Insights/Technology & Analytics/CodeCraft™

Websites and digital products that carry the strategy through to code

ORVO Editorial29 April 20269 min read
The short answer

Most website projects fail on content and ownership, not design. Fix the message architecture before layout, treat performance and accessibility as requirements rather than optimisations, build for search and answer visibility from the first page, and name the person who owns the site after launch. Design taste is necessary and rarely the constraint.

What is a content-first website process?

A content-first process establishes the message architecture — what each page must say, to whom, in what order, with what proof — before visual design begins. Layout is then designed to carry real content rather than placeholder text, which prevents the late-stage redesigns caused by content that does not fit.

Who this is for
Companies whose website contradicts the positioning they have just paid for
Businesses with slow, unmaintainable sites accumulated over years
Teams planning a CMS migration or a first product interface
Organisations needing a site that stands up to investor and enterprise scrutiny
Key takeaways
+Write the pages before designing them. Placeholder copy hides every structural problem.
+Performance is a business requirement: slow pages lose revenue and rankings simultaneously.
+Accessibility is a legal and commercial baseline, not a phase-two enhancement.
+Build SEO and answer-engine foundations into the templates, not into a post-launch checklist.
+Name a site owner before launch, with a budget and a monthly rhythm, or the site decays.

Message architecture comes before layout

Website projects that begin with wireframes usually end with a content emergency. Three weeks before launch, somebody discovers that the eight-word headline slot cannot hold the actual proposition, that the case studies do not exist, and that the product pages need technical detail nobody has written.

Content-first inverts that. Each page gets a short specification: the audience, the single job of the page, the argument in order, the proof required, and the one action it should produce. That specification is written and approved before any visual work starts.

The design brief then becomes concrete and much easier to execute well, because the designer is arranging real material with real lengths and real hierarchy rather than guessing at it.

Treat performance and accessibility as requirements

Both get labelled as technical concerns and postponed, and both have direct commercial consequences.

On performance, the discipline is budgetary: a page weight and interaction-latency ceiling agreed up front, images and fonts controlled, third-party scripts individually justified. Most heavy sites are heavy because of accumulated tags nobody owns, not because of the design.

On accessibility, the baseline is colour contrast, keyboard navigation, semantic structure, real focus states, labelled forms and meaningful alternative text. Meeting it widens the addressable audience, improves machine readability, and removes legal exposure. Retrofitting it costs several times more than building it in.

Build search and answer visibility into the templates

Search foundations are structural, which means they are cheap at build time and expensive later. What belongs in the templates from day one:

·A URL and information architecture that reflects real topics, with no orphan pages.
·One indexable page per meaningful intent, rather than a single page trying to rank for a dozen.
·Structured data — Organisation, Article, FAQPage, Product, BreadcrumbList — generated from content, not hand-maintained.
·Question-shaped headings with self-contained answers, so pages can be quoted by answer engines.
·Server-rendered content: anything requiring client-side execution to be readable is a retrieval risk.
·Clean internal linking that expresses topical relationships rather than navigation convenience.

Scope discipline and sequencing

The single biggest predictor of a bad build is an undisciplined scope: every stakeholder's requirement admitted, no phase boundary, launch date defended by cutting quality rather than scope.

The alternative is a defined first release — the pages and functions that carry the commercial argument — followed by planned increments. That keeps the launch date honest, gets real user behaviour into the roadmap, and prevents the pattern where six months of work goes live untested.

Choosing the platform follows the same logic: pick the simplest system that meets the first release and the next twelve months of publishing needs. Headless architectures are powerful and impose a permanent engineering dependency; a well-configured conventional CMS is the right answer more often than vendors admit.

Ownership after launch

A website is an operating asset. It needs a named owner, a monthly rhythm — performance, error monitoring, content freshness, search visibility, conversion review — and a small standing budget for iteration.

Without that, decay is measurable within two quarters: dead links, stale claims, forms that quietly stopped delivering, and a content set that no longer matches the sales story. Every project we see rebuilt from scratch was originally a good site with no owner.

Design-first build vs. content-first build

Starting artefact
Wireframes and moodboards
Page specifications and real copy
Late-stage risk
Content does not fit the design
Layout already tested with real content
Search foundations
Post-launch checklist
Built into templates
After launch
Unowned, decays
Named owner, monthly rhythm, iteration budget

Frequently asked questions

How long does a corporate website build take?

A focused build with content-first discipline typically runs ten to sixteen weeks: two to three for architecture and specifications, three to four for content and design, four to six for build and QA, then launch. Content readiness is almost always the critical path.

Should we build on a headless CMS?

Only if you have engineering capacity to maintain it and genuine multi-channel content needs. For a single marketing site with a small team, a conventional CMS with good templates ships faster and costs far less to run.

How do we keep the site fast over time?

Set a performance budget, review it monthly, and require justification for every added third-party script. Degradation is almost always caused by accumulated tags and unoptimised media rather than by the original build.

Does accessibility affect SEO?

Indirectly and substantially. Semantic structure, meaningful text alternatives, clear headings and keyboard-navigable interfaces all make content easier for machines to parse, which supports both traditional ranking and citation by answer engines.

CodeCraft™ — Technology & Analytics
Want this run inside your business?

A short diagnostic conversation is usually enough to tell you whether there is a real opportunity here — and what it would take.

Book a conversation Explore Technology & Analytics
Related reading
MarTechMind™
Martech that earns its licence fee: stack design, automation and AI-led operations
9 min read
InsightEdge™
Marketing measurement that survives scrutiny: ROI, mix modelling and useful dashboards
9 min read
PlatformOne™
Building proprietary products inside a consulting practice
8 min read