Use-Cases for New Tech Stacks (Beyond Pure Static)

Moving off plain WordPress isn’t one decision — it’s four different decisions, depending on what the site actually has to do. According to W3Techs, WordPress’s share of the content-management market slipped from 43.2% to 41.9% over six months in 2026, the first sustained decline it has recorded, even as it still powers roughly 59% of CMS-identified sites. That bend in the trendline is the market motion worth studying — not because “WordPress is dying,” but because real companies have already made each of the four moves off it publicly enough to learn from.

This piece surveys documented, named deployments across four stacks — headless WordPress + WooCommerce, Cloudflare Pages and Workers, Astro/Next.js static-JS, and full custom apps — and it’s honest about the fifth answer that keeps coming up: for plenty of businesses today, WordPress is still the right call. What follows is evidence, not a sales pitch. Every result below is drawn from a source the company or its agency published themselves.

Headless WooCommerce: when the store is big enough to earn it

Headless WooCommerce keeps the WordPress and WooCommerce back end a merchant already knows, and bolts on a separate, pre-built front end for speed. It’s the pattern NW eSource is currently evaluating for ATM MegaStore, an internal e-commerce client — that’s the extent of our own stake in this section, named once and left there. The evidence that it works comes from other people’s stores.

Gourmet Basket, a luxury gift-hamper retailer, is the clearest anchor. In the headless rebuild documented by Blaze Commerce, the store kept its full WooCommerce backend and added a headless front end — and the numbers it published are the kind speed advocates rarely get to name out loud.

Page loads hit 0.223 seconds at peak, average order value rose 30% within days, and the investment was recouped in roughly 60 days — on a store the agency describes as well above $3 million a year in revenue. — Blaze Commerce, Gourmet Basket case study

Bear & Bear, built by the agency Byte9 on a comparable stack (a Blaze front end over a WooCommerce backend on AWS infrastructure), reported a 47% reduction in page-load time, a 78% increase in page views, a 37% increase in transactions, and a 60% year-over-year revenue lift in its published case study. Two different stores, two different agencies, the same shape of result.

It’s also not a bespoke, build-from-scratch pattern anymore. Production-ready templates exist to prove it’s repeatable: the open-source w3bdesign/nextjs-woocommerce (Next.js + TypeScript against the WooCommerce REST API, with Algolia search) and Railway’s official one-click next-woo template, which deploys a WordPress + WooCommerce backend and a Next.js storefront together as a single stack.

Here’s the honest caveat this section needs, and it comes from the same people who build these stores. In Blaze Commerce’s own piece, “When NOT to Go Headless with WooCommerce,” the stated heuristic is that headless pencils out somewhere around $2–5 million a year in revenue, commonly cited near $3 million — because the conversion lift from speed has to outrun a $50–70K-plus rebuild inside roughly 12 to 24 months. Below that line, plain WooCommerce, or even plain WordPress, usually still wins. If your store isn’t near that revenue band, this section is a preview of a later decision, not this year’s.

Cloudflare Pages and Workers: cheap, fast, and almost nothing to attack

The second decision is for the sites that don’t need a store — or a CMS — at all. Cloudflare Pages and Workers host static files and edge functions on Cloudflare’s network, with a generous free tier and effectively no server to break into.

The clearest named, vendor-verified production example is bigger than a small-business site, and it’s worth being upfront about that. Cybernews, a cybersecurity news publisher serving around 150,000 daily readers — spiking to millions on viral stories — runs its site and edge functions on Cloudflare Pages and Workers, explicitly eliminating self-hosted servers, per Cloudflare’s own case study. Cybernews is a mid-size publisher, not a corner-shop marketing site. Use it for what it proves: the platform is real and battle-tested at scale.

