A two-week sprint so the call-now button loads before the caller gives up.
A tow site that takes 4.2 seconds on a 4G shoulder is a tow site the caller never saw. The sprint cuts LCP under 1.6s, INP under 200ms, CLS under 0.1, and moves the call-now button into the above-the-fold HTML so it is tappable before any JavaScript runs.
Hudson Valley Towing, before and after.
Fictional operator, representative numbers. The site shipped with a 3.2 MB hero image, three untreated jQuery plugins, and a chat widget that blocked the main thread for 1,180ms. Two weeks later: 220 KB AVIF hero, inlined critical CSS, chat widget deferred to idle. The call-now button went from paintable at 3.6s to paintable at 0.9s.
Core Web Vitals · field data
Core Web Vitals · field data
Call-now time to tap: 3.6s → 0.9s
Illustrative mockup · fictional operator · your report uses your field data in CrUX
Everything in the two-week sprint.
- Day 1 audit: Lighthouse mobile 4G, CrUX field pull, synthetic WebPageTest run
- Hero and gallery images converted to AVIF with WebP fallback, srcset per breakpoint
- Critical CSS inlined for the above-the-fold path on every service page
- Third-party scripts quarantined: chat widget, pixel, review feed deferred to idle
- Call-now button lifted into the above-the-fold HTML so it is tappable before JS executes
- Font subsetting and self-hosting for the two display weights actually used
- Server config: Brotli compression, HTTP/2, cache-control headers, Early Hints where supported
- Performance budget committed to the repo with a GitHub Action that fails deploys over 1.6s LCP
- Before/after Lighthouse report plus a 90-day CrUX trend monitor
- Field testing from a real phone on 4G at a real yard, not from a dev laptop
Google's 2.5s LCP threshold is a passing grade, not a competitive one.
A shoulder caller gives you about 11 seconds. If your call-now button paints at 3.6s, every one of your competitors who paints at 1.1s wins the call before the caller sees your logo.
LCP under 1.6s is the real tow-site target
Google tells you 2.5s is "good." But the median tow landing page in a competitive metro paints at 1.2 to 1.6s. The paid-traffic landing pages we ship hit 1.4s. Pass by Google's rules, lose by the market's. Operators who moved from a passing 2.3s to a 1.4s page saw call-form conversions climb 24% without changing a single word of copy.
INP is the hidden killer on form-fill scenes
The caller taps the equipment-class radio and nothing happens for 540ms because the thread is busy. They tap again, double-firing the handler, and the form state goes sideways. INP under 200ms is why the intake submits cleanly on a Pixel 6 at 2:14am. Fixing INP is mostly deferred scripts, chunked tasks, and taking the chat widget off the main thread.
CLS wrecks the thumb-tap at the worst moment
Your hero image loads, the layout shifts 120px, and the thumb that was aimed at call-now now hits "see our equipment." The caller taps away. CLS under 0.1 is a width/height on every image, aspect-ratio boxes on every embedded element, and font-display:optional on display weights. Boring work. Big difference. Pair this with the scene intake form and the whole hand-off stops failing on cheap Androids.
What dispatchers ask before signing the sprint.
What Lighthouse scores do you target for tow operators?
Performance 90+, Accessibility 95+, Best Practices 100, SEO 100. On the field metrics: LCP under 1.6s, INP under 200ms, CLS under 0.1 on a mid-tier Android over 4G.
Why is 1.6s the target instead of Google's 2.5s?
A shoulder caller gives you about 11 seconds before tapping the next result. Google's 2.5s LCP threshold is a passing grade, not a competitive one. A tow page that loads in 1.6s wins calls a 2.4s page loses, even though both pass.
How long does the sprint take?
Two weeks. Week 1 is audit, image conversion, critical-CSS inlining, and third-party script quarantine. Week 2 is server config (Brotli, HTTP/2, cache headers), font subsetting, and the before-after Lighthouse report.
Can you fix a site built on a tow-specific CMS?
Yes, with caveats. We have worked on Towbook Marketing, DRN, and Autura's hosted sites. Server-side fixes depend on what the CMS exposes; if it does not let us set cache headers, we document the ceiling and prioritize the client-side wins.
Do the gains survive a theme update?
Yes, if your theme is version-pinned. The sprint adds a performance budget to the repo and a GitHub Action that fails the deploy if LCP regresses over 1.6s. On hosted CMSes we hand you a 10-item checklist to run after any theme update.
What about mobile call-now button placement?
The sprint moves the call-now button into the above-the-fold HTML so it is tappable before any JavaScript executes. On most tow sites this is the single biggest conversion fix of the engagement.
Services that pair well with the sprint.
Sprint it a la carte or bundle it with a plan.
$495 per month, two-week opening sprint. Bundled with Live Transfer or Dispatch Ready on the tow operator lane, the sprint runs on the subscription site at no additional charge.