alloy digital
Thompson Machinery — Current Site vs. Concept Redesign
Honesty guardrail. Left = the measured problems behind each pin. Right = concept art (illustrative, not a live build). Lead with measured findings only — never claim the site is slow without field data to back it.
🔒 thompsonmachinery.com
Current state · measured1
2
3
Live screenshot — Thompson Machinery's current homepage, captured at desktop resolution (2880×2000px). Click each numbered pin for the measured finding.
CrUX field bucket is SLOW Measured
Over the trailing 28-day window, real visitors land in the worst-performing CrUX tier — the page underdelivers for the actual customers measured, not just a lab simulation (measured via PageSpeed Insights / CrUX, 2026-06-19).
Source: Thompson Machinery — Discovery Findings.md
Slow Largest Contentful Paint (LCP 3.2s) Measured
Real-user LCP measures 3.2s — beyond the 2.5s "good" threshold — so the main hero content is slow to appear for the visitors measured (measured via PageSpeed Insights / CrUX, 2026-06-19).
Source: Thompson Machinery — Discovery Findings.md
Desktop layout shift (CLS 0.36) Measured
Desktop Cumulative Layout Shift measures 0.36 — above the 0.25 "poor" threshold — so content jumps around as the page loads and buyers mis-tap or lose their place (measured via PageSpeed Insights / CrUX, 2026-06-19).
Source: Thompson Machinery — Discovery Findings.md
🔒 thompsonmachinery.com
Concept art · illustrativeConcept direction
One unified site — built around how buyers actually shop.
An illustrative redesign that resolves each measured finding below. Concept art — not a live build or quoted scope.
Findings, before → after
Finding 01
CrUX field bucket is SLOW
NowOver the trailing 28-day window, real visitors land in the worst-performing CrUX tier — the page underdelivers for the actual customers measured, not just a lab simulation (measured via PageSpeed Insights / CrUX, 2026-06-19).
ConceptAlloy direction: a performance-first rebuild that moves the site out of the SLOW bucket toward Core Web Vitals "Good" on real devices. (concept — illustrative).
Measured: Thompson Machinery — Discovery Findings.md
Finding 02
Slow Largest Contentful Paint (LCP 3.2s)
NowReal-user LCP measures 3.2s — beyond the 2.5s "good" threshold — so the main hero content is slow to appear for the visitors measured (measured via PageSpeed Insights / CrUX, 2026-06-19).
ConceptAlloy direction: optimize the hero image, fonts, and render path so the largest content paints under 2.5s on real connections. (concept — illustrative).
Measured: Thompson Machinery — Discovery Findings.md
Finding 03
Desktop layout shift (CLS 0.36)
NowDesktop Cumulative Layout Shift measures 0.36 — above the 0.25 "poor" threshold — so content jumps around as the page loads and buyers mis-tap or lose their place (measured via PageSpeed Insights / CrUX, 2026-06-19).
ConceptAlloy direction: reserve space for images, fonts, and embeds so the layout locks in place on first paint and never shifts under the user. (concept — illustrative).
Measured: Thompson Machinery — Discovery Findings.md