ALM Website Performance Audit
Technical findings and remediation plan for Overfuel
Mobile shoppers wait 10.5 seconds for lead inventory to appear.
ALM’s digital showroom experiences acute mobile execution gridlock. While high-end desktop hardware masks these delays, mobile shoppers experience over 10 seconds of rendering latency, resulting in Google Core Web Vitals failure and preventable ad-click lead abandonment.
This technical audit is engineered exclusively for implementation within Overfuel’s existing code framework and platform architecture. It does not propose, require, or advocate for replatforming, rebuilding the site, or migrating to Flow Web Design’s technology stack. Overfuel’s engineering team will execute these targeted optimizations directly within their current codebase and build pipeline, preserving all existing inventory syndication, dealer workflows, and platform tools.
1. Executive Assessment and Evidence
Section 01Lighthouse throttles processor and network conditions to represent a standard consumer smartphone on mobile cellular data. Under these real-world mobile constraints, Overfuel's unoptimized script load completely halts rendering: Score 34/100, LCP 10.5s, and TBT 5,700ms.
Real-world field data aggregates luxury buyers on top-tier smartphones ($1,200 iPhones) over fast home/office Wi-Fi in metro Atlanta. While powerful hardware conceals some lag, the site nonetheless fails Core Web Vitals with LCP (2.5s) and CLS (0.10) hanging on a precarious razor's edge.
Mobile Laboratory Results (Synthetic Lighthouse Simulation)
Moto G Power / Throttled NetworkAcute degradation under simulated mobile CPU/network throttling
Hero inventory asset discovery blocked by script parsing and layout tasks
Long tasks (>50ms) monopolize main thread during initial page load
Script evaluation, hydration, and compilation of client-side bundles
Dominated by script parsing, layout recalculations, and styling passes
Initial render delayed by render-blocking resources
Layout stability is well within Google recommended threshold
Visual completion heavily deferred for mobile viewports
4.6 MB total transfer size over mobile connections
Real-User Field Results (Chrome User Experience Report — CrUX)
28-Day Rolling Aggregate (75th Percentile)Site fails Google 75th-percentile field compliance thresholds
Borderline value resting on the exact good/needs-improvement boundary
Needs Improvement range (200ms–500ms); input delays during user interactions
Resting precisely on the rounded boundary of the Good threshold
Server response / edge delivery requires optimization
This audit evaluates the baseline performance dataset provided in the supplied PageSpeed Insights report, rather than a freshly scheduled independent run. The supplied text omitted the exact tested URL and whether field metrics reflect URL-level or origin-level CrUX collection; that qualification is preserved in all subsequent analyses.
2. JavaScript Execution and Third-Party Loading
Section 02Mental Model: "The Single-Lane Highway Bottleneck" & "The Frozen Steering Wheel"
A smartphone processor has only one main thread—a single lane of highway. For the first 13.5 seconds of page load, Overfuel fills that lane with 11.0 seconds of uncoordinated third-party JavaScript freight (chat widgets, finance engines, valuation estimators, tracking tags). When a mobile car buyer attempts to scroll or tap the screen, their actions are frozen for 5,700 milliseconds because the main thread is locked.
The Automotive Vendor Overload
Uncontrolled third-party scripts injected into the critical rendering path without deferred facades:
| Vendor Tool | Typical Weight | Performance Pathology | Required Architecture Fix |
|---|---|---|---|
| Digital Retailing & Finance | 450–900 KiB | Executes payment calculators before shopper taps CTA. | Facade button; load SDK only upon explicit click. |
| Trade-In Valuation Tools | 300–600 KiB | Synchronously bootstraps nested iframe handshakes. | Lazy-load iframe when viewport scrolls within 300px. |
| Live Chat & SMS Widgets | 350–700 KiB | Initiates WebSockets and polling immediately on load. | Static SVG icon; hydrate chat library only on user tap. |
| Marketing & Pixel Beacons | 500–1,200 KiB | Unthrottled tag firings compete for mobile main thread. | Offload to server-side GTM or edge Cloudflare Zaraz. |
| Overfuel Core Framework | 800–1,500 KiB | Monolithic client-side hydration of static page components. | In-place code-splitting within Overfuel's existing build system; tree-shake unused modules and lazy-load interactive components. |
Synchronous execution blocks initial paint before shopper expresses purchase intent.
WebSockets and background audio assets bootstrap immediately on page load.
Heavy iframe handshakes initiate during critical first seconds of rendering.
Ad pixels unthrottled in GTM continuously execute on client devices.
Monolithic client-side hydration of static layout components.
3. Largest Contentful Paint (LCP) and Rendering Critical Path
Section 03Mental Model: "The Empty Showroom Bay"
When an automotive buyer clicks an ad for a $70,000 Mercedes-Benz at ALM, they expect to see the car instantly. Instead, they stare at an empty showroom bay for 10.5 seconds. Overfuel delays the discovery of the primary vehicle photograph by prioritizing render-blocking stylesheets and deferred script evaluation ahead of asset discovery.
Primary LCP Failure Mechanisms
- Render-Blocking Resource Chains: Synchronous CSS files delay initial layout, preventing the browser from initiating hero image requests.
- Late JavaScript Discovery: The lead vehicle image is injected dynamically via client-side JavaScript rather than declared in raw HTML.
