1. Install the script once
Each website gets its own snippet: a single script tag you paste into the header (or tracking-code) setting of your website or funnel builder. The Installation page in Flowvert has step-by-step guides for common builders, including GoHighLevel, ClickFunnels, WordPress, Shopify, Webflow and Framer.
Flowvert only marks a page as installed after the script on that page actually checks in. Until then the page shows as waiting for verification.
2. Pick an element and create versions
Open your page in the visual editor and click what you want to test. You can test:
- One element — a headline, paragraph or button label.
- A section — up to 12 text elements changed together as one version, such as a heading, sub-heading and button.
- Styling — colour, highlight, bold and underline on selected words, plus hiding or showing an element.
Each test has the original plus up to four alternatives. Every version can be previewed on your real page before launch; previews are never counted in results.
3. Choose a conversion goal and check it
A goal is what counts as success. Flowvert supports thank-you-page visits, button or link clicks, form submissions, supported booking and form embeds (such as Calendly, Typeform and HubSpot forms), and custom events sent from your own code or Google Tag Manager.
Before launch you can run a goal check: complete the action on your page once and Flowvert confirms it saw it.
4. Launch and split traffic
Each new visitor is assigned to one version and keeps that version on return visits. Traffic is split evenly by default; on plans with custom splits you can set your own weights.
Only one running test can change a given element at a time, so overlapping tests can't overwrite each other.
5. Measure fairly
- A visitor only counts for a version once that version was actually shown to them.
- First visits wait at most 0.8 seconds for the tested element. Visits where a version appeared too late are excluded instead of being credited.
- Results stay marked as preliminary until the test has enough visitors, has run for a minimum number of days, and passes a traffic-split check.
How long that takes depends on your traffic, your current conversion rate and how big the difference between versions is. Some tests end without a clear winner — that is a valid result too.