Because it's made up of people who don't know you yet, often on a phone, in the middle of something else.
That gap is normal, and it's also poorly measured, because some of the sales your ads drive end up credited to "everything else." The useful question isn't why paid converts lower, but whether it converts enough for what it costs.
The benchmark
Compare like with like. In the 2026 Contentsquare benchmark, based on 99 billion sessions, a new visitor converts at 1.7% and a returning visitor at 2.9%. Judge your paid discovery traffic against the first number, since your site's average is pulled up by your loyal customers.
In practice, the gap is often much wider. In two stores we audited this winter, returning visitors converted nearly five times better than new ones in the first, and thirteen times better in the second. Even so, paid wasn't one uniform block.
In the first store, paid search actually converted slightly better than organic search (3.5% versus 3.3%). In both, the main Meta source fell according to GA4 to between 0.04% and 0.05%, barely one sale per two thousand visits.
What the number doesn't see
Faced with a result like this, we first labeled the source "catastrophic" in one of these audits, before adding a whole box of caveats.
The mechanism explains why. When someone taps an ad on Instagram or Facebook, the page opens in the app's built-in browser, which keeps its own cookies.
An Apple engineering team member points this out on the developer forum about these embedded web views: there is no way to share their cookies with Safari's.
People look, close the app, then come back two days later from Safari, by typing your name into Google or opening your newsletter.
To your analytics tool, that return visit comes from a stranger who arrived through another channel. Ads paid for the first date, and another channel cashes in the sale. We conclude that part of the strong conversion of everything else belongs to paid, and no default report shows it.
At a third brand, Meta's ad platform reported a positive return on spend for campaigns that GA4 judged nearly sterile.
The rule
Separate three families of campaigns: discovery, retargeting and searches for your brand name. Each has its own benchmark, and blending them produces a number that describes none of them.
Check your campaign tags first. At a brand we advise, three conventions coexisted for the same ad platform (meta / paid, facebook / paid and fb / paid), so no report showed paid social in full.
The steps
- Conversion by family. Calculate it separately for discovery, retargeting and brand, over the same period.
- The return within the month. In a GA4 exploration, cross the "First user primary channel group" with the "Session primary channel group," then look at where visitors who first came through paid social end up buying. This number remains a floor, since returns from another browser slip past it.
- The landing page. Does it answer, within the first screens, the questions of someone who doesn't know you: who you are, when the package arrives, how to send it back?
- The comparison over time. Track each family separately from one period to the next.
- The sentence for leadership. Fill it in with your numbers:
"Our paid discovery converts at X%, versus 1.7% for new visitors across the market, and Y% of the people it brings in then buy through another channel."
The trap
The trap is optimizing the site for an average visitor, who doesn't exist.
Your loyal customers already know who you are and how returns work, whereas someone arriving from an ad needs to read that right away.
And if you judge discovery by last click, you risk cutting the campaigns that bring in your future customers.
