A Core Web Vitals audit on an Adobe Commerce storefront is a different kind of engagement from a general performance audit. The CWV audit is anchored on specific, measurable, Google-published metrics: Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift. The deliverable is not “your site is slow.” It is a prioritized list of interventions, each with a measurable target, an estimated effort, and an expected metric impact. This article describes how a serious Magento CWV audit actually runs, what gets checked, and what gets fixed first when the audit produces findings.
The framing here reflects how Bemeir’s Magento performance practice actually structures these engagements. The work is methodical, the findings are evidence-based, and the recommendations are ordered by impact-per-effort rather than by what is easiest to talk about.
What CWV audits are anchored on
The three Core Web Vitals metrics are defined by Google’s Web Vitals initiative and they are the metrics that affect both user experience and SEO ranking signals. The thresholds matter:
- LCP: “Good” is under 2.5 seconds. “Needs improvement” is 2.5-4.0 seconds. “Poor” is over 4.0 seconds.
- INP: “Good” is under 200 ms. “Needs improvement” is 200-500 ms. “Poor” is over 500 ms.
- CLS: “Good” is under 0.1. “Needs improvement” is 0.1-0.25. “Poor” is over 0.25.
The CWV audit is fundamentally about moving each of these metrics into the “Good” band for 75% of the retailer’s real users, measured at the 75th percentile on mobile devices over a rolling 28-day window. That is the bar Google measures against for the page experience signal, and it is the bar the audit’s prioritization framework anchors on.
Mid-market Adobe Commerce storefronts typically arrive at the audit with LCP in the 3-5 second range on mobile, INP in the 300-600 ms range, and CLS variable but usually 0.05-0.20. The audit’s job is to identify the specific interventions that will move each metric into the Good band, ranked by the effort required and the impact per unit of effort.
What the audit actually checks
A complete Magento CWV audit covers eight distinct technical surfaces, each of which can contain meaningful CWV-affecting issues:
| Surface | What gets examined | Common findings |
|---|---|---|
| Server response time (TTFB) | Backend Magento controller response, Varnish hit rate, database query patterns | N+1 queries, cold caches, slow controllers on PDP |
| Critical rendering path | HTML structure, CSS delivery, blocking JavaScript | Render-blocking CSS, synchronous JS, missing preload hints |
| LCP element | What is the LCP element on each key page, and how is it delivered | Unoptimized hero image, lazy-loaded LCP image, missing fetchpriority |
| Image delivery | Format, compression, dimensions, responsive sizing, CDN behavior | Missing WebP, missing srcset, oversized images, no CDN |
| JavaScript execution | Main thread blocking, third-party scripts, bundling | Heavy KnockoutJS payload (on Luma), unbounded third-party scripts |
| Layout stability | Image dimensions in HTML, ad slot sizing, font loading | Missing width/height on images, FOIT/FOUT, dynamic content injection |
| Interaction handlers | Click handlers, form interactions, cart updates | Long event handlers, blocking API calls during interaction |
| Third-party scripts | Analytics, chat, personalization, tag managers | Unbounded third-party JavaScript, late-loaded scripts |
Each finding is captured with evidence: a Chrome DevTools trace, a CrUX field measurement, a Lighthouse score breakdown. The evidence matters because the remediation prioritization depends on knowing the actual impact, not the theoretical impact.
What we fix first
The CWV audit produces a prioritized remediation list. The prioritization is not arbitrary. Across hundreds of mid-market Adobe Commerce audits, the following sequence has produced the largest impact per dollar spent.
1. LCP element optimization. The single most impactful intervention on most Adobe Commerce sites. Identify the LCP element (almost always the hero image on the homepage, the product image on the PDP, the category banner on the CLP), ensure it is preloaded with `<link rel=”preload”>`, ensure it is served in WebP or AVIF format at the right dimensions, ensure it has `fetchpriority=”high”`, and ensure it is not lazy-loaded. The typical impact is 500-1,500 ms of LCP improvement on mobile.
2. Render-blocking CSS reduction. Magento’s default Luma CSS is heavy. Inlining critical CSS for the above-the-fold experience and deferring the rest typically saves 300-800 ms of LCP. On Hyvä-based storefronts this is largely handled by the theme itself, which is one of the reasons Hyvä migrations produce such large CWV improvements.
