How to A/B test a landing page without resetting its analytics

An A/B test should answer a new question without erasing what happened before the experiment. In QuickPages, the underlying pages continue to collect their own views and conversions while the A/B test report measures only traffic that entered through the shared experiment URL.

This separation lets you compare variants fairly and still open either page later to review its complete history.

1. Start with a clear hypothesis

Do not create a second version simply because the first page feels old. Write down the decision the test should inform. Good hypotheses compare one meaningful difference:

• A specific outcome headline versus a problem-focused headline.
• A short form versus a form with one qualifying field.
• Case-study proof versus a benefits checklist.
• A direct call to action versus a lower-commitment invitation.

If the variants change the offer, layout, form, and destination at the same time, a winner will not tell you which change mattered.

2. Duplicate the page as an A/B variant

Open the existing page settings and choose Duplicate as A/B Variant. QuickPages creates a new published page with a safe unique name and URL, copies its design and settings, and opens the A/B test setup with both pages already selected.

The copy does not inherit the original page's traffic, leads, or previous A/B test membership. Edit the duplicated page and make the one change your hypothesis requires.

3. Create the shared experiment URL

Give the experiment a recognizable internal name, choose the publishing domain, and enter a short test link name. Select at least two eligible published entry pages and set the initial status to active when you are ready to distribute traffic.

4. Send all experiment traffic to one address

Use the shared experiment URL in ads, emails, social posts, and campaign links. QuickPages assigns visitors to active variants and keeps them on the selected version during the experiment.

Do not split campaign links manually between the individual page URLs. Direct visits still count on each page, but they are intentionally excluded from the A/B test comparison because QuickPages did not distribute that traffic.

5. Read visitor and conversion metrics separately

The report shows experiment visits, estimated unique visitors, conversion submissions, and visit conversion rate. One visit can submit more than once, so QuickPages preserves both facts:

• Conversions count successful submissions.
• Visit conversion rate measures how many experiment visits converted at least once. Repeat visits count separately.
• Estimated unique visitors is an IP-based approximation, not a claim that QuickPages can identify individual people.

Use the variant table to compare traffic allocation and results. A variant with little traffic should not be declared a winner simply because one early visitor converted.

6. Review the actual leads, not only totals

Click View Leads to open the lead inbox with the A/B test filter clearly displayed. This is where you can inspect the people and submitted fields behind the conversion totals.

The experiment's statistics CSV contains variant totals. The lead inbox export contains individual submissions. Choose the lead export when you need email addresses, form answers, source details, or a client handoff.

7. Wait for a decision-quality result

There is no universal number of visits that makes every test conclusive. The required sample depends on the normal conversion rate and the size of the difference you care about. As a practical rule, do not stop after a few conversions or after checking repeatedly for whichever version is temporarily ahead.

Also review lead quality. A shorter form may increase submissions while producing fewer qualified conversations. The winning variant should improve the business outcome, not merely the easiest metric.

Do not start an A/B test when the page has almost no traffic, when a destination is broken, or when the variants change several unrelated ideas at once. Fix technical problems directly and use testing for a specific decision that future traffic can answer.

8. Pause or end the experiment deliberately

Pause traffic when a variant is broken or when the campaign is temporarily stopped. When the test has answered the question, select a winner and end it. The shared URL can then route visitors to the winning page.

Deleting an A/B test soft-deletes the experiment relationship and report; it does not silently destroy the pages. QuickPages also prevents deleting a page while it belongs to an active test, so the experiment cannot be left with a missing variant.

Use the reporting guide when you need individual leads rather than variant totals, and the UTM guide to keep acquisition context attached to those conversions.