Can a landing page really be built in 48 hours?

Yes, with two conditions. First, 48 working hours across two to three calendar days, not a weekend during which nobody answers email. Second, the scope is locked in writing before the clock starts: one offer, one audience, one action, a named section list and a named approver. Roughly 34 of the 48 hours are work; about 4 are client review gates and the rest is overlap and slack. Execution was never the slow part — unresolved decisions are.

Is this a case study of a real client?

No, and deliberately so. This is a documented breakdown of our standard process with hours attributed from builds we have run, not an account of one named project. We are not going to invent a customer to make the narrative tidier. Where a figure is an average across builds rather than a single measurement, the article says so.

What most often makes a 48-hour build miss its deadline?

Client review gates, by a wide margin — not the technical work. A two-hour approval window assumes a named person is free, and calendars do not know about your launch. The fixes are procedural: name the approver in the scope lock, book both windows as calendar invitations before the clock starts, insist on one consolidated feedback list rather than three contradictory ones, and agree in writing up front that an unanswered gate counts as approved.

What gets sacrificed at this speed?

Audience research, A/B testing, original photography of a physical product, multiple language versions, long compliance review, and consensus-based approval. The page is built on what you already know. If more than two of those are load-bearing for your launch, buy a two-week build instead — it will be cheaper and better.

Should I write the copy before the design?

Always. Designing first and pouring words in afterwards is the most common reason a landing page has to be rebuilt: the layout gets shaped around a headline length nobody chose, and every real sentence then breaks it. Write a skeleton first — headline, subhead, one sentence per section argument, and the exact CTA wording — then design against it and polish the prose once line breaks are visible. Never use placeholder text on a compressed timeline.

What is the cheapest way to get a good launch page?

Do the scope lock and the copy yourself, then buy only the build. Those two blocks need you most and cost the most to outsource. Below that, a decent template with a locked scope and written copy is about a day of your own time and low-hundreds a year in software — most launches do not fail because the page was templated. Pay for compressed delivery only when a date exists that you do not control.

What has to be ready before the clock starts?

The signed one-page scope lock, a named approver with two review windows booked, existing brand assets at usable resolution with clear rights, access to the domain and DNS, access to the analytics and form destinations, and any legal-approved claim wording. The asset audit at hour 6 exists to catch gaps early; discovering at hour 30 that the only product photo is a 900-pixel JPEG is how a 48-hour build becomes a 96-hour one.

We Built a Launch Landing Page in 48 Hours. Here's Every Hour of It.

By Ryan Cline · August 7, 2026

Published: 7 August 2026 · Last reviewed: 7 August 2026

A launch landing page can be built in 48 hours, and roughly 34 of those hours are work while the rest is waiting on someone. This is an hour-by-hour account of how that fits together. Read the next paragraph before the table, because it matters.

What this article is, precisely. This is a documented breakdown of our standard 48-hour landing page process, with hours attributed to each phase from builds we have run. It is not a case study of one named client, and no client is described here. We are not going to invent a customer to make a story read better. Where a number is an average across builds rather than a single measurement, it says so. Where something routinely goes wrong, it says that too, because the failures are the useful part.

The 48 hours, at a glance

Block	Hours	Phase	Who is working	What exists at the end of it
0–2	2	Scope lock	Producer + client	A written, signed scope. One offer, one audience, one action.
2–6	4	Message and copy skeleton	Writer	Headline, subhead, the five section arguments, the CTA wording.
6–8	2	Asset audit	Producer	A list of what exists, what is missing, what will be substituted.
8–10	2	Client review gate 1	Client	Copy approved, or the whole thing slips. This is the real risk.
10–16	6	Design pass	Designer	Full-page layout, type scale, colour, mobile behaviour decided.
16–22	6	Build	Developer	Responsive page, real content, no placeholders.
22–26	4	Media production	Producer	Hero visual, supporting imagery, any short motion piece.
26–30	4	Plumbing	Developer	Form, destination, notification, consent banner, analytics, events.
30–32	2	Client review gate 2	Client	Staging link reviewed. One consolidated list of changes.
32–38	6	Revision round	Whole team	The agreed list, done. Not a second design direction.
38–42	4	Hardening	Developer	Performance, accessibility, cross-browser, real-device checks.
42–45	3	SEO and metadata	Developer	Title, description, canonical, social card, schema, sitemap entry.
45–47	2	Launch	Developer + producer	DNS, SSL, redirects, form tested from a real phone on cellular.
47–48	1	Handover	Producer	Edit access, tracking documentation, what to watch in week one.
Total	48			~34 hours of work, ~4 hours of client gates, ~10 hours of parallel overlap and slack

Two caveats on that table before anyone plans against it. First, blocks overlap — design and copy are not strictly sequential, which is where the apparent extra hours come from. Second, 48 hours means 48 working hours across two to three calendar days, not two days including a weekend when nobody answers email. Anyone quoting the second thing is quoting a slogan.