3. Third-party script audit and bound. Most Adobe Commerce storefronts have accumulated five to fifteen third-party scripts: analytics, chat, A/B testing, personalization, social pixels, tag manager. Each one is a candidate for retirement or deferred loading. According to Web Almanac’s research on third-party JavaScript, the median eCommerce site loads roughly 700 KB of third-party JavaScript, much of which is non-critical. The audit identifies which scripts are pulling weight and which are not, and the remediation defers or removes the latter.
4. Image delivery optimization at the CDN layer. Modern image CDNs (Cloudinary, Imgix, Fastly Image Optimizer, Cloudflare Images) can serve right-sized, modern-format images on the fly. Most mid-market Adobe Commerce sites are still serving JPEGs at full resolution, or have a CDN configured suboptimally. The remediation usually delivers 200-600 ms of LCP improvement and meaningful bandwidth reduction.
5. Database query optimization on critical pages. TTFB matters because it sets the floor for LCP. Audits routinely find Magento controllers with N+1 query patterns on PDPs, slow EAV attribute joins on category pages, and missing indexes on customer-specific queries. The remediation is targeted and effective: 100-300 ms TTFB reduction is typical, which translates roughly 1:1 into LCP reduction.
6. INP optimization on interactive surfaces. INP became a Core Web Vital in March 2024 and it has become the metric that separates “good” from “great” Adobe Commerce sites. The remediation targets long event handlers on cart updates, search interactions, and filter applications, breaking them into smaller chunks with `scheduler.yield()` or `requestIdleCallback` patterns where appropriate. The work is more nuanced than LCP optimization but the impact is meaningful: typical INP improvements are 150-400 ms.
7. Layout stability fixes. CLS issues on Adobe Commerce are usually caused by three patterns: images without dimensions in HTML, ad slots or banners that load dynamically, and font loading that triggers reflow. The remediations are mechanical and the impact is binary: most CLS issues either fix completely or don’t fix at all.
8. Hyvä migration evaluation. If the audit reveals that LCP improvements are bottlenecked by Luma’s inherent JavaScript weight, the remediation roadmap includes a Hyvä migration evaluation as a follow-on engagement. Bemeir’s Hyvä migration team typically runs this as a separate scope after the CWV audit identifies whether the migration math works.
What the audit produces
The deliverables are three documents: a technical findings report (15-30 pages, written for the engineering team), an executive summary (4-8 pages, written for the C-suite), and a prioritized remediation roadmap (the operating document for the next quarter). Each finding includes evidence, expected impact on the relevant CWV metric, estimated remediation effort, and dependencies on other findings.
The executive summary translates the technical findings into business impact. According to the Google Core Web Vitals research and corroborating retail benchmarks, every 100ms improvement in mobile LCP correlates with roughly 1% improvement in mobile conversion. The summary uses the retailer’s own conversion baseline and mobile revenue share to translate the projected CWV improvements into revenue impact, which is the language the budget conversation actually happens in.
How long the audit takes
A complete Magento CWV audit takes two to three calendar weeks: roughly 60-80 hours of senior performance engineering effort, plus time for the retailer’s team to provide access and answer follow-up questions. The audit cannot be compressed below this floor without sacrificing depth, because the most valuable findings come from running real traffic simulations, examining actual user-experience data from the CrUX dataset, and tracing specific page loads end-to-end.
Audits that promise CWV findings in three days are running Lighthouse and producing the Lighthouse output as a deliverable. Lighthouse is a useful tool but it is not an audit, and its output is not a remediation plan. The difference between a Lighthouse run and a real audit is the difference between a thermometer and a medical examination.
What the audit costs
Mid-market Adobe Commerce CWV audits typically cost $20,000 to $50,000. The variation reflects the platform’s complexity and the depth of the engagement. At the typical price point, with a typical remediation outcome of 5-8% conversion lift on mid-market traffic, the audit pays back inside three to four months, and the remediation work that follows pays back inside six to nine months.
The follow-on work
The audit’s job is to identify and prioritize, not to remediate. The remediation work that follows is a separate engagement, typically scoped against the prioritized roadmap. The retailers who get the most out of the audit treat the deliverables as an operating plan rather than a binder, working through the prioritized list in sprints over the following quarter. Bemeir’s Adobe Commerce performance practice sometimes runs the remediation work as a follow-on engagement and sometimes hands it off to the retailer’s incumbent agency, depending on the right fit. The audit’s value does not depend on who does the follow-on; it depends on whether the follow-on actually happens.
That follow-through is the difference between a CWV audit that produces a temporary improvement and one that produces a permanent change in the platform’s performance posture. The platforms that maintain strong CWV scores over three years are the ones with audit-driven discipline applied as a recurring practice, not as a one-time event.
