Launch pages in hours, not sprints

A partner signs on Thursday. The joint announcement goes out Tuesday morning, and it needs a landing page — logo lockup, what the partnership means for customers, a proof point, a form. On most startup marketing sites, this is where the trouble starts: the page is a dev ticket, the ticket goes into a sprint, and the sprint ends after the news cycle does. Or it’s an agency request, which means a quote, a kickoff, and two to three weeks. The announcement ships as a blog post with no call to action, and everyone quietly agrees the website is a bottleneck.

We think marketing velocity is a website property. Not a team property, not a heroics property — a consequence of how the site is built. Built one way, a landing page is an afternoon. Built the other way, it’s a project.

Pages as composition, not construction

The architecture that makes the difference: every page is a typed document — an ordered list of sections drawn from a library the site already has. Hero, feature grid, pricing table, FAQ, testimonial and proof blocks, call-to-action band. Each section has a schema: these fields, these shapes, these constraints. Building the partnership page means choosing six sections and writing the copy that fills them. No new components, no new CSS, no deploy.

That reframes what a landing page is. It stops being engineering work — there is no code in it — and becomes what it always should have been: an editorial act. The marketer who knows what the page should say is the person who ships it. Realistic time from “we need a page” to “the page is live” is two to four hours, and most of that is writing copy, which you’d have to do anyway.

Why fast doesn’t mean sloppy

Speed without guardrails just produces broken pages faster. Three properties make composition safe at speed:

• Schema validation. Every page is checked against the section schemas before it ships. A pricing table missing a tier name, a hero pointing at a dead image, a FAQ item with no answer — rejected at save time, not discovered by a prospect. A malformed page structurally cannot go live.
• Versioning. Every publish is an immutable revision. That kills the QA paralysis that makes teams slow: you don’t need a three-person review to ship a page when reverting it is one click. The cost of a mistake drops from “incident” to “undo,” and shipping speed follows the cost of mistakes.
• Design tokens. Sections consume the brand’s colors, type scale, and spacing from one source. There is no way to compose an off-brand page, because the section library is the brand. Nobody reviews for visual consistency, because consistency isn’t a property of the page — it’s a property of the system.

The pages you’d actually ship this quarter

Concretely, the pages this makes cheap are the ones growth teams keep not shipping:

• The partnership announcement that has to be live while the news is news. The window is measured in days; a sprint doesn’t fit inside it.
• The pricing experiment. One campaign’s traffic goes to a variant page with different packaging or an annual-only offer. It runs for three weeks and disappears. Nobody builds this if the page costs $5,000 — the experiment isn’t worth it. At two hours, it obviously is.
• Vertical and localized variants. The services page recomposed for dentists, then for law firms; the same page in German for the DACH campaign. Same sections, different copy and proof — ten variants become an editorial batch job instead of ten tickets.

Why the agency quote is thousands — and when you don’t need this

The classic quote for “one landing page” — several thousand dollars, two to three weeks — isn’t gouging. On a blob-architecture site (page builders, hand-built templates, WordPress plus plugins), every new page really is bespoke: custom markup, custom responsive behavior, a real chance of breaking something adjacent, a round of cross-browser QA. The price reflects genuine risk. Composition doesn’t make agencies cheaper; it removes the reason the work was expensive.

The honest caveat: this only pays if you actually run campaigns. If your site changes quarterly — a copy tweak, a new team member, an annual pricing update — you don’t need composition speed, you need reliable maintenance, and investing in a section library would be over-engineering. The test is a number: count the pages marketing wanted in the last six months against the pages that shipped. If the gap is zero, keep what you have. If marketing has stopped asking because they already know the answer, the architecture is the bottleneck — and no amount of process fixes an architecture problem.

The sites we build and manage work this way — a typed section library, validated pages, one-click reverts — and the retainer is what keeps a two-hour landing page an actual two hours, month after month.