Tim Kencang.in
Changing the landing page of a live ad carries a risk: the ad platform's algorithm relearns and cost per result jumps. A link rotator solves this. Your ad keeps pointing to one short link, which divides visitors between two or more pages. You can test headlines, prices, or designs without changing the ad.
How the rotator works in Kencang.in
When creating a link, open the A/B Testingtab, add alternative URLs, and set percentages. If percentages are blank, traffic is split equally. The link's main URL is always one variation. The server randomly assigns visitors; cookies mean the same visitor tends to receive the same variation for a consistent experience.
Designing a valid test
One change per test
If version B changes the title, image, and price all at once, you don't know which one has the effect. Test one variable, record the results, then test the next.
Define metrics before starting
Clicks on the buy button, form submissions, or WhatsApp chats. Measure the metric the same way in both versions, for example through Google Analytics events or pixels. Use different UTM for each variation URL (utm_content=a and utm_content=b) to keep sessions and conversions separate in analytics.
Sample size
A 2% difference among 50 visitors proves nothing. As a rough guide, detecting a conversion increase from 3% to 4% at a 95% confidence level requires around 4.000 visitors per variation. Use a sample-size calculator such as the one provided by Evan Miller before starting, and do not stop early simply because one version "looks" like it is winning.
Duration
Run at least one full weekly cycle, as weekday and weekend visitor behavior differs.
Example scenario
- Plan pricing: version A shows monthly prices, version B annual prices with a discount. Metric: clicks on the subscribe button.
- Page length: version A is one screen with a form; version B is a long page with testimonials. Metric: form submissions.
- Marketplace vs. your own site: version A leads to Amazon, version B to your own store. Metric: orders, matched through different coupon codes.
- Language: if audience is mixed, language targeting is more suitable than a rotator.
Reading results
The Kencang.in dashboard shows clicks for each variation so you can check the split follows your weights. Read conversions in your analytics tool. Calculate each version's conversion rate, then test significance; the calculator above provides this test. An insignificant result means there is not enough data, rather than proving the versions are equal.
Once a winner is found
Change the link's main URL to the winning version and remove other variations. The short link and ad stay the same, keeping the campaign running smoothly. Record what you tested, for how long, the results, and the decision. These records become a team asset.
Common mistakes
- Stopping the test when one version leads on the first day.
- Changing a page during the test.
- Sending different traffic sources to different versions, so the test actually compares sources.
- Forgetting conversion tracking on one version.
Other uses for a rotator
Besides testing, rotators help distribute workload: one registration link can rotate among several customer-service WhatsApp numbers, or distribute leads among salespeople. We cover the WhatsApp approach in the article on short links for WhatsApp marketing. A/B testing is available on Pro and Business plans; see the details on the pricing page.
Frequently asked questions
Do the same visitors always get the same version?
Yes, as long as the cookies in the browser are saved. This keeps the experience consistent and prevents one person from being counted on two versions. Visitors who delete cookies or use other devices may obtain different versions.
Can I test more than two versions?
Yes. Add as many alternative URLs as needed and set each percentage. Remember that more versions require more visitors for each version's results to be significant.
What if one of the versions leads to a site that I don't manage?
Traffic splitting still works, but you cannot add conversion tracking there. Use another indicator, such as a different coupon code for each version, to compare sales.
When is a rotator not suitable for use?
When the destination should depend on visitor characteristics such as country, device, or language. Use targeting for these cases, because a rotator splits visitors randomly without considering their profiles.
