The launch checklist: what to check before a new website or campaign goes live
The checks that get skipped because the site looks fine to the people who built it: whether search engines are allowed in, whether the forms actually arrive, whether anything is being measured, and whether the rest of the internet points to the new pages. In order, with a printable checklist, and one launch mistake of our own.
Launches fail quietly. The site looks right in the browser, the owner is pleased, the designer moves on, and for weeks nobody notices that Google has been told to stay away, that the contact form sends enquiries nowhere, or that the old pages people had bookmarked now lead to an error.
None of these checks is difficult. They get skipped because a site that looks finished feels finished. Here they are in the order we would run them, with a printable version to tick off.
Our own launch mistake
For at least twelve days in September 2026, Google was told not to crawl oxocreatives.com at all.
The new site's own robots.txt, the small file that tells crawlers where they may go, was fine: it has allowed everything since the day it was written. But until the domain was moved across to the new site on 23 September, the address was still answered by the old platform the business had used before, and that platform's robots.txt ended with two lines that close a site to every crawler:
User-agent: *
Disallow: /
Google's robots.txt report in Search Console shows it: a fetch of that old file on 11 September, the block in force until our own file was first fetched on 23 September at 16:10. The site looked perfectly normal to any person who visited it. Only a crawler would have noticed, and only if someone had asked it.
Google is plain about what robots.txt does: "A robots.txt file tells search engine crawlers which URLs the crawler can access on your site" (Google, introduction to robots.txt).

What we changed: every build of this site now fails if its robots.txt would shut Google or Bing out, and our regular checks read the live file the way a crawler does. The lesson for anyone else is simpler. Check what search engines are being told at your real address, before and after the switch, not what your new site says it will tell them.
1. Can search engines get in?
- Open yourdomain/robots.txt on the live address. If it contains
Disallow: /underUser-agent: *, crawlers are being told to stay out of the whole site. - Check for a leftover noindex. Sites are often built with a noindex tag so that the unfinished version stays out of search. Google says that when it finds the tag, it "will drop that page entirely from Google Search results" (Google, block indexing with noindex). Make sure it came off at launch.
- Redirect every old address to its new one. If page addresses have changed, Google recommends "HTTP permanent redirects if possible, such as 301 and 308" and advises against chains of redirects (Google, site moves with URL changes). A page that used to bring in enquiries should not become an error page.
- One address for the whole site. The site should load at one version of the address (with or without www, always https), with the others redirecting to it.
- Submit the sitemap in Google Search Console and Bing Webmaster Tools, and verify the site in both if it is not already.


2. Does it work for a real customer?
- Send a real enquiry through every form, from a phone, and check it arrives, including in the spam folder. A form that looks successful but emails nobody is the most expensive fault on this list.
- Tap every phone number and email address on a phone. They should open the dialler and the email app.
- Go through the site on a phone, not just a laptop, including the menu, the forms and any booking system.
- Try a page that does not exist. It should show a helpful not-found page with a way back, not a blank screen.
- Click every link in the main menu and footer. Links copied from the old site often point at pages that no longer exist.
3. Can you tell whether it is working?
- Analytics records enquiries, not just visits. Set up and test at least one key event for your enquiry form; our guide to GA4 key events walks through it.
- The cookie banner does what it says. Nothing that needs consent should run before a visitor agrees; we cover the rules in do you need a cookie banner for analytics?
- You know your starting numbers. Note last month's enquiries and where they came from, so you can tell in three months whether the new site made a difference.
4. Does everything point at the new site?
- Your Google Business Profile has the right website address, and your other listings match; see our Google Business Profile guide.
- Social profiles, email signatures, directories and any ads point to live pages, not old addresses.
- Printed material and signage with a web address still leads somewhere sensible.
5. The legal basics
- A privacy notice that describes what the site actually collects, including through its forms and analytics.
- Company details, if you are a limited company. GOV.UK says that on "business letters, order forms and websites, you must show" the company's registered number, its registered office address, where it is registered, and that it is a limited company (GOV.UK, running a limited company).

6. If it is a campaign, not a site
- The landing page is live, fast and says what the ad promised.
- Conversion tracking is tested with a real enquiry before any money is spent.
- Daily budgets and end dates are set, and someone will look at the account in the first few days.
- Someone is ready to answer. A campaign that brings enquiries nobody replies to in time is worse than no campaign. If it is Google Ads, our ads audit checklist covers the account itself.
7. After launch
- Day one: check the live robots.txt again, send another test enquiry, and look at Realtime in analytics.
- Week one: in Search Console, open the robots.txt report and the page indexing report. Google's robots.txt report lets you see previously fetched versions from the last 30 days, which is how we found our own problem (Google, robots.txt report).
- Month one: compare enquiries with your starting numbers, and look for pages that bring visitors but no enquiries. If pages are still missing from Google, our guide to why a website is not showing up goes through the causes.
Where the evidence is thin
- How often launches go wrong. We know of no reliable data on how many small business launches break search visibility, forms or tracking. What we can offer is one real example, our own, and the checks that would have caught it.
- What twelve days cost us. We cannot measure what the block cost in visibility or enquiries, because nothing about the old site was being measured by the same tools.
- The order. The order above is our judgement of what does most damage if missed, not a ranking from any study.
If you are commissioning a new site, how to choose a web designer covers what to agree before the work starts, and what a website costs covers the budget. If you would rather have it built, our websites service starts with a landing page sprint at a fixed £3,500, and full sites run from £3,500 to £15,000.

I run oXo Creatives full-time: an independent, founder-led marketing agency and consultancy that I founded in 2013. Until 2026 I ran it alongside senior in-house roles at Google, NatWest, King, PlayStation and Meta.
Want help with websites?
New sites, landing pages, and fixes to the pages you already have, so more of the people who visit actually get in touch. Typically a few weeks, depending on the number of pages and whether a developer is involved.