The site is working against the parents it’s trying to win.

We didn’t rely on opinion. The audit triangulates three lenses, and every external claim is linked so you can check our work.

1. Peep Laja’s CXL heuristics. We walked every screen through the entry point of the ResearchXL framework — scoring Clarity, Relevance, Friction and Distraction. Laja’s core law: “clarity trumps persuasion,” and a value proposition must be understood in ~5 seconds.

2. Modern UX & typography research — Nielsen Norman Group, Baymard Institute, the WCAG guidelines and Butterick’s Practical Typography.

3. First-party measurement. Real Google Lighthouse runs, and contrast ratios computed from actual pixels sampled out of the live screenshots — not eyeballed.

A visitor can’t answer “what is this and why choose it” in five seconds.

The hero spends its most valuable real estate on decoration instead of a claim. The differentiator that should be impossible to miss — true Spanish immersion, ages served, two locations, what it does for your child — is never stated plainly above the fold.

Users form a visual first impression in ~50ms, and that impression sets a credibility “halo” for everything after it. A low-complexity, conventional-looking design wins that judgement; a busy, unfamiliar one loses it.

Sources: CXL — First impressions · CXL — Value proposition & the 5-second test · NN/g — be specific, not generic

color is doing a job only shape and convention can do.

People know “a strongly colored stripe across the top is the nav” by convention. These blocks break the convention: they read as decorative banners, so a meaningful share of visitors won’t realise the site’s sections live inside them. NN/g’s eye-tracking found weak clickability signifiers make users work measurably harder.

It also fails accessibility outright: WCAG 1.4.1 Use of Color forbids color as the only way to signal an action — and the red and green blocks sit next to each other, the single most common form of color-blindness.

When everything is loud, the button can’t shout.

Good hierarchy comes from contrast in value and saturation, not from more colors — “if everything is contrasted, nothing stands out” (NN/g). The page saturates every band, so the dark-navy CTA pills are actually quieter than their surroundings — the accent rule run backwards. The discipline to aim for is the 60-30-10 rule: a dominant neutral, a secondary, and a single 10% accent saved for the primary action.

And these aren’t just stylistic gripes. We sampled the real pixels and ran the WCAG contrast maths:

Whitespace isn’t empty — it’s what makes the rest legible.

Full-bleed saturated backgrounds run edge to edge with content packed against them. NN/g calls whitespace the negative space that “strengthens figure-ground separation” and lets elements be seen; CXL calls adding whitespace the “golden rule of simplicity.” The repeating wave divider and stacked color bands do the opposite — they raise the mental effort needed to process every screen, and high cognitive load measurably costs conversions.

This very page is the counter-example: one typeface, a calm neutral background, a capped reading column, and space around each idea. It isn’t harder to build — it’s a decision.

Sources: NN/g — Whitespace · NN/g — Visual hierarchy

~130 characters per line is about double the readable limit.

The evidence is unusually unanimous here. Baymard puts the ideal measure at 50–75 characters; Butterick says 45–90; and WCAG 1.4.8 caps blocks at 80. The articles here run past all three. Combine that with three loaded display fonts — Montserrat, Roboto Slab and ABeeZee, plus Font Awesome — and the page asks readers to parse several typographic voices at once.

It matters because people barely read online to begin with: on an average visit users read at most ~28% of the words, and 79% scan rather than read. Over-wide, undifferentiated text means the scanned fraction misses the message.

Sources: Baymard — Line length · Butterick · NN/g — How little users read · NN/g — Scanning

A 12.7-second mobile LCP loses roughly half the visitors before they see a thing.

Parents research daycare on their phones. Our Lighthouse run clocked the homepage’s Largest Contentful Paint at 12.7s on mobile — Google’s “good” bar is 2.5s, and anything over 4s is “poor.” Google’s own data: 53% of mobile visitors abandon a page that takes over 3s, and bounce probability climbs 123% as load goes from 1s to 10s.

The cause is page weight, not the network. The site is built on WordPress + Beaver Builder, which ships 18 stylesheets, 25+ scripts (885 KB of JS), three font families and Font Awesome, and 82 “!important” overrides — with 292 KB of JavaScript and 70 KB of CSS never used on the page.

Sources: web.dev — LCP · Google/SOASTA — 53% · Think with Google · Portent

The desktop compromises get worse on a small screen.

The four color blocks stack into a tall wall the visitor must scroll past before reaching the actual pitch and the buttons. The amber-on-blue hero copy (right) is the same 3.57:1 contrast that fails WCAG for normal text — harder still on a sunlit phone. And every extra second of that 12.7s load disproportionately raises the odds a parent leaves first.