Analytics

How to test two versions of an element on your page (Experiments)

How to test two versions of an element on your page (Experiments)

An A/B test in TeleportHQ shows two versions of one element, the original and a variation, each to half of your site's visitors, and counts which version leads to more conversions, such as orders, bookings or sent forms. Select the element, click Try a variation in the right panel, change the copy, then publish. Read the result in Growth → Analytics → Experiments and keep the winner.

Before you start

1. Make the variation

  1. On the canvas, select the element to test, for example your homepage's Hero section.
  2. In the right panel (Inspector), find the A/B test section, below Attributes. It reads "Make a second version of …".

    The A/B test section of TeleportHQ's right panel for a selected section, with the Try a variation button highlighted
  3. Click Try a variation.

TeleportHQ places a copy of the element right after the original, named after it with " · Variation" at the end (for example Hero · Variation), and selects the copy. A message confirms: "The variation is selected. The original stays on the page, slightly blurred. Hover either version to see what it is, and click the blurred one to switch. Publish, then read the results in Growth → Analytics."

  1. Edit the variation like any other element. For example, change its headline to Fresh bread, delivered by 8 am or give its button another color.

Nothing changes on your live site yet: visitors are split from your next publish.

2. Switch between the two versions while you edit

In the TeleportHQ editor, both versions of a test stay on the page. The one you're not viewing is slightly blurred, with a dashed outline.

  • Hover either version to see which one it is, for example "This is the original of …".
  • Click the blurred version to view and edit it.
  • Or select either version: in the right panel, the A/B test section says "This is the original." or "This is the variation." Click Edit the variation or Edit the original to switch.

Every test on the page switches together, so you see the page the way a visitor of each version would.

3. Start the test

  1. Publish your site. See How to publish your site.

From then on, each visitor of your published TeleportHQ site sees one of the two versions, picked at random, so about half see each. A visitor who accepted your cookie banner sees the same version again when they come back with the same browser.

4. Read the results

  1. In the left sidebar, click Growth, then Analytics. Or, with a tested element selected, click View results in the A/B test section of the right panel.
  2. In the column on the left of Analytics, click Experiments ("A/B tests").
  3. At the top right, pick a date range, for example Last 30 days.

Each card on the TeleportHQ Experiments tab is one test, named after the tested element. It says Running on the page, or No longer on the page — past result once the test has ended. Side by side, Original and Variation each show:

  • Sessions: the visits that saw that version.
  • Conversions: how many of those visits then converted.
  • Rate: conversions ÷ sessions.

Show on page next to a version opens its page and selects it on the canvas. A version that was deleted reads "removed".

A test's card on the Experiments tab of TeleportHQ Analytics, with the Original and the Variation side by side and their sessions, conversions and rate

A Winner badge appears on a version once both versions have at least 100 sessions and the difference is unlikely to be chance (statistically significant). The line under the two versions says where the test stands:

  • "No visitors have seen this test yet. Visits count once the site is published with it."
  • "Not enough visitors yet: 60 more sessions needed on the smaller version before a verdict."
  • "No conversions on either version yet."
  • "No clear winner on lead yet — the versions convert about the same."
  • "Variation converts 18% more on purchase — unlikely to be chance."

TeleportHQ compares the two versions on orders or bookings when the test's visitors made any. Otherwise it compares them on the outcome those visitors produced most often, such as sent forms (lead). The line names the outcome it compared.

Visits show on the Experiments tab a few minutes after they happen.

5. End the test

On the test's card on the Experiments tab, click one of:

  • Apply winner · keep the variation (or · keep the original): shown once there's a winner. Keeps the winning version.
  • Keep the variation: keeps the variation, whatever the result.
  • Keep the original: keeps the original and ends the test without a change.

You can also click Keep the original or Keep the variation in the A/B test section of the right panel.

TeleportHQ deletes the other version, turns the one you kept into a normal element (a kept variation loses " · Variation" from its name) and selects it. A message confirms, for example: "The variation stays on the page. Publish to show it to everyone." Publish your site so every visitor sees it.

