Visitor Type

The visitor type condition targets your experiment at new visitors or returning visitors. A new visitor has never been to your site before. A returning visitor has visited at least once. The two groups often behave differently, so the message that works for one can fall flat for the other.

A New/Returning Session condition in the Audience Builder with two radio options (New Visitor and Returning Visitor) with Returning Visitor selected. — The visitor type condition: a single choice between new and returning visitors.

Visitor types

New visitors

A new visitor has no A vs B visitor cookie in their browser yet. This is usually their first session on your site. They have never used your product before, so they are judging your brand for the first time.

Returning visitors

A returning visitor already has an A vs B visitor cookie in their browser, set on an earlier visit. They know your product to some degree. They may have browsed before, or they may already be a customer.

How detection works

The snippet decides new versus returning on its own, using a cookie it sets in the visitor's browser. You do not configure anything: this happens automatically for every visitor.

Cookie name: _avsb_visitor. It holds a unique visitor ID plus the visitor's current experiment assignments.
Where it lives: a first-party cookie in the visitor's own browser. Nothing is stored on our servers.
How long it lasts: one year, and it renews to a fresh year on every visit. An active visitor's cookie never expires on its own. Someone who stays away for a full year looks new again on their next visit.
First visit: no cookie exists yet, so the visitor counts as new. The snippet sets the cookie itself, as soon as it loads. You do not need to do anything.
Later visits: the cookie exists, so the visitor counts as returning.

Visitor type relies on a cookie. Anything that removes or blocks that cookie makes a returning visitor look new again. This happens when someone clears their browser cookies, switches to a different browser or device, or browses in private or incognito mode. It also happens when a visitor blocks cookies entirely, for example with a strict privacy setting or an ad blocker. In that case every visit from that browser reads as a brand new visitor, forever, because the cookie can never be set. This is a known limit of cookie-based tracking. It applies across the industry, not only to A vs B.

Common use cases

Welcome offers for new visitors

Test a first-visit discount banner or a welcome message exclusively for new visitors. Returning visitors (who may have already seen or dismissed the offer) are not enrolled, so they never see the test banner.

Loyalty messaging for returning visitors

Test whether a "Welcome back" message or a loyalty-themed hero section improves conversion for returning visitors. New visitors do not know your brand yet, so keep them in the control group. The loyalty message would only confuse them.

Onboarding flow testing

Run an onboarding flow experiment for new visitors only. A returning visitor already knows your product, so the same flow would feel condescending or confusing to them. Limiting the experiment to new visitors keeps the results meaningful.

Running parallel experiments

Sometimes you want the same experiment idea, with different messaging for new and returning visitors. Create two experiments: one targeting new visitors, one targeting returning visitors. Write copy suited to each group. Then compare the results separately to see what works for each.

Combine visitor type with a traffic source condition for sharper targeting. An audience is a named group of visitors defined by rules like these two combined. "New visitors from paid traffic" is a good one for top-of-funnel experiments. Those visitors are seeing your brand for the first time, and they arrived with intent from an ad.