that reached 99+, ten were static builds; only two ran WordPress, one of those behind NitroPack, an optimization layer that effectively staticizes pages1. That's the whole argument in one sentence: WordPress can be made fast, but static architecture is fast by default, and in the field the default wins. WordPress still earns its place on editing experience and ecosystem — this article says exactly where.
The finding this article hangs on
Platform debates usually run on vibes, so we'll run this one on a measurement we took ourselves. On August 18, 2026 we ran Google's PageSpeed Insights API against 1,080 insurance agency homepages and published the aggregate dataset CC01. Three numbers set the stage: the median mobile performance score was 61; the middle half of the industry sat between 52 and 70; and the median site took 7.9 seconds to paint its largest element on a mobile connection, against Google's 2.5-second "good" threshold for LCP2.
Then we fingerprinted the top tier. Among 1,046 substantive agency sites, exactly 12 — 1.15% — scored 99+. Ten of the twelve were static: hand-built pages or website-builder static export. Two ran WordPress, and one of those two reached the tier via NitroPack, an optimization service that effectively serves a staticized version of each page. That distribution — ten native-static, one staticized-WordPress, one disciplined-WordPress — is the spine of everything below.
What WordPress actually is (and isn't)
WordPress is a dynamic content management system: pages live in a database, and PHP assembles the HTML when a visitor (or crawler) requests it. Its genuine strengths follow from that design. The editing experience is mature — any licensed agent can log in and fix a phone number. The ecosystem is unmatched: themes, form builders, SEO plugins, membership tools, someone-already-built-it for nearly anything. And the talent pool is enormous; you will never struggle to find someone who "does WordPress."
What WordPress isn't is inherently slow — and our own data proves that with the two 99+ exceptions. The honest framing is statistical, not categorical: the typical WordPress install, as actually configured by the typical agency's vendor, stacks a heavyweight theme, a page builder, and a dozen plugins, and that stack is what lands sites in the median-61 band. Why the typical install ends up there has a specific mechanism, not a mystery.
Why the typical install ends up slow
Each layer of the standard agency WordPress build adds render-blocking work to every page. The theme ships its own CSS and JavaScript bundles sized for every feature the theme supports, not the ones your site uses. The page builder (Elementor, Divi, WPBakery) adds its runtime on top — markup wrappers, builder CSS, builder JS. Then plugins: a form plugin, a slider, a chat widget, an SEO plugin, an analytics injector — many of which enqueue scripts site-wide even on pages that never use them. None of these choices is individually fatal; the sum is a phone parsing hundreds of kilobytes of JavaScript before the largest element paints — the exact cost the mobile PageSpeed score prices in.
What a static site actually is
A static site does the assembly once, at build time. Every page is rendered to plain HTML before deployment, served from a CDN, with JavaScript added only where a specific page earns it (a quoting widget, a calculator). There's no database query and no PHP execution at request time — which is also why there's no WordPress-style patching treadmill on the public surface. Modern static frameworks (ours is Astro) generate the whole machine layer too: full JSON-LD schema on every page, a generated llms.txt, and agent-callable WebMCP endpoints — the parts an AEO audit scores that a plugin stack controls only partially.
The honest cost: editing is developer-gated or pipeline-driven. There's no admin login where a non-technical office manager rewrites a page on a Saturday. For us that trade works because content ships through a generator; for some agencies it genuinely doesn't — which is the next section's point.
The comparison, factually
| Dimension | WordPress (typical install) | Static build |
|---|---|---|
| Rendering | PHP + database assemble each page per request | Pre-rendered HTML at build time, served from CDN |
| Measured field results | 2 of the 12 sites at 99+ in our 1,080-site sample; one via NitroPack1 | 10 of the 12 sites at 99+1; our own client builds measure 99–100 from clean builds |
| Speed ceiling | Reachable with discipline or an optimization layer that staticizes output | Fast by default; effort goes to not regressing |
| Editing | Mature in-browser editor, non-technical staff friendly — its strongest card | Developer- or pipeline-gated; headless/git CMS closes some of the gap |
| Ecosystem | Enormous: themes, plugins, integrations, talent pool | Smaller; custom work replaces plugins |
| Maintenance & security | Core, theme, and every plugin need patching; server-side execution is an attack surface | No request-time execution on the public site; near-zero patching duty |
| Schema / llms.txt / WebMCP | Schema via plugins (partial control); root files possible; agent-callable endpoints require custom server work | Full control generated per build: JSON-LD everywhere, llms.txt, live /mcp endpoints3 |
| Cost structure | Hosting + premium theme/plugin licenses + maintenance time or retainer | Build cost upfront; hosting roughly free at agency scale; no license stack |
The two WordPress sites that made 99+
The exceptions are worth dwelling on, because they keep this article honest in both directions. One of the two WordPress sites in the 99+ tier got there behind NitroPack — which is less a counterexample than a concession: the optimization layer wins by serving pre-optimized, effectively static output in front of the dynamic stack. The other got there the hard way, which proves disciplined WordPress can compete at the top. What the dataset says is not "WordPress can't" — it's that in a field of over a thousand real agency sites, almost nobody's WordPress vendor actually does1. Betting your agency's web presence on being the rare exception is a choice you should at least make knowingly.
