When a Custom Website Is Worth the Investment — and When It Isn’t

The four thresholds

1. The site is a primary acquisition channel. Not “has a website” — the pipeline runs through it. When organic traffic, conversion rate, and page performance connect directly to revenue, small controlled improvements compound into infrastructure-grade returns. The canonical evidence for the mechanism: Deloitte Digital & Google, “Milliseconds Make Millions” (2020 study, 37 brands, 30M mobile sessions) measured ~+8% retail conversions from a 0.1-second mobile-speed improvement in its cohort — directional, pre-CWV-era, and still the clearest demonstration that web performance is a revenue variable, not an aesthetic one. Infrastructure that carries revenue justifies engineering; brochures don’t.

2. The roadmap contains software. Portals, quoting engines, calculators, integrations as features, member areas — application-shaped items that platforms host awkwardly through plugin lattices and external services. Count your two-year roadmap: when the software items outnumber the content items, forcing them through a content platform charges a permanent friction tax that engineering removes.

3. Structured scale is the strategy. Programmatic SEO architectures — hundreds or thousands of genuine, data-driven pages with exact schema — are custom territory by elimination: builders resist dozens of pages, platform CMS caps bound thousands, and code generates them consistently. If search-as-a-system is the growth model, the architecture is the product.

4. Performance has commercial consequences you must control. The field data (HTTP Archive's Core Web Vitals Technology Report, November 2025 snapshot) shows platform ceilings are real but survivable; the threshold case is needing the ceiling — and the floor — as enforceable properties. That means performance budgets in code review, not hopes about the next theme update.

The not-worth-it list, stated plainly

This list is roughly half the businesses that ask us for custom quotes. We tell them the same thing this page does, because a client who buys the wrong thing from us becomes a case study for a competitor.

Sequencing and procurement: crossing the threshold well

Sequence is legitimate strategy. Platform-now-custom-later is how many businesses correctly cross: validate on a builder, grow on WordPress or Webflow, engineer when the thresholds arrive — migrating each time with a complete redirect map so SEO equity survives. The failure mode is drift: staying two years past the threshold because migration feels scary, paying the workaround tax annually.

Procure against capture. When you do buy custom, the contract terms matter more than the pitch: code assigned to you, standard technology (the hiring math: React at 46.9% of developers, Stack Overflow Developer Survey, 2025 edition), ordinary hosting, and documentation a successor shop could pick up. The custom market’s real risk isn’t overpaying — it’s a build only its builder can maintain. Two questions filter it: could a thousand other shops take this over? and does the contract say it’s ours?

Keep going