Why does my new landing page get traffic but zero add-to-carts?

Is your new landing page actually broken?

You launch a new landing page on Tuesday. By Thursday Shopify shows 400 sessions and not one add to cart. The sessions report says a big share came from Sweden and Ireland, almost all of them bounced, and most were on desktop. Someone asks whether the add-to-cart button is broken. Someone else wants to kill the page and send the ads back to the product page.

Before anyone touches the page, answer a narrower question: how many of those 400 sessions were people who saw your ad? On a page that is a few days old, the honest answer can be a small fraction of what the report shows.

A new page meets two kinds of visitors in its first days. The first kind is software: Meta fetching, rendering and reviewing the page because ads point at it. The second kind is shoppers. Only shoppers can add to cart, so the page should only be judged on them.

Skipping straight to step four is how good pages get killed in their first week, and how a redesign gets blamed for numbers that were never about the design.

Has any ad that uses the page actually delivered?

Start in Ads Manager, not Shopify. Find every ad whose main destination is the new URL. Not a site link, not a duplicated ad sitting paused, but the link the ad sends people to when they tap it. Then check whether those ads have impressions, and from what time.

If no ad that uses the page as its main link has impressions yet, then none of the sessions came from your ad being shown to someone. They came from somewhere else. In our audits, the usual somewhere else on a brand-new URL is Meta itself.

If sessions started before the first impression, you have found your first batch of non-shoppers. Everything before that moment comes out of the read. If sessions keep arriving in bursts after delivery starts, keep going: the next section explains where they come from.

Where do sessions come from before an ad has spent anything?

Meta looks at your landing page as part of reviewing your ad. Its Advertising Standards say the review process "may include the specific components of an ad, such as images, video, text and targeting information, as well as an ad's associated landing page or other destinations". Review "relies primarily on automated tools", starts automatically before ads begin running, and ads "remain subject to review and re-review at all times" (Meta Transparency Center). When you edit an ad to fix a rejection, Meta says it is treated as a new ad and reviewed again (same source).

Meta also documents the crawlers it sends to websites. FacebookExternalHit crawls content shared on Meta's apps and "gathers, caches, and displays information about the app or website such as its title, description, and thumbnail image", which is how link previews get built. Meta adds that it "might bypass robots.txt when performing security or integrity checks" (Meta for Developers). The same page lists meta-externalads, which crawls for "improving advertising and other business-related products and services", and meta-externalagent, which crawls for "training foundation AI models or improving products".

Put those together and a new landing page gets fetched by software whenever ads that use it are created, edited or rechecked. That software runs somewhere. Meta's own list of data centers includes Luleå, Sweden and Clonee, Ireland (Meta Data Centers). Meta does not publish which site serves which fetch, so treat the country as a strong clue, not proof. It fits the Sweden and Ireland pattern founders see the week a page goes live.

What do Meta crawler sessions look like in Shopify?

In our own audits, review and preview visits carry a consistent signature. No single signal proves it. Together they are hard to miss.

Two details catch people out. First, the visits come through site links and paused ads too, so "the page has no live ads" does not mean the page has no Meta visits. Second, they can show up as sessions in Shopify and as visitors in A/B testing tools. A split test that started the day the ads were built can carry crawler visits in one arm and not the other, and declare a loser before a single shopper arrives.

What about user agents?

If you can see user agents (server logs, a CDN, some testing tools), look for facebookexternalhit, meta-externalads and meta-externalagent, the names Meta publishes. Most founders only see the Shopify sessions report, so the signature above matters more than the user agent.

How do you judge the page on real visitors only?

Cut the data down in this order. Each cut removes a kind of visit that could never have bought.

1. After real delivery. Drop every session before the first impression on an ad that uses the page as its main link.
2. Outside the edit windows. Drop bursts in the hour after ads were created, duplicated or edited, even once delivery has started.
3. Your target country. If you sell mostly in the US, read the US alone.
4. Real placements. Keep sessions that carry your ad UTMs. Drop untagged bursts and stray right-column hits that line up with edits.

