How to track conversions on a Framer site with funnels and A/B tests
Set up tracking IDs, funnels and A/B tests in Framer to see which pages and buttons bring in leads, and read the results honestly.

Most Framer sites launch with a nice design and no idea whether it works. The page gets traffic, a few leads come in, and nobody can say which button, section or headline did the job. When a client asks "is the new homepage better?", you end up guessing.
Framer now has the tools to answer that question without a pile of third-party scripts: tracking IDs on links and forms, funnels, and A/B tests that run on the page itself. This guide walks through how to set them up in a sensible order, what the numbers mean, and where the limits are.
Start with one conversion, not ten
Before you touch any settings, decide what a "win" is for this site. A portfolio might care about the contact form. A SaaS landing page might care about clicks on "Start free trial". A template shop cares about the purchase link.
Pick one primary conversion and, at most, two supporting ones. For example:
- Primary: the contact or demo form is submitted.
- Supporting: a visitor clicks "See pricing" from the homepage.
- Supporting: a visitor opens a case study page.
If you track everything, the reports turn into noise and nobody looks at them after week two. A short list keeps the dashboard useful for you and for the client.
Add tracking IDs to links and forms
Framer counts page views automatically. Clicks and form submissions are only counted when you give them a tracking ID. When you set up a link, add a tracking ID in the link settings. For a form, select the form on the canvas and add the tracking ID there. Every submission is then recorded.
A few naming rules save a lot of confusion later:
- Use one ID per action, and keep it unique. If the hero button and the footer button share an ID, you can never tell them apart.
- Use a clear pattern such as home-hero-trial, pricing-cta-trial and contact-form-submit.
- Use the same ID for the same action across page variants. This matters for A/B tests, as you will see below.
Once IDs are in place and the site is published, the Analytics overview shows a Tracking section with the top tracked events on the site.
Build a funnel that matches the real journey
A single conversion rate tells you the result. A funnel tells you where people drop off. In the site's analytics view, open Funnels from the left menu, add a new funnel and name it after the journey, for example "Homepage to demo request".
Then add steps. Each step can be a page view, a tracked link click or a tracked form submission. A typical lead-generation funnel looks like this:
- Page view: the homepage.
- Link click: home-hero-pricing.
- Page view: the pricing page.
- Form submission: contact-form-submit.
If several buttons lead to the same place, you can add more than one tracking ID to a single step. The step counts when any of them fires. The funnel view then shows the conversion rate for each step, and you can filter by device type or country to see whether, say, mobile visitors drop off at the pricing page.
The step with the biggest drop is your next design task. Fix that before you redesign anything else.
Set up an A/B test on the page that matters
Once you know where people leave, you can test a fix. In Framer, right-click the page in the Pages panel and choose New A/B Test. Framer creates a variant of the page that you can edit freely. You can have up to five variants per page, but two is the right number for most sites, because each extra variant needs more traffic to reach a result.
The setup has three parts:
- Make sure the conversion link or form carries the same tracking ID in every variant.
- Click Configure in the test banner, then add the conversion event as a step. You can add several steps, but only the last one decides the conversion rate.
- Click Start and publish the site (and deploy it, if you use staging).
Traffic is split evenly between variants. Change one meaningful thing per test: the headline, the order of sections, the call to action, or the length of the form. If you change five things at once and the variant wins, you will not know which change mattered.
Read the results without fooling yourself
The test view shows a funnel, a table of variants and a graph over time. Three numbers matter:
- Conversion rate: unique visitors who completed the conversion step divided by all unique visitors to that variant.
- Lift: how much better or worse a variant performs compared with the original.
- Probability of being best: how likely each variant is to be the strongest. Framer treats a variant above 90% with enough data as a winner.
Framer also tells you how many days remain before a winner can be declared, and it marks results as inconclusive when traffic is low or the variants perform about the same. Take that label seriously. A test that is inconclusive after a few weeks is still useful: it tells you the change does not move the number, so you can stop arguing about it.
Do not edit a variant while the test is running. If you need a change, stop the test, make the edit and start a new one.
Know the limits before you promise anything
A few details are worth explaining to clients up front:
- Plans: funnels and A/B tests are part of Framer's Convert add-on, which is available on Pro and Enterprise site plans and is billed by analytics events per month (page views, tracked clicks and tracked form submissions).
- Same-day measurement: Framer assigns variants without cookies, using an anonymous ID that resets daily. A visitor sees the same variant all day, but a conversion that happens on a later day is not attributed to the test.
- Traffic: there is no fixed test length. Low-traffic pages may take a long time to reach a result, so test your busiest pages first.
If you also use Google Analytics or Segment, Framer emits browser events such as framer:click and framer:formsubmit, which you can forward with a small listener script in the site's custom code settings.
Apply the winner and keep a log
When a test finishes, click Stop Test. Framer asks which variant should replace the original. Choose the winner, publish the site again, and delete the variants you no longer need.
Then write one line in a simple test log: the date, the page, what changed, the result and what you will test next. After a few months this log becomes one of the most convincing things you can show a client, because it turns "I think this design is better" into a record of decisions backed by data.
Conclusion
Tracking on a Framer site does not need to be complicated. Choose one primary conversion, give your key links and forms clear tracking IDs, build a funnel to find the biggest drop-off, and run focused A/B tests on your busiest page. Read the probability and lift honestly, apply the winner, and log what you learned. Done consistently, this turns every redesign into an informed decision rather than a guess.
Sources: Framer Help: Run an A/B test on your Framer site, Framer Help: Track links and forms, Framer Help: Set up a funnel.