Hours 0–2 · Scope lock, and why it is the whole trick

Nothing about a 48-hour build is fast execution. The execution was always this fast. What makes it possible is that the scope is locked before the clock starts, and locked means written down and agreed.

The lock document is one page and answers exactly six things:

One offer. Not a product family. One thing being sold or signed up for.
One audience. If two audiences need different arguments, that is two pages, and two pages is not a 48-hour job.
One action. One primary CTA, repeated. A page with a buy button, a demo request, a newsletter box and a Discord link has no CTA.
Section list. Usually five to seven, named, in order, agreed.
Who approves, by name, and their availability across the two review gates.
What happens to anything new. New ideas after hour 0 go on a list for after launch. This is written down so that saying it later is not a confrontation.

This is the single highest-value part of the process and it is free. If you take one thing from this page and never speak to us, take this: most landing page projects that take six weeks contain about four days of work and five weeks of unresolved decisions. Writing the six answers above before anyone opens a design tool is the entire difference, and you can do it yourself this afternoon.

Hours 2–8 · Copy before design, always

Designing first and pouring words in afterwards is the most common reason a landing page has to be rebuilt. The layout ends up shaped around a headline length nobody chose, and every real sentence then breaks it.

So the writer goes first, and produces a skeleton rather than finished prose: the headline, the subhead, one sentence stating the argument each section must make, and the exact CTA wording. Roughly four hours. Polishing happens later, against the real layout, where line breaks are visible.

The asset audit that follows is unglamorous and saves the project more often than anything else: what images actually exist, at what resolution, with what rights. Discovering at hour 30 that the only product photo is a 900px JPEG pulled off a marketplace listing is how a 48-hour build becomes a 96-hour build. Two hours of looking, at hour 6, is cheap insurance.

Hours 8–10 and 30–32 · The gates, where it actually goes wrong

Here is the honest failure mode, and it is not a technical one. In our experience the overwhelming majority of missed 48-hour deadlines are missed at a client review gate. Not because clients are difficult — because a two-hour approval window assumes a named person is free, and calendars do not know about your launch.

What we do about it, having lost hours to this repeatedly:

The approver is named in the scope lock, with their two windows booked as calendar invitations before the clock starts.
Feedback comes back as one consolidated list. Three people sending contradictory notes separately is not feedback; it is a second project.
A gate that passes without response after its window is treated as approved, and this is agreed up front in writing rather than sprung later.
Gate 2 is a staging link on a real phone, not a screenshot. People approve screenshots and then object to the real thing.

If your organisation cannot guarantee a named approver in two two-hour windows, do not buy a 48-hour build. Buy a two-week one and be happier. That is a genuine recommendation, not a humblebrag.

Hours 10–22 · Design and build

Six hours of design, six of build. The design pass is a full page, mobile decided at the same time rather than after — retrofitting a desktop layout to mobile is where the schedule is usually lost, and most of the traffic is mobile anyway.

The build uses real copy from hour one. No lorem ipsum, ever, on a compressed timeline: placeholder text hides every layout problem you are about to have and then reveals them all at once at hour 34.

What routinely takes longer than planned in this block: the hero section, every time. It is usually a third of the design hours for a tenth of the page. That is correct allocation, not overrun — it is the only section most visitors read — but it should be planned for rather than discovered.

Hours 22–30 · Media and the plumbing nobody budgets

Media production runs in parallel where it can: hero visual, supporting imagery, and a short motion piece if the offer needs one.

Then four hours of plumbing, which is the most underestimated block on the page. It is not just a form:

Form submission with validation and a visible error state that a real person can act on.
Where the submission actually goes, and who is notified, and a test proving both.
Consent banner and analytics that respect the answer.
Conversion events defined and firing, verified in the destination rather than assumed.
A thank-you state that confirms something happened.
Spam handling, because an unprotected launch form fills with junk within a day.

The one that gets cut and should not: testing the form from a real phone on cellular data, logged out, as a stranger would. Desktop-on-office-wifi testing has passed forms that were quietly broken for half the audience.

Hours 32–45 · Revision, hardening, metadata

Six hours for the agreed revision list. To be blunt about what that does and does not cover: it is the list from gate 2. A new creative direction at hour 33 is a new project, priced separately, and saying so is the reason the date holds. Unlimited revisions and a guaranteed date cannot both be true.