Here is what those cuts do to the 400 sessions from the opening.

Thirty-eight real sessions and zero add to carts feels damning. It is not, yet. If your control page adds to cart on 6% of sessions, a page exactly as good would still show zero on 38 sessions about 1 time in 10. At 150 real sessions with zero add to carts, that chance falls below 1 in 10,000, and you have a real problem. (Illustrative rates.)

What should you compare the new page against?

Not an industry benchmark and not last month. Compare it to a control page: your current best page for the same product, getting traffic from the same campaign or audience, in the same hours. Same hours matters because the day of the week, the auction and the audience all move add-to-cart rates on their own. Our guide on Shopify conversion rate benchmarks for supplements explains why outside averages mislead here.

Then read each step for both pages. Shopify's funnel counts sessions with cart additions, sessions that reached checkout and sessions that completed checkout. A session only counts as reaching checkout when there was "user input (for example, a key press or mouse click) during checkout" (Shopify Help Center).

A page as good as the control would produce about 11 add to carts on 180 sessions. Three or fewer would happen by chance well under 1 time in 100. So the problem sits between landing and the buy button, and the rest of this guide is about that step. (Illustrative rates.)

If the add-to-cart rate holds and orders do not come, the leak is later, in the cart or checkout. Our guide on add to carts but no purchases covers that half. If the whole site slipped at the same time as the new page launched, start with a Shopify conversion rate drop instead.

Does the page say what the ad promised?

Once the visitors are real, the first thing to check is message match. The ad makes one promise: a specific situation, a specific person, a specific offer. If the page opens on something else, the visitor has to work out whether they landed in the right place, and on a phone most will not bother.

A typical version: the ad is a creator talking about a 30-second morning routine, and the page opens on the founder story and a wall of ingredient science. Nothing on the page is wrong. It is answering a question the visitor did not ask yet. Put the ad and the first screen of the page side by side and check four things: the same problem, the same words, the same offer and the same product shot or face.

Page type matters here too: a cold visitor from a story-led ad may need a presell before a buy box. Our guides on landing page vs product page for Meta ads and whether advertorials work cover that choice.

Price or offer shock

The ad says one thing about the offer, or nothing at all, and the page reveals something else: a higher price than expected, a bundle-only deal, a subscription that looks mandatory, or shipping costs that appear late. People who were curious at the ad leave at the price.

Baymard Institute's abandonment research puts extra costs (shipping, tax, fees) at the top of the reasons shoppers abandon, at 40% once "just browsing" shoppers are excluded (Baymard Institute). That research is about checkout, but the reaction does not wait for checkout when the surprise sits on the landing page. Check that the offer in the ad is the offer at the buy box, and that the total cost is clear before the tap. Our guide on subscribe and save discounts covers how the subscription option itself can scare people off.

Can a visitor on a phone actually add to cart?

Is the buy button where people scroll?

Long landing pages often put the first buy button far down the page. Attention does not reach that far evenly. In Nielsen Norman Group's 2018 research, users "spent about 57% of their page-viewing time above the fold" and 74% of it in the first two screenfuls (Nielsen Norman Group). Put scroll depth from session recordings or heatmaps next to where the first button sits. If most real visitors stop before it, they never had the chance to add to cart.

Does the button work for every variant and plan?

New landing pages often use a rebuilt buy box: a custom section, a page builder block or a bundle widget. Test it on real phones, iPhone and Android, for every variant, bundle and purchase option. Subscriptions are the fragile part. Shopify's theme documentation says that to add a selling plan to a cart, "it's necessary to include an input with the name 'selling_plan', which will carry the selling ID as its value" (Shopify.dev). A buy box that drops that input can break the subscribe option while one-time purchase still works.

