How an A/B split link works
The link’s normal destination is variant A. You add up to five more destinations, each with a weight from 1 to 100, so one link can have six variants in total. On every click, urlf.ly picks one variant at random in proportion to the weights. With weights of 50 and 50, each page gets about half of the clicks; with 90 and 10, the second page gets about one click in ten.
The choice is made again on each click, so the same person can land on a different variant if they open the link twice.
How to set up an A/B test link
- Sign in and open the dashboard.
- Paste the first page into Destination URL. This is variant A.
- Open Advanced options and find A/B split test.
- Set the Weight of the destination above (variant A), for example
50. - In Other variants, add one line per extra page: the weight, a space, then the URL, for example
50 https://example.com/b. - Press Create link and share the short link.
You can change weights or variants later on the link’s Edit page. The short link stays the same.
Reading the results
The link’s stats page has an A/B split results table with one row per variant: Variant, Destination, Weight, Clicks, Unique visitors, and Share of the clicks. It follows the date range you pick, and bots are left out of the counts.
Which rule wins
A link can combine several routing rules. A matching country rule from geo-targeting is applied first, then mobile targeting, and only then the A/B split. Keep this in mind if a test link also has other rules.
Good uses for A/B test links
- Landing page tests. Send half of a campaign to each version of a page and compare results in your own analytics.
- Offer tests. Try two prices, bundles, or sign-up pages behind the same link.
- Gradual rollouts. Send 10 percent of visitors to a new page first, then raise its weight.
Limits to know
urlf.ly counts clicks per variant; it does not track what visitors do on the destination page, calculate statistical significance, or pick a winner for you. With small numbers of clicks, the split can differ noticeably from the weights because each click is chosen at random. To measure sign-ups or sales, use the analytics on the destination pages.
A/B split testing is part of the paid plan. You can upgrade from your dashboard.
Frequently asked questions
How many destinations can one A/B link have?
Up to six: the main destination (variant A) plus up to five other variants.
Does the same visitor always see the same variant?
No. A variant is picked at random by weight on every click, so a returning visitor can land on a different one.
Where can I see which variant performs better?
On the link’s stats page, in the A/B split results table, with clicks, unique visitors, and share for each variant.
Does urlf.ly track conversions for each variant?
No. It counts clicks per variant. Use the analytics on your destination pages to measure conversions.