Four hours of hardening: image weight and formats, layout stability, real-device checks across at least three phones, keyboard navigation, focus states, colour contrast, and alt text that describes the image rather than naming the file. Accessibility here is four hours; retrofitted later it is days.

Three hours of metadata, which is the block most likely to be skipped under pressure and the one with the longest tail: title and description written for a human, canonical set, social card that renders correctly when pasted into the two places it will actually be pasted, structured data, and the sitemap entry. A launch page that is invisible for its first fortnight has wasted the compressed timeline it was built on.

What gets cut, honestly

A 48-hour build is a real constraint, and pretending nothing is sacrificed would be dishonest. What does not fit:

Audience research. The page is built on what you already know. If you do not know who is buying, 48 hours will produce a well-made page aimed at a guess.
A/B testing. You cannot test a variant of a page that does not exist yet. Testing starts after launch.
Original photography of a physical product. Shipping and shooting does not compress. If you need it, it must already exist.
Multiple language versions. Possible afterwards, not inside the window.
Long legal review. If a compliance team needs five days, that is the timeline; no build process changes it.
Endless consensus. Named approver or nothing.

If more than two of those are load-bearing for you, the compressed timeline is the wrong purchase, and we will say so before taking the work.

What this costs, and the cheaper routes

Compressed delivery carries a premium because it consumes a team's full attention for two days and leaves no room to absorb someone else's overrun. Current prices and delivery dates for this are on the landing pages service page, and the wider market context — DIY through agency — is in our landing page cost guide for 2026.

The cheaper routes are real and worth stating plainly:

Build it yourself on a template. With a locked scope and written copy, a decent template page is a day of your own time and low-hundreds a year in software. If the offer is good, it will convert. Most launches do not fail because the page was templated.
Hire a freelancer on a normal two-week timeline. Materially cheaper than compressed delivery, and the extra fortnight costs you nothing if the launch date is soft.
Do the scope lock and copy yourself, buy only the build. Copy and scope are the two blocks that need you most and cost the most to outsource. Doing them yourself is the biggest single saving available.

Pay for compressed delivery when a date exists that you do not control — an event, a store approval, a funding announcement, a partner's campaign. That is the only good reason, and it is a common one.

Frequently asked questions

Can a landing page really be built in 48 hours?: Yes, with two conditions. First, 48 working hours across two to three calendar days, not a weekend during which nobody answers email. Second, the scope is locked in writing before the clock starts: one offer, one audience, one action, a named section list and a named approver. Roughly 34 of the 48 hours are work; about 4 are client review gates and the rest is overlap and slack. Execution was never the slow part — unresolved decisions are.
Is this a case study of a real client?: No, and deliberately so. This is a documented breakdown of our standard process with hours attributed from builds we have run, not an account of one named project. We are not going to invent a customer to make the narrative tidier. Where a figure is an average across builds rather than a single measurement, the article says so.
What most often makes a 48-hour build miss its deadline?: Client review gates, by a wide margin — not the technical work. A two-hour approval window assumes a named person is free, and calendars do not know about your launch. The fixes are procedural: name the approver in the scope lock, book both windows as calendar invitations before the clock starts, insist on one consolidated feedback list rather than three contradictory ones, and agree in writing up front that an unanswered gate counts as approved.
What gets sacrificed at this speed?: Audience research, A/B testing, original photography of a physical product, multiple language versions, long compliance review, and consensus-based approval. The page is built on what you already know. If more than two of those are load-bearing for your launch, buy a two-week build instead — it will be cheaper and better.
Should I write the copy before the design?: Always. Designing first and pouring words in afterwards is the most common reason a landing page has to be rebuilt: the layout gets shaped around a headline length nobody chose, and every real sentence then breaks it. Write a skeleton first — headline, subhead, one sentence per section argument, and the exact CTA wording — then design against it and polish the prose once line breaks are visible. Never use placeholder text on a compressed timeline.
What is the cheapest way to get a good launch page?: Do the scope lock and the copy yourself, then buy only the build. Those two blocks need you most and cost the most to outsource. Below that, a decent template with a locked scope and written copy is about a day of your own time and low-hundreds a year in software — most launches do not fail because the page was templated. Pay for compressed delivery only when a date exists that you do not control.
What has to be ready before the clock starts?: The signed one-page scope lock, a named approver with two review windows booked, existing brand assets at usable resolution with clear rights, access to the domain and DNS, access to the analytics and form destinations, and any legal-approved claim wording. The asset audit at hour 6 exists to catch gaps early; discovering at hour 30 that the only product photo is a 900-pixel JPEG is how a 48-hour build becomes a 96-hour one.

Hour allocations are averages across builds we have run and are practitioner records, not audited data. This article describes a standard process and does not describe any individual client engagement. Next scheduled review: February 2027.