The technical checklist before you launch a landing page

The message is good, the social proof is in place, the call to action is impossible to miss — and the page still fails on launch day. Not because of the content: because of what happens behind it, invisible to the eye, and that nobody checked before opening the traffic. A form that submits nowhere, a conversion pixel that was never installed, a consent banner blocking measurement, a redirect left over from the old page: these are technical failures, not copywriting failures, and they're fixed with a list, not with talent. This article covers the technical side of a launch — tracking, forms, GDPR, speed, what Google sees — as a companion to our 25-point CRO checklist, which covers structure, message and proof.

Why a list beats memory, even for an experienced team

The idea that a checklist reduces errors isn't specific to marketing. A study by Alex B. Haynes and coauthors (including Atul Gawande), published in 2009 in the New England Journal of Medicine under the title "A Surgical Safety Checklist to Reduce Morbidity and Mortality in a Global Population," tracked the introduction of a surgical safety checklist across eight hospitals on four continents: the rate of major complications dropped from about 11% to 7%, and mortality from about 1.5% to 0.8% (Haynes et al., 2009). The gap wasn't a lack of skill on the teams' part: it came from steps everyone already knew, but that a heavy workload makes people skip roughly one time in ten. Publishing a landing page obviously doesn't carry the same stakes, but the mechanism is identical — under deadline pressure, a competent team forgets the same two or three things, always the same ones, unless something forces it to check them one by one instead of keeping them in mind.

Tracking: if nobody checks, nothing comes through

• The GA4 tag (or whichever tool you use) sits on the actual published page, not just the staging environment — verified in real time, not just read in the source code.
• The Meta pixel and the Google Ads conversion fire on the action that matters (form submission), not on the mere display of the thank-you page.
• The tag manager points to the production container, not the test workspace used while building the page.
• Consent mode is active and the cookie banner doesn't block baseline measurement — see our article on Consent Mode v2.
• The UTM parameters for the campaigns that will drive traffic to the page are defined and consistent before the first euro is spent — see our guide to UTM tracking.

The form, and what happens after it's submitted

• Submit the form yourself, end to end, from a device that isn't the one it was built on.
• Check that the lead actually reaches the destination tool (CRM, shared inbox, spreadsheet) and not just the server logs — see our guide to connecting a form to a CRM.
• Check what shows up after submission: a badly configured thank-you page is also the moment the conversion should fire on the measurement side.
• Test at least one invalid submission (an empty required field, a malformed email) to confirm the error messages display correctly — our form guide covers the details.
• Make sure no test email address or personal account is still wired in as the sole recipient of leads.

GDPR and legal notices: what gets an ad campaign rejected

• The consent banner displays, offers a refusal option as easy as acceptance, and the form's checkbox isn't pre-ticked — see our article on GDPR-compliant forms.
• The legal notice and, depending on the offer, the terms of sale are reachable from the footer and up to date — see what's legally required.
• The purpose of the data collection and how long it's kept are stated, even briefly, near the form.
• A regulated sector (health, finance, professional training) shows the disclosures specific to its activity, not just the generic notices of a brochure site.

Speed, mobile, several browsers: test on real devices

Checking a page yourself, on your own computer and your own connection, is never enough: it's the best-configured device and the least representative of your real traffic. Yet the number of people needed to catch most problems stays surprisingly small. Jakob Nielsen and Thomas K. Landauer published a mathematical model of usability problem discovery in 1993, in the proceedings of the ACM CHI conference; Nielsen later drew a conclusion that became a classic of the field: a small group of testers — around five people — is enough to catch roughly 85% of an interface's usability problems (Nielsen & Landauer, 1993). Applied to a landing page about to go live: asking four or five colleagues to browse it, each on their own phone and their own browser, before opening the traffic, catches nearly every display bug that a single tester, however careful, will never see on their own machine.

• Open the page on at least one Android phone and one iPhone, not just the browser's responsive mode.
• Check the display on Chrome, Safari, and at least one third browser (Firefox, Edge).
• Measure real mobile speed, not just the office's fiber connection — see our article on load speed and Core Web Vitals.
• Click every link on the page, including the footer, to catch 404s and broken anchors.
• Make sure no redirect loop or link to the old version of the page is still lingering.

What Google needs to see on the first crawl

• The HTTPS certificate is active on the final domain, not just the test environment — see our article on HTTPS and trust.
• The exclusion tag set during staging has actually been removed before opening traffic — unless the page is a deliberately non-indexed ad landing page, in which case it's the opposite you need to verify.
• The title, meta description and share image are customized for this specific page, not inherited from a generic template — see our guide to the Open Graph image.
• Any structured data matches exactly what the page actually displays — see our article on structured data.
• The page appears in the sitemap if it's meant to be indexed.

The 48 hours after launch

1. Send one real test lead within the hour after opening traffic, and confirm it reaches its destination.
2. Check the analytics tool's real-time reports to confirm conversion events are coming through, not just pageviews.
3. Check the ad status if a campaign points to the page (a technical disapproval from Google Ads or Meta).
4. Watch for 404 errors and server response times during the first hours of real traffic.
5. Don't make any structural change in the first 48 hours: a real change should be tested, not decided on the fly after a handful of visits.

None of this makes a page more persuasive: it only stops a persuasive page from losing leads for a reason it should never have run into. For the part that actually drives conversion — structure, message, form, social proof —, our 25-point CRO checklist is available for free by email. And to start from a base that has already passed both checklists before entering the catalogue, our 10 LanderKit templates are sold at €89 each, or €229 for the full bundle, each with a live demo you can browse before buying on its demo page.