A/B testing

Setting up your first split

Open the link's details panel and hit Split under A/B testing. You'll see two URL fields:

• Variant A — pre-filled with your link's current destination. Edit it if you want a different baseline.
• Variant B — empty. Paste the URL of the page you want to test against A.

Submit and you have your first split. Both variants start at an even 50/50 share. Add more variants the same way.

Weights

Each variant has a weight (0 to 100). The share of traffic a variant receives is proportional to its weight relative to the others. Three variants at weight 1 each get 33%. Variant A at 8 and B at 2 gives A 80% of the traffic.

Set a variant to weight 0 to take it out of rotation without deleting it — you can switch it back on later. If all variants are at 0, all visitors go to the link's default destination URL.

How visitors are split

Each visitor is consistently assigned to the same variant every time they click — so your conversion data isn't distorted by the same person seeing different pages. No cookies or tracking scripts are involved.

"Why does my own browser always go to Variant A?"

Because you're always assigned to the same variant — and that's intentional. Your own clicks always land on the same page, regardless of incognito mode.

To verify the other variant actually works:

• Switch to mobile data (different connection).
• Use a VPN with a different exit point.
• Ask someone on a different network to click.
• Check the variant's click count in the panel — that's the real signal.

Reading the results

Each variant row shows its 30-day click count and a sparkline. The Control row shows clicks that went to the link's base destination — visitors who clicked before the split was set up, or when all variants were off.

For conversion data, read your affiliate network's reports filtered by the variant label — Linksaff passes it through to your postback.

Things to keep in mind

• Max 8 variants per link.
• Changing weights re-assigns some visitors. Keep weights stable while a test is running if you want clean data.
• Deleting a variant is permanent. Its click history stays in the breakdown, but the variant row is gone.
• Geo targeting and A/B testing are separate modes. Use one or the other per link. See Geo targeting.