When WordPress is genuinely fine
Three cases, stated without hedging. First: your team edits the site constantly and non-technical editing is non-negotiable — WordPress's editor is still the best-in-class experience, and a lean build (no page builder, minimal plugins, or a NitroPack-style layer) can be made fast enough. Second: you're deep in a WordPress-native ecosystem — a membership area, an existing content library, integrations that would cost real money to rebuild — and the migration cost outweighs a speed-tier gain. Third: your growth engine isn't organic — if referrals and paid leads fill the book and the site is a business card, the median-61 band costs you little. In all three cases the right move is the FAQ's "measure, then subtract" path, not a rebuild.
What we build, and why
Our position follows from the dataset rather than preceding it: we build static (Astro, pre-rendered, CDN-served) because the top speed tier of our own industry sample is where our client builds have to measure — 99–100 mobile from clean builds — and because the AI-search layer (the AEO half of the job) rewards exactly what static generation makes free: complete schema, generated llms.txt, callable agent endpoints, and pages a crawler parses without executing anything. When a client is better served staying on WordPress, the honest advice is the tune-up path above — and we've given it.
Questions agencies ask
Is WordPress bad for SEO?
No — WordPress powers an enormous share of well-ranking sites, and its ecosystem (Yoast, Rank Math, sitemap tooling) makes on-page SEO hygiene easy. The problem our data shows isn't SEO capability, it's the performance ceiling of the typical theme-plus-page-builder-plus-plugins install: a median-61 industry where the top speed tier is almost entirely static. Speed is one ranking input among many, and it's the one the platform choice most directly determines.
Can WordPress reach 99+ on mobile PageSpeed?
Yes — two of the twelve 99+ sites in our 1,080-site study ran WordPress, and one of them got there behind NitroPack, an optimization layer that effectively serves a staticized version of each page. So it's achievable, but in this sample it required either unusual discipline (lean theme, few plugins) or a caching/optimization product doing the staticizing that a static architecture does natively. Reaching 99+ on a dynamic stack without such a layer was, in this dataset, rare to the point of near-absence.
Don't I need WordPress to edit my own site?
That used to be the trade, and it's the most legitimate reason to choose WordPress. It's also narrowing: static sites can pair with git-based or headless CMS editing, and on our own builds the content pipeline (blog posts, answer pages, location pages) is generated and published without anyone touching code. If daily hands-on editing by non-technical staff is a hard requirement, WordPress's editor is still the more mature experience — that's an honest point in its favor.
What about WordPress security and maintenance?
A dynamic CMS runs PHP and a database on every request, so it carries a patching duty: core, theme, and every plugin are code that must be kept current. A static site has no server-side execution at request time — there's nothing to inject into and no admin login on the public site — which is why the maintenance line in our comparison reads the way it does. This is a real, structural difference, not fear marketing; a well-maintained WordPress site can be run safely, it just has to actually be maintained.
If I stay on WordPress, what should I do first?
Measure, then subtract. Run PageSpeed Insights mobile on your homepage and your highest-traffic service page. If you're in the industry-typical 50-70 band, the wins in order: a lean theme without a page builder, removing plugins that inject scripts site-wide, right-sizing images, and a caching/optimization layer (the NitroPack route one of our 99+ WordPress sites took). Then re-measure — the same-tool before-and-after is the only evidence that matters.
Sources
- Strategic AI Architects. "We Ran PageSpeed on 1,080 Insurance Agency Websites," original research via the Google PageSpeed Insights API v5, August 18, 2026 — median 61, middle half 52–70, median LCP 7.9s, 12 of 1,046 substantive sites at 99+ (ten static; two WordPress, one via NitroPack). Aggregate dataset published CC0. strategicaiarchitects.com.
- web.dev (Google). "Largest Contentful Paint (LCP)" — the 2.5-second good threshold at the 75th percentile. Verified live 2026-08-18. web.dev.
- Strategic AI Architects. Live WebMCP endpoint verification, 2026-08-18: JSON-RPC tools/list calls against bigsiouxlife.com/mcp, ichrasavings.com/mcp, bigsiouxbenefits.com/mcp, and strategicaiarchitects.com/mcp. See What is WebMCP?
Talk it through
Want a second pair of eyes on it?
Free 30 minutes. Bring what you found, or bring nothing and we will look together at how AI engines read your site and which fixes move first.
Find out which band your site is in
Run the free Audit — a live AEO audit plus a HIPAA tracking scan — and see your mobile score the way Google and the AI engines see it.
Related reading: GoHighLevel website vs. custom-built site · the 1,080-site PageSpeed study · why WordPress sites are often slow · what causes layout shift