Deleting either version on the canvas also ends the test: the version left becomes a normal element.

Check that it worked

After you publish, the test's card on the Experiments tab reads "No visitors have seen this test yet. Visits count once the site is published with it." until the first visitors come. A few minutes after their visits, it shows Sessions for both versions.

Troubleshooting

There's no A/B test section in the right panel Visitor analytics is off, your workspace's trial or subscription has ended, or the element can't be tested (the page itself, or an element that belongs to a feature). Turn visitor analytics on in Growth Settings, or select another element, such as a section inside the page.

"No visitors have seen this test yet. Visits count once the site is published with it." No visitor has seen the test yet. If you haven't published your site since you started the test, publish it. Then open your live site and check again after a few minutes.

"Not enough visitors yet: … more sessions needed on the smaller version before a verdict." Both versions need at least 100 sessions before TeleportHQ names a winner. Let the test run, usually a week or two.

"No clear winner on … yet — the versions convert about the same." The difference is too small to rule out chance. Let the test run longer, or end it with Keep the original or Keep the variation.

The card says "Original converts …% less" but the Winner badge is on the Original When the original wins, the line under the two versions reads "converts …% less": the percentage is how much less the variation converts. Go by the Winner badge: it's on the version that converts better.

"That test is no longer on the page." The test was already ended, for example from the right panel. Reopen the Experiments tab.

Good to know

  • Test several elements at once. A page can run as many tests as it has elements. Each test splits its own visitors in half, separately from the others.
  • A visit counts once the visitor has seen the tested element. The version they saw stays with them on the next pages, so an order placed on the checkout page counts for the hero that led there.
  • The version is remembered only after the visitor accepts cookies. Once a visitor clicks Accept on your cookie banner, your site saves the version they see in the browser's local storage (a small storage area in the browser, never sent to servers). Until then, the version holds only while they browse your site, and a reload or a later visit can show the other one. See Does TeleportHQ analytics use cookies?.
  • The split happens even with analytics off, so a page never shows both versions, but nothing is measured.
  • Only sites TeleportHQ publishes report results.
Share

Related Articles

How to see which visits lead to sales and sign-ups (Conversions)

In TeleportHQ, open Growth → Analytics → Conversions to see how many visits reached your site's goal (an order, a booking, a sent form, a sign-up or a newsletter sign-up), with each rate and what it was counted from, every goal your site recorded, and the goal broken down by channel, campaign, page, country or device, credited by the attribution model you choose.

How to see where visitors click and scroll (Heatmaps)

In TeleportHQ, open Growth → Analytics → Heatmaps and pick a page and a device to see which elements visitors of your published site click, where they click in frustration or on things that do nothing, and how far down they scroll, then draw the map over the page on the canvas or on your live site.

How to see who visits your site (Analytics)

In TeleportHQ, click Growth in the left sidebar, then Analytics, to see your published site's visitors, page views, traffic sources, countries, devices, most-viewed pages, clicks, conversions, heatmaps, returning visitors, funnels and A/B tests, for the last 24 hours, the last 7, 30 or 90 days or a custom range.

How to turn visitor analytics on or off and choose tracked pages

In TeleportHQ, open Growth, click the Growth Settings gear and flip the Visitor analytics switch, then publish so the change reaches your live site; Tracked pages below the switch lists the pages you want to record visits on.

Does TeleportHQ analytics use cookies?

No. TeleportHQ's built-in visitor analytics never sets cookies and, like A/B tests and popups, stores nothing in visitors' browsers unless they accept your cookie banner; it records anonymous page views and clicks, doesn't keep IP addresses with visit records, and deletes detailed visit records after 90 days.

How to publish your site

In TeleportHQ, click Publish at the top right of the editor, keep Production selected, and click Publish to Production; a progress bar beside the button shows each step, and after a few minutes your site is live on a free TeleportHQ address that you can share.