65 Add-to-Carts, 0 Conversions - Is It the Audience or the Page?

A few weeks back, we spent an entire meeting on a case that drives you nuts. A jewelry brand. An audience launched in early September on Meta, calibrated for fine jewelry, with cart values between 1,000 and 1,700 euros. A month later, the verdict lands. 65 add-to-carts. Zero conversions. Not a single one.

And here's the kicker, the soft metrics were solid. A 2.63% click-through rate since launch, which is very good. So everyone starts eyeing the audience sideways, figuring it just doesn't work. Except it's almost never that simple. Before you kill an audience, you need to look at three things. The audience, the creative, and the landing page. And above all, you need to stop thinking in terms of volume.

65 add-to-carts means nothing

That's the first reflex we killed in the meeting. The account manager was thinking in cost per add-to-cart and in volume. 65 add-to-carts, that's his number. And that's exactly where he was going wrong.

65 add-to-carts, whether you spent 3,000 euros or 10 euros, is a completely different story. The raw volume tells you nothing. Zero info. Until you convert it into a rate, you're flying blind.

What you want to see is the add-to-cart rate relative to sessions, and then the drop-off rate from add-to-cart through to checkout, through to conversion. If your add-to-cart rate is fine but everything collapses afterwards, the problem isn't at the top of the funnel. If the add-to-cart itself is weak, the issue is elsewhere. Before you even ask yourself audience or page, think in rates.

In this specific case, 65 add-to-carts and 0 conversions smells like a structural problem. Not a question of insufficient volume. There's something else going on.

The audience wasn't the real culprit (but almost)

First argument in the audience's favor. It's a legacy audience. We use it successfully on other jewelry clients. So on its own, it's barely a suspect. If it had come out of nowhere, we'd have started there. But no, it runs well elsewhere.

Except there was a massive mismatch. This audience is calibrated for high jewelry. 14,000-euro necklaces, that kind of thing. And what did we put in it? Creatives for far more accessible products. A 130-euro ring, the price of a simple necklace, dropped into an audience built for people who buy at 14,000.

Even if the brand's target average cart value stays high, it makes no sense product by product. Meta learns from signals. You ask it to find high-jewelry buyers, and you show it a 130-euro ring. It brings in people who click (hence the nice click-through rate), but who aren't there to buy that. People click, add to cart out of curiosity, and leave.

So the audience wasn't bad in absolute terms. The audience-plus-product combination was incoherent, that's the real problem.

Why segmenting by average cart value tanked the performance

And this is the point we spent the most time on. We had a logic of splitting products by average cart value. High jewelry on one side, more accessible products elsewhere. On paper, airtight logic. You isolate, you control, you measure.

In real life, it destroys performance.

The 130-euro ring had no reason to be isolated in a separate audience. By pulling it out to stick it in the luxury silo, we stripped it of the context where it could have converted. We over-segmented for an internal accounting logic, and Meta ended up with an unreadable signal.

It's a bit like in a store. You take your accessible product, the one that moves, and you put it in the high-jewelry section next to the 14,000-euro pieces. No one who comes for the high end is going to look twice at your 130-euro ring, and the people who would have bought it never set foot in that aisle. You've made it invisible to the right audience.

The decision was simple. We kill the isolated luxury audience. We fold the 130-euro ring back into an audience that already works, the one where the product makes sense relative to the price point. And we test whether that change of context unlocks conversion. This is exactly the kind of reasoning about targeting control that I break down in stop trying to control everything on Meta Ads. By trying to wall everything off, you stop the algo from doing its job.

Better to leave an accessible product in the audience that performs than to isolate it to satisfy a theoretical segmentation.

The triptych to check before you conclude

Before saying "the audience doesn't work," you review three things. In order.

The audience first. Does it run elsewhere? If it performs on other accounts or other products, it's probably not the problem. If it's new and never tested, then you can legitimately suspect it.

The creative next, and especially its coherence with the audience. That's the point we missed at the start. Putting an accessible-product creative into a high-end audience is asking Meta two contradictory things. The click-through rate can stay good, because people click. But they aren't qualified for that purchase.

The landing page last. Where you send the traffic changes everything. A ring sent to a clean product page is one thing. The same ring buried in a high-jewelry collection is another. And beyond the product page, think about the product's price against the audience's calibration. That's the filter we forget most often.

You check the audience, then the creative, then the page, and you finish with the price. You tick all four before killing anything.

And if you've got add-to-carts that don't convert on site, also think about what happens afterwards. A well-built abandoned cart flow recovers some of these people who added and left. But watch out, it doesn't fix an audience-product mismatch. It recovers the ones who were genuinely interested.

What this case teaches you

An audience that generates add-to-carts and zero conversions is rarely the audience alone. It's almost always an inconsistency somewhere in the chain.

Here's what I take away from this case.

Think in rates, never in volume. 65 add-to-carts means strictly nothing until you look at the spend and the rates that go with it.

Check the coherence between product price and audience calibration. A 130-euro product in an audience built for 14,000 is doomed, even if the nice click-through rate gives you the illusion it's working.

Stop over-segmenting by average cart value. The accounting split that seems logical internally often breaks real performance. Leave your accessible products in the audiences where they make sense.

So next time you've got add-to-carts with no conversions, start there. What exactly did you ask that audience to do.