• Tap add to cart on each variant and each purchase option, on a phone, logged out.
• Check that the right variant, plan and discount land in the cart or drawer.
• Check that no popup, chat bubble or sticky bar covers the button on small screens.
• Check the page in the Instagram and Facebook in-app browsers, not only Safari or Chrome.

Does the page load fast enough on mobile?

Google's Core Web Vitals thresholds for a good experience are a Largest Contentful Paint of 2.5 seconds or less, an Interaction to Next Paint of 200 milliseconds or less and a Cumulative Layout Shift of 0.1 or less, measured at the 75th percentile of page loads and segmented across mobile and desktop (web.dev). An LCP above 4 seconds counts as poor (web.dev).

New landing pages are often heavier than the product page: a hero video, large images, extra apps for reviews, timers and bundles. Test on a phone over mobile data, not on the office wifi. A layout shift that moves the button just as someone taps costs add to carts directly.

Is the page getting the wrong traffic?

Sometimes the page is fine and the visitors are not the ones you meant to pay for. Meta lists Audience Network as its own publisher platform next to Facebook, Instagram, Threads and Messenger, and the right-hand column as its own Facebook position (Meta for Developers). Audience Network is how app publishers show Meta ads and monetize their apps (Meta Audience Network). A tap from inside another app, or from a desktop sidebar, can behave very differently from someone in Feed or Reels who tapped on purpose.

• Placement mix. If your URL parameters carry the placement, split sessions and add to carts by placement. If they do not, compare Ads Manager's placement breakdown with Shopify sessions by device.
• Campaign and audience. A new page launched on a fresh cold campaign, judged against a control fed by warm retargeting, is not a fair fight.
• Country mix. If the campaign targets several countries, the new page may be getting a different mix from the control. Read your main market alone.
• Clicks vs sessions. If Meta reports far more link clicks than Shopify sees sessions, the leak is between the ad and the page, before anyone sees the buy button.

The fix is to compare like with like. If the new page sits in a different campaign from the control, the add-to-cart gap may be the audience, not the page. Our guide on Intelligems vs Shoplift covers how to split traffic so both pages see the same visitors.

How we judge a new landing page at Succession

When a founder tells us a new page failed, we do not open the page builder first. We rebuild the read.

1. Filter Meta's crawlers first. We cut out the bursts from Meta's data centers that line up with ads being created or edited: mostly Sweden and Ireland, desktop or other devices, no UTMs, near 100% bounce, never a purchase.
2. Judge after real delivery. Nothing before the first impression on an ad that uses the page as its main link counts toward the verdict.
3. US first. We read one market alone, usually the US, before any rate, so mixed international traffic does not drag the numbers down.
4. Per-page step rates. Session to add to cart to checkout to order for this page alone, not the sitewide rate, so one page's leak does not hide in the average.
5. A control page in the same hours. Same product, same campaign, same window. The new page is judged on the gap to the control, not on a benchmark.
6. Scroll depth against the buy button. Recordings and heatmaps show whether real visitors ever reached the first button.

That read tells you which of three things the page needs: more real traffic, a fix to one step, or a rewrite. A full CRO audit goes further, across every page and the checkout.

What should you do this week?

1. In Ads Manager, list every ad that uses the new page as its main link, and note when each was created, last edited and first got impressions.
2. In Shopify, turn on the Human or bot session filter and pull the page's sessions by hour, country and device since it went live.
3. Cut everything before real delivery, every burst in the hour after an ad edit, every country you do not target and every session without your ad UTMs.
4. Count what is left. If your normal add-to-cart rate would not yet have produced several add to carts on that many sessions, keep counting.
5. Once there are enough real sessions, put the page's step rates next to a control page in the same hours and the same campaign.
6. If the first step leaks, check message match with the ad, the offer at the buy box, the button on real phones for every variant and plan, scroll depth and mobile speed.

If you would rather have someone do the split with you, our free growth audit runs this read on your new page, live.

FAQ