For the smaller-site case this section is actually arguing for, there’s no single famous small-business brand to name — so rather than invent one, the honest move is to point at the pattern, which is well documented by the practitioners running it. It looks like this: keep authoring content in the familiar WordPress editor, fire a webhook on publish to trigger a static rebuild (often with Astro), and deploy the result to Cloudflare Pages. The client never leaves the editor they know; the public site is served entirely from Cloudflare’s edge. Write-ups of exactly this setup include Dragos Roua’s “Moving From WordPress to Cloudflare Static Pages” and Easton’s static-blog-to-Cloudflare-Pages deploy guide.

The honest tradeoff: Pages and Workers win on cost and on attack surface for sites that don’t need a CMS of their own — but the platform is not, by itself, a content-editing solution. Pair it with a deliberate, client-directed editing layer — the control spectrum covered in Should You Retire Your WordPress Site? — for anything the owner needs to change themselves. That’s where hosting that fits the stack, not the other way around stops being a slogan and starts being a line item that shrinks.

Astro and Next.js: static-JS, and yes, at real scale

The third decision is for content sites — marketing sites, documentation, publications — that rarely change per-visit and don’t need logins. Static-JS frameworks like Astro and Next.js compile pages ahead of time and ship little or no JavaScript by default. The named examples here aren’t scrappy startups; they’re some of the largest software companies on earth.

Firebase (Google) migrated its blog from Blogger to Astro and, in its case study, reported build times dropping about 75% (from 6 minutes to 1.5 via GitHub Actions), Lighthouse performance scores up 71%, and publishing time falling from hours to minutes.

Microsoft built the 200-plus-page Fluent 2 Design System website in Astro over ten months, citing zero-JS-by-default output for DDoS resistance and Azure scalability, plus the ability to keep its existing React component library. Per Microsoft’s case study, pages now ship in “half the time” versus the prior stack, with simple pages taking 20–30 minutes through a Figma-to-Markdown workflow.

And the beat that’s hard to argue with:

WP Engine — a WordPress hosting company — rebuilt its own Atlas documentation site in Astro instead of WordPress. Development costs dropped over 50%, content operations more than doubled in speed, and the team’s time split flipped from roughly 70% maintenance / 30% content to 95% content / 5% maintenance. — WP Engine, Atlas docs case study

That last one isn’t editorializing; it’s WP Engine’s own published result. When a company that hosts WordPress for a living picks a static framework for its own docs, the use-case has stopped being theoretical. And the same static-generation approach that scales to Microsoft is exactly what makes it a strong fit for a local-service or marketing site: the same near-zero attack surface and sub-second loads, at any scale.

Full custom app or portal: when the site is the software

The fourth “stack” isn’t really a website decision at all, and naming that early saves a lot of misscoped projects. Client portals, dashboards, booking and scheduling systems, anything with per-user logins and real application logic — those are software builds (typically Next.js or React against a real backend), not marketing-site rebuilds. There’s no honest “case study” to drop in here without manufacturing one, so the framing is the point: if the site is the software, price it as a product build, not a website. Scope it that way from the start and the client neither over-builds a brochure nor under-builds an application.

When WordPress still wins

None of the above is an argument that everyone should leave WordPress. There are real cases where it remains the right tool, and a survey that won’t say so isn’t a survey — it’s a pitch. WordPress still wins when:

• You depend on a specific plugin ecosystem with no practical headless equivalent, and rebuilding that functionality would cost more than it returns.
• Non-technical staff need to self-publish constantly through a familiar editor, and the migration cost and timeline don’t pencil out yet.
• The site already runs on a well-maintained, actively-patched WordPress install with a competent host — here the risk is maintenance discipline, not the platform itself.
• The business is budget-constrained and a full migration isn’t this year’s highest-ROI spend.
• You run a store under roughly the $2–5M/year band, where a headless rebuild’s payback period stretches past what the numbers justify (per the Blaze Commerce heuristic above).