- Absence of Priority Hints: The browser treats the vehicle photo as low priority, queuing it behind tracking beacons and widget stylesheets.
Visual Velocity Metrics
- First Contentful Paint (2.1s): Initial paint exceeds Google's 1.8s threshold due to font rendering delays and CSS parsing.
- Speed Index (12.7s): ALM shoppers experience 12.7 seconds of visual churn before the mobile viewport stabilizes.
- Layout Stability (CLS 0.023 vs 0.10): Real-user CLS rests right on the edge of failure due to dynamic promotional banner injection.
Overfuel Engineering Fix: Priority Hinting & Preload Pipeline
The primary hero vehicle image must be preloaded in the document <head> and declared with fetchpriority="high" and loading="eager":
View Mandatory HTML & Preload Implementation ▾
<!-- 1. Document <head> Preload in Overfuel Template -->
<link rel="preload" as="image" href="/images/alm-hero-mobile.webp" fetchpriority="high" />
<!-- 2. Hero Vehicle Image Declaration -->
<img
src="/images/alm-hero-mobile.webp"
srcset="/images/alm-hero-mobile.webp 640w, /images/alm-hero-desktop.webp 1200w"
sizes="(max-width: 768px) 100vw, 1200px"
fetchpriority="high"
loading="eager"
decoding="async"
alt="Atlanta Luxury Motors Featured Vehicle"
width="1200"
height="675" />4. Delivery, Caching, and Interaction Responsiveness
Section 04When an ALM shopper taps a vehicle filter (*"Under $45k"* or *"AWD"*), the interface hesitates for nearly a quarter second before confirming the tap. This delay makes the dealership's digital presence feel loose, unresponsive, and unrefined.
Transferring 4.6 MB over cellular connections forces the shopper’s phone to download the equivalent of three full digital magazine editions just to view a single pre-owned vehicle listing.
Time to First Byte: 1.4s (Target: ≤ 0.5s)
Needs OptimizationA real-world aggregate TTFB of 1.4 seconds indicates that Overfuel is dynamically querying databases and re-rendering HTML on every request. Overfuel must configure full-page edge caching at the CDN layer (e.g. Cloudflare Edge Cache) with automatic cache invalidation upon inventory updates.
5. Target of 100 and Proposed Performance Budgets
Section 05Targeting 100/100: Engineering Ambition vs. Mathematical Reality
ALM has targeted a perfect 100/100 mobile score. While enforcing stringent performance budgets is the only reliable way to push scores into the high 90s, performance budgets alone cannot mathematically guarantee a perpetual 100/100 score. Lighthouse scoring curves are non-linear, and external ad tags routinely induce swings of ±5–10 points. The commercial goal remains: Passing Google's Core Web Vitals in the field.
Engineering Remediation Gaps (Current vs. Enforced Budget)
Visualizing the magnitude of optimization required across Overfuel's codebase:
Full Metric Reference
6. Validation Protocol and Overfuel Deliverables
Section 06Vendor remediation is structured for execution directly by Overfuel’s engineering team within their existing code framework and development workflow. Overfuel retains complete code ownership and platform stewardship; zero replatforming or technology stack migration is required. Progress and deliverables must be verified through a 5-stage milestone roadmap with contractual Go / No-Go sign-off gates:
Stage 1: Production Parity in Staging
Gate: Flow Advisory Sign-offStaging must reflect 100% production parity. Third-party marketing pixels, financing tools, and chat widgets must be loaded via their optimized facade patterns, not artificially removed to inflate test scores.
Stage 2: Repeated 5-Run Lab Benchmarking
Gate: Overfuel Engineering LeadA minimum of five (5) consecutive automated runs across Homepage, Search Results Page (SRP), and Vehicle Detail Page (VDP), discarding outliers and submitting the median report showing Performance ≥ 90 and TBT ≤ 150ms.
Stage 3: End-to-End Functional QA
Gate: ALM Operations & CRM LeadVerify that all lead forms, credit applications, and trade-in calculators transmit cleanly into ALM's CRM (VinSolutions/DealerSocket) with 100% tracking tag fidelity.
Stage 4: Controlled Production Rollout
Gate: Overfuel DevOpsDeploy optimizations behind synthetic performance monitoring (SpeedCurve/Datadog) to verify zero production regressions in live user traffic.
Stage 5: 28-Day Real-User Field Certification
Gate: ALM Executive CommitteeOfficial project acceptance is granted only when the rolling 28-day Chrome User Experience Report (CrUX) confirms that ALM passes Google's Core Web Vitals assessment at the 75th percentile.
7. Secondary Quality Findings and References
Section 07Strikethrough discounts and payment terms must meet WCAG AA 4.5:1 color contrast to prevent legal exposure under ADA digital accessibility guidelines.
Over 1,800 DOM nodes on vehicle listing pages multiplies browser reflow calculations. Pruning component wrappers to < 800 nodes saves 30% layout computation.
Set Cache-Control: public, max-age=31536000, immutable with unique build hashes for all static scripts, stylesheets, and vendor bundles.
Authoritative Performance Documentation
This technical performance audit and remediation architecture was prepared by Flow Web Design for the executive and engineering leadership of Atlanta Luxury Motors (ALM) and Overfuel. All findings, budgets, and protocols remain binding recommendations for platform optimization.