New vs returning visitors: two websites inside the same website

Of all the audience conditions available to you, this is the one with the largest behavioural gap and the one small brands use least. A first-time visitor and someone on their fourth visit are not the same customer at a different moment. They need opposite things from the page.

What each one is actually doing#

The first-time visitor is deciding whether you are real. Is this a legitimate shop? Will the product arrive? Can I return it? Who else has bought it? Almost all of their friction is trust and comprehension.

The returning visitor has already resolved those questions or they wouldn't be back. Their friction is different: they're comparing a specific option, checking whether the price changed, looking for the size they wanted, or deciding whether today is the day. What helped the first visitor — the explanatory hero, the brand story, the trust badges — is now an obstacle between them and the thing they came for.

Showing both the same page means either under-serving the newcomer or slowing down the buyer. Usually both.

The changes that pay off#

For new visitors: lead with social proof and the guarantee. Make shipping and returns explicit above the fold. Explain what the product is before explaining why it's better. Keep the primary action obvious but don't push for the purchase in the first viewport.

For returning visitors: compress everything explanatory. Surface what they looked at last time. Lead with availability, size and price rather than with the story. If they've added to cart before without buying, that's the single most actionable state on your whole site.

The returning-visitor version is usually the one that produces the bigger measured effect, and it's almost always the one nobody builds, because the homepage gets designed for the person who's never seen it.

How to test it without wrecking your sample#

The honest caveat: segmenting by visitor type splits your traffic, and for most brands new visitors dominate. If returning traffic is 25% of sessions, a returning-only experiment has a quarter of the sample.

That's fine if you choose the hypothesis accordingly. Returning-visitor changes tend to produce large effects precisely because the mismatch is large, so they often clear the detectability bar that a site-wide microcopy test never would.

In Bixlow, new-versus-returning is a standard audience condition, and it composes with the others — so "returning visitors on mobile who arrived from email" is a single configuration rather than a custom script. Each condition you add shrinks the sample, so add them deliberately: start with visitor type alone and only combine when you have a specific reason.

The definition matters more than people think#

"Returning" has to mean something concrete. A visitor who came once three months ago is not the same as one who came yesterday. If your definition is "has ever visited", you're pooling browsers, comparison shoppers and loyal customers into one bucket and calling it a segment.

Decide the window that matches your purchase cycle. For a consumable with a monthly repurchase, thirty days is meaningful. For furniture, ninety days is still "in the consideration set". Bixlow lets you set the window rather than assuming one, which matters because the wrong window produces a segment that behaves like the average — and then the test comes back flat and you conclude the idea doesn't work.

The trap: don't personalise into a dead end#

There's a failure mode where returning-visitor personalisation narrows the experience until the customer can only see what they've already seen. Someone who looked at one product three times starts getting a site that only talks about that product, and never discovers the thing they'd actually have bought.

Guard against it with a metric that can detect it. Revenue per session catches this where conversion rate on the viewed product does not. If your personalisation lifts conversion on the surfaced item but flattens revenue per session, you've made the site better at selling one thing and worse at selling.

Key takeaways#

• New visitors need trust and comprehension; returning visitors need speed and specifics. The same page can't do both well.
• The returning-visitor experience is usually the unbuilt one, and often produces the larger measured effect.
• Define "returning" with a window that matches your purchase cycle, or the segment behaves like the average.
• Guard personalisation with revenue per session so you catch narrowing before it costs you discovery.