Guide
How to test a website before launch: a checklist
A website launch fails in boring ways: a form that emails nobody, a staging setting that hides the site from search, a sign-up nobody outside the team can finish. This checklist covers what to test before launch day, which kind of tool fits each check, and which checks need people.
How do I test a website before launch?
Work through six areas: links and forms, speed on a phone, accessibility basics, search and sharing settings, analytics, and whether a newcomer can do the main task. You or automated tools can run the first five. The last needs someone who has never seen the site, real or simulated, trying to use it.
Why launch-day bugs are always the boring ones
Teams test the features they built and skip the plumbing around them, because it worked last time. The contact form still sends to an old address, the staging block on search engines ships to production, the analytics tag fires twice. None of it is hard to catch. It just needs a list and an afternoon.
- A checklist grouped by area, with the kind of tool for each check.
- The settings that most often ship by mistake from staging.
- Which checks a tool can do, and which need a person.
- Where simulated newcomers fit, and where they are the wrong tool.
How it works
- 01
List the key pages and one core task
Usually the homepage, pricing, the sign-up or contact page, and the main thing a new visitor should be able to finish, such as signing up or booking a demo.
- 02
Run the technical checks
Crawl for broken links, submit every form, and load the key pages on a phone over mobile data.
- 03
Check accessibility basics
Run an automated checker, then go through the core task with the keyboard alone and check that images and form fields are labeled.
- 04
Check search, sharing and tracking
Remove any staging block on search engines, check titles and share previews, and confirm that your analytics record a visit and a sign-up exactly once each.
- 05
Watch newcomers try the core task
Have people who have never seen the site try it, real or simulated, and fix the step where they get stuck.
- 06
Watch the first days
After launch, check error logs, form submissions and analytics daily for a while. That is when the problems nobody predicted show up.
Links, forms and email
Crawl the site with a link checker and fix every broken link and redirect loop. Then do by hand what a crawler cannot: submit every form with an address you can read, and check that the thank-you page appears and the confirmation email arrives, outside the spam folder.
Check where form submissions go. A contact form wired to a test inbox or a colleague who has left is one of the most common launch-day failures, and nobody notices until a customer asks why you never replied.
Try the error cases too: an invalid email, an empty required field, a password that breaks the rules. Each message should say what is wrong and how to fix it.
Speed and devices
Load the key pages on a real phone over mobile data. Lighthouse, built into Chrome's developer tools, reports speed, layout shifts and basic accessibility in one run, and points at the heaviest images and scripts.
Check the pages in at least one browser nobody on the team uses every day. Layout bugs like to hide there.
Accessibility basics
Automated checkers catch a useful share of problems, such as missing alt text, unlabeled form fields and low contrast. They cannot tell you whether the site is usable with assistive technology.
So also do a manual pass: complete the core task with the keyboard alone, check that the focus is always visible, and check that headings describe the page in order. For a real answer on screen readers, test with people who use them every day. Simulated people, including Mimiq's, cannot stand in for that.
Search, sharing and tracking
Staging settings ship by accident. Check that no page carries a noindex tag or a robots.txt rule left over from staging, that each key page has its own title and description, and that a sitemap exists. Then paste a link into a chat app to see the share preview.
Open your analytics in its debug or real-time view, visit the site, and complete the core task once. Each step you care about should record exactly once. Double-counted sign-ups make a launch look better than it is, and missing ones make it look worse.
Check the legal basics that apply to you, such as a privacy policy and, in some countries, consent before analytics cookies are set.
Can a newcomer do the main task?
Every check above can pass and the site can still lose people, because the team cannot see it fresh. The only test for this is someone who has never seen the site trying the main task, with nobody explaining anything.
A few real people from your audience are the gold standard. Simulated people are the fast version. In a Mimiq flow, each one opens your site in their own real browser and tries a task such as 'Sign up and start using it', clicking, typing and filling in forms for up to 40 steps, with an email inbox of their own for verification links. The report shows how far each one got, and where and why they stopped.
For the first impression of each key page, a Mimiq page test shows it to 25 simulated visitors in about 90 seconds and ranks their objections. Neither replaces the technical checks above: simulated people find confusion, not broken tracking.
What simulated newcomers cannot check
They cannot log in to an existing account, sign in with Google or another account, or pay, so a flow starts from a public page and ends before payment. A CAPTCHA can stop them. Anyone stopped by one is marked as blocked by the site, and if it stops everyone, the flow fails and nothing is charged. A staging or preview address without bot protection works, as long as it is public.
They do not measure speed, accessibility, search settings or tracking. And they are simulated: a flow shows where people like your visitors can get stuck, not how many real visitors will finish.
Pre-launch checks by area
| Area | What to check | Who or what checks it |
|---|---|---|
| Links and forms | Every link resolves; every form submits and reaches the right inbox; errors say how to fix them | A link checker, then you by hand |
| Speed | The first screen on a phone over mobile data; heavy images and scripts | A speed report such as Lighthouse, and a real phone |
| Accessibility | Keyboard-only use, labels, contrast, alt text | An automated checker, a manual pass, and people who use assistive technology |
| Search and sharing | No leftover noindex; titles, descriptions and a sitemap; share previews | The page source, and a link pasted into a chat app |
| Tracking | Each key step records exactly once | Your analytics debug or real-time view |
| Newcomer usability | A stranger can finish the main task and understands each key page | Real people from your audience, and simulated people for a fast first pass |
Questions
How long before launch should I test?
Run the full list a few days before launch, so there is time to fix what it finds, and the quick checks again on launch day after the final deploy. Settings change between staging and production.
What do teams forget most often?
A staging block on search engines left in place, and forms that send to the wrong inbox. Both are invisible from inside the team, and both cost real traffic and real leads.
Can I test a website before it is public?
The technical checks, yes, on a staging site. Mimiq needs a public address, because a real browser opens it on Mimiq's side; a staging or preview address works as long as it is public and has no login or bot check in front of it.
Is a usability test the same as QA?
No. QA checks that the site works as built. A usability test checks whether a newcomer can use it. A site can pass every QA check and still lose people at a confusing step, so most launches need both.
What does a simulated check cost?
Your first test is free: sign in with Google and it starts right away. A free account includes 50 credits, no card. After that, a 25-person page test uses 25 credits, a flow uses 5 credits a person, and it is a $9 Starter (120 credits that never expire) or monthly plans from $49.
Keep reading
See what 25 people make of your page.
Your first test is free: sign in with Google and it starts right away. Last checked 2026-10-04.