MediaTrack.ai Docs

Your Testing Results

Why You Have Four Variants

Every campaign launches with four completely different page designs — Authority, Split, Narrative, and Creative. No one can predict which design will convert best for your specific audience and offer. So instead of guessing, all four run simultaneously and the data decides.

This is standard practice at the world's best CRO agencies. The difference is that your testing happens automatically, from day one, with no manual setup required.

How the Testing Works

When your pages go live, visitors are split evenly across all four variants. As conversion data accumulates, the system identifies which variant is winning and gradually sends more traffic to it.

This is called a multi-armed bandit — a method used by companies like Amazon and Booking.com to continuously optimise which version of a page visitors see.

PhaseWhat's Happening
First few daysEqual traffic to all four variants — gathering baseline data
After ~100 visitors per variantThe system starts shifting traffic toward better performers
Clear winner identifiedThe winning variant receives the majority of traffic

You don't need to do anything. The optimisation happens in the background.

Reading Your Results

In your report, the Page Variants section shows how each variant is performing.

What You SeeWhat It Means
Conversion Rate (CVR)Percentage of visitors who submitted the form
VisitorsHow much traffic this variant has received
Traffic shareWhat percentage of current traffic is going to this variant

A variant with a higher conversion rate will gradually receive more traffic. If one variant is clearly winning, you'll see its traffic share growing week on week.

What Gets Tested

Each variant tests a fundamentally different approach to the same offer:

  • Authority — Does leading with trust, credentials, and social proof convert better?
  • Split — Does a clean, balanced layout with equal weight on copy and visuals work better?
  • Narrative — Does a story-driven, editorial approach that builds desire over the page convert better?
  • Creative — Does a bold, visually distinctive design that stands out convert better?

These are not minor tweaks. They are fundamentally different conversion strategies. The winner tells you something important about your audience's psychology.

What the Winner Tells You

Beyond conversion rate, the winning variant gives you strategic insight:

  • If Authority wins — your audience needs reassurance before they'll act. More social proof and trust signals in your ads may help.
  • If Narrative wins — your audience responds to the story of transformation, not just the offer. Your ad creative should reflect this.
  • If Creative wins — your audience is responding to visual distinctiveness. Bold creative in your ads will resonate.
  • If Split wins — your audience wants clarity and balance. Clean, structured ad creative will align best.

How Long Does It Take?

Meaningful results require approximately 100 visitors per variant — so 400 total visitors to see early patterns. At typical paid traffic volumes:

  • At 50 visitors/day: meaningful data in 8–10 days
  • At 200 visitors/day: meaningful data in 2–3 days
  • At 1,000+ visitors/day: clear winner emerging within 24–48 hours

Your report will tell you where you are in this process.

Can You Override the Winner?

Yes. If you have a strong reason to favour a particular variant — for example, it better represents your brand even if conversion rates are close — you can request this through the feedback section of your report. We'll configure traffic to favour that variant.

A Note on Statistical Confidence

The system only declares a winner when it is at least 95% confident the result is real and not random variation. Until that threshold is reached, the result is a strong signal but not a certainty.

If two variants are very close in conversion rate, the system may never declare a decisive winner — which is itself useful information. It means both approaches work for your audience.

Automatic Winner Promotion

When a variant reaches 95% confidence and has accumulated at least 100 visitors, the system automatically promotes it — shifting the majority of traffic to the winning variant without any action required from you.

This threshold is configurable per campaign. If your traffic volumes are high, we can lower the minimum visitor count so promotion happens faster. If your conversion rates are low and you need more data before committing, we can raise the threshold.

The promotion event is logged in your report so you can see exactly when it happened and what triggered it.

Funnel Tracking

The A/B test measures conversion across five stages of the visitor journey:

StageWhat It Captures
VisitVisitor landed on the page
ScrollVisitor scrolled past the fold
Form startVisitor began filling in the form
Form submitVisitor submitted the form (a lead)
Thank youVisitor reached the confirmation page

The funnel view in your report shows where visitors drop off for each variant. A variant may win on overall conversion rate but lose at a specific funnel stage — this gives you insight into where each design is strongest.

On this page