Landing Pages with Poor Web Vitals, Criteo

Card class: Hero • Category: Cross-Channel

​

At a glance

A cross-channel table that joins the product-detail pages Criteo dynamic ads send shoppers to against your web-vitals measurement, and lists every landing page that is both slow (poor Largest Contentful Paint and related vitals) and converting below your account average. Criteo’s whole purpose is to bring a warm, high-intent shopper back to a specific product page. If that page loads slowly, the retargeting cycle pays to deliver the click and then loses the shopper to a sluggish load before they can buy. This card finds the leak at the point it actually happens, the destination page, rather than blaming the campaign. It is where ad performance and site performance meet, and it only exists because Vortex IQ sees both the Criteo serving layer and your web-vitals data.

​

Calculation

Calculated automatically by joining your Criteo landing-page data with your connected web-vitals measurement. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.

​

Worked example

A US outdoor-gear DTC retailer runs Criteo dynamic retargeting that sends shoppers to product-detail pages. A recent PDP template change added a heavy hero-video block that slowed load on several pages. The join lists Criteo landing pages by vitals and conversion rate against the account average. Window is the rolling 30 days to 19 Jun 26. Account average conversion rate on Criteo PDPs is 3.0%. What the pattern tells you:

1. Three PDPs are both slow and converting at well under half the account average. These are the leaks. Criteo delivered warm, high-intent clicks to them, and the pages lost those shoppers before they could buy. The campaign is doing its job; the destination is failing.
2. The healthy boot page proves it is the page, not the audience. The /boots/alpine-gtx page loads in 2.1s and converts slightly above account average on the same kind of retargeting traffic. Same audience, same platform, good page, normal conversion. That contrast localises the fault to the slow pages.
3. The flagged pages share the new heavy template. All three slow pages use the PDP template with the hero-video block; the fast page predates it. That common factor points the engineering fix at the template, not at individual products.
4. The wasted cost is concentrated on the highest-traffic slow page. The 4-season tent page took 4,200 retargeting clicks at a 1.1% conversion rate; a page converting at account average would have produced far more sales from the same spend. Prioritise the fix by clicks-times-conversion-gap, not by which page is slowest in isolation.
5. The fix lives in engineering, not in the ad account. Defer or lazy-load the hero video, compress and correctly size hero images, reserve layout space to cut the CLS shift, and re-measure. As LCP improves toward the good band, conversion rate on retargeting traffic should climb back toward the account average.
6. Re-measure after the fix before touching budget. Do not pull spend off these products; fix the page and let the recovered conversion rate restore the ROAS. Pausing the campaign would forfeit warm demand that the page fix recaptures.

Quick sanity tests:

• Page slow + conversion well below average = genuine destination leak, fix the page.
• Page slow + conversion near average = the speed is tolerable for that product; deprioritise.
• Page fast + conversion below average = not a vitals problem; look at price, stock, or offer.
• Several flagged pages share a template = template-level fix, highest leverage.
• Conversion recovers after a vitals fix = confirmed the page was the cause.

​

Sibling cards merchants should reference together

​

Reconciling against Criteo

Where to look in Criteo’s own dashboard:

Criteo Management Centre → Reporting → Performance Report for clicks and conversions by landing page or product, and a web-vitals source (your real-user monitoring, Google PageSpeed Insights, or a CrUX-based report) for the LCP and CLS readings on those same URLs.

Criteo cannot produce this card on its own. Criteo sees the clicks it delivered and the conversions it attributed per landing page, but it has no idea how fast those pages load, that lives in your web-vitals measurement. The cross-channel join of Criteo landing-page performance against web-vitals data is what Vortex IQ assembles. To reconcile manually, export Criteo’s per-landing-page clicks and conversion rate, pull the LCP and CLS for the same URLs from your vitals source, and the pages that are both slow and below your account-average conversion rate are the ones this card flags. Why our number may legitimately differ from a manual check: Cross-connector reconciliation: This card is inherently cross-channel, joining ad landing-page data to web-vitals measurement:

​

Known limitations / merchant FAQs

Why is a slow landing page Criteo’s problem and not just a site problem? Because Criteo’s entire model is to pay to bring a warm, high-intent shopper back to a specific product page. If that page loads slowly, you have spent the CPC to deliver the click and then lost the shopper to a sluggish load before they could buy. The retargeting cycle’s whole purpose is defeated at the final step. The cost is real ad spend, which is why this sits in the Criteo manifest, not only in a site-health report. My site passes Core Web Vitals overall, so why are pages flagged here? Site-wide averages hide page-level problems. The specific product-detail pages Criteo targets, often newer templates, image-heavy hero blocks, or product-configurator pages, can be much slower than your homepage or category pages that dominate the site-wide average. This card looks page by page at exactly the URLs Criteo sends paid traffic to, which is where the money is leaking. Should I pause campaigns to the slow pages? No, fix the pages. Pausing forfeits warm, high-intent demand that a page fix would convert. The right order is to fix the vitals (defer heavy media, compress and size images, reserve layout space to cut layout shift), re-measure, and let the recovered conversion rate restore ROAS. Reserve any budget reduction for cases where an engineering fix is genuinely impossible in the near term. How do I prioritise which page to fix first? By impact, not by raw slowness. Rank the flagged pages by clicks multiplied by the conversion gap, so the highest-traffic, worst-converting page comes first. A page that is the slowest on the list but takes little Criteo traffic matters less than a slightly-slower page absorbing thousands of paid clicks. Fix for recovered conversions, not for the worst single LCP number. Could ATT distort this card? Partly, on the conversion side only. Web vitals are measured on the device and are unaffected by ATT. The conversion-rate comparison uses Criteo-attributed conversions, which iOS ATT can understate on iOS-heavy pages, so a page might look slightly worse-converting than it truly is. The vitals reading itself is reliable; weigh the conversion gap with the traffic’s iOS share in mind. Why combine slow load and low conversion instead of flagging on speed alone? Because speed alone over-flags. Some slow pages still convert acceptably for their product, and fixing them yields little. The card requires both poor vitals and a conversion rate well below the account average, which isolates the pages where the slowness is actually costing conversions. That keeps the list short and every entry worth an engineer’s time.

---

​

Tracked live in Vortex IQ Nerve Centre

Landing Pages with Poor Web Vitals is one of hundreds of KPI pulses Vortex IQ tracks across Criteo and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. Start for free or book a demo to see this metric running on your own data.