What the label actually checks
Most analytics tools decide "new" or "returning" by looking for some form of identifier left on a previous visit (a cookie, a device ID, or an equivalent) and checking it against a lookback window. Google Analytics 4, for example, uses a persistent client identifier and keeps counting a visitor as returning as long as that identifier survives on the device. The label is really answering a narrower question than it sounds like: not "has this person ever been here before" but "did this tool recognise them within the window it checks."
Why the split matters more than it looks
In one benchmark of direct-to-consumer ecommerce brands, returning visitors made up only 18.8% of sessions but drove 36.5% of revenue. That kind of gap is common in businesses with repeat purchases or ongoing relationships: a small share of recognised visitors contributes well beyond its share of traffic, so a site that only watches total sessions can miss where the value is actually concentrated.
The profit case for paying attention to it
The split also connects directly to one of the better-known findings in retention research: a 5% increase in customer retention can increase profits by 25% to 95%, a figure that traces back to research from Bain & Company and still holds up as a rough order of magnitude across industries. A rising returning-visitor share over time is one of the few analytics signals that maps fairly directly onto that kind of profit impact, which is part of why it gets tracked even on sites with no ecommerce checkout at all.
Where cookie-free tools draw the line differently
Not every analytics method defines "returning" the same way. A tool built around a rotating daily hash instead of a persistent cookie deliberately caps how long it can recognise the same visitor, since the value mixed into that hash changes on a fixed schedule. A visitor who reads one article today and comes back in three days can show up as two separate new visits in daily reports under that method. That is an intentional privacy trade-off, not a bug, but it means a new-versus-returning number from a cookieless tool and one from a tool using long-lived identifiers are not directly comparable.
How to read the split without over-interpreting it
- Compare it over time, not as a single snapshot. A returning share that climbs after a content push or an email campaign is a clearer signal than the raw percentage on any one day.
- Read it next to engagement, not alone. A high returning share paired with a weak engagement rate usually means people are coming back out of habit, not genuine interest.
- Do not assume a low returning share means a weak site. A business built around one-time purchases or a single project (a roofer, a wedding photographer) will naturally see more new visitors than returning ones, and that is expected, not a problem to fix.
- Remember what "new" costs to measure. Longer identification windows need more persistent identifiers, which is its own privacy trade-off worth weighing against how much the extra precision is actually worth to the business.
Keep it as one signal among several
The new-versus-returning split is useful for spotting direction, a rising or falling returning share says something real about whether people are coming back. Treat the exact percentage with more caution, since it depends on the tracking window and method behind it, and read it alongside how unique visitors are counted rather than as a number that means the same thing everywhere.