The decision framework

Strip away the brand names and the four decisions reduce to a short path most owners can walk in under a minute:

1. Do you sell products online with real inventory, variant, and checkout complexity — and is the store doing roughly $3M+/year? → Headless WooCommerce is worth pricing out.
2. Do you sell online but below that revenue line, or with simple checkout needs? → Plain WooCommerce or a simpler cart usually wins.
3. Do you just need pages that rarely change, with no logins? → Static (Astro / Next.js).
4. Do you need a handful of simple supporting pages, fast and cheap, with no CMS at all? → Cloudflare Pages.
5. Do users log in, and is the site itself the product? → A custom app or portal — a software build, scoped as one.

The right answer is whichever line describes your business, not whichever stack an agency happens to specialize in.

What our own scan of the market shows

The public case studies say what’s possible. Our own data says what’s already common. The NW eSource 2026 Small-Business Website Audit reviewed 1,800 U.S. small-business websites across more than ten verticals; 1,709 were reachable and scored. Of those reachable sites, 34% run WordPress — still the largest single platform, but no longer the default. Of the 66% that are not WordPress, 55% run custom or agency-built stacks, 30% run hosted site builders (Squarespace, Wix, and the like), and 13% run modern static or JavaScript-driven stacks.

Two things worth sitting with. First, “everyone uses WordPress” is already false — a clear supermajority of reachable small-business sites run on something else. Second, the same audit found the softer reason owners start looking in the first place: among WordPress business sites with a blog, 48% hadn’t published in 12 months or more (a figure that held at 47% and 49% across two independent samples), the median time since the last post was 11.5 months, and 63% weren’t adding new content at all. The move off WordPress is rarely a dramatic decision. More often it’s what happens when the old system quietly stopped being used and the next rebuild simply picks a different foundation.

NW eSource — web design We don’t sell one stack to every client. See our web design process for how we match the foundation to the business.

The Agency Lens

Agency Lens We’ve watched Gourmet Basket, Bear & Bear, Firebase, Microsoft, and WP Engine each pick a different stack and win with it for a different reason — and that’s the whole lesson. There is no single “correct” WordPress alternative, only a correct match between what a site has to do and what a stack is good at. Our job on every engagement is to make that match honestly, including the times it points back to WordPress. We scope the stack to the business, never the business to the stack we already know how to build.

Frequently asked questions

Is WordPress dying?

No. Its CMS market share slipped from 43.2% to 41.9% over six months in 2026 per W3Techs — the first sustained decline on record — but it still powers roughly 59% of CMS-identified sites. For a growing set of use-cases, though, real companies are publicly documenting faster, cheaper alternatives.

Do real companies actually use headless WooCommerce, or is it just an agency pitch?

Real companies do. Gourmet Basket and Bear & Bear both published results — a 0.223-second peak load time and a 60% year-over-year revenue lift, respectively — from real headless-WooCommerce rebuilds, documented by the agencies that built them.

How much revenue do I need before a headless WooCommerce rebuild pays off?

The commonly-cited working threshold, per the agencies doing this work, is roughly $2–5 million a year in store revenue, often quoted near $3 million. Below that, the conversion lift from speed usually can’t outrun the cost of the rebuild within a reasonable payback window.

Is Astro or Next.js just for tech companies, or does it work for a normal business site?

Both. Microsoft and Firebase used it for large sites, but the same static-generation approach is what makes it a strong fit for local-service and marketing sites: the same near-zero attack surface and sub-second loads, at any scale.

What if my site needs logins or is itself the product?

That’s a different build category — a custom application, not a website-stack decision — and it should be scoped as one from the start so you don’t over- or under-build.

The lesson under all five decisions is the same: there’s no universal WordPress replacement, only the right foundation for what your site actually has to do. If you’re weighing a rebuild — a store chasing speed, a marketing site chasing simplicity, or a portal that’s really a product — talk to NW eSource and we’ll help you match the stack to the business before a line of it gets built.

Written by Claude and the team at NW eSource — a Portland AI consulting and web design firm.