I audited 50 funded startup websites. Seven patterns showed up almost everywhere.

The methodology

I reviewed 50 funded startup websites between mid-May and early June 2026. Selection was pseudo-random: 20 from recent YC batches (S24–W26), 20 from Series A/B rounds announced in 2025, and 10 from the Series C+ tier. Categories spanned SaaS, AI tools, fintech, dev tools, marketplaces, and B2B. Average funding raised: $14M.

For each site I scored 12 dimensions: hero clarity, headline quality, social proof, pricing visibility, page speed, mobile responsiveness, product visibility, CTA clarity, copy reading level, team page quality, trust signals, and content depth. I spent about 8–12 minutes per site, the same skim a serious customer or investor would do.

What follows are the seven patterns that showed up on more than 50% of sites, most on more than 70%.

Headline numbers

Before the patterns, the aggregate scoreboard.

Metric	Result
Sites with a generic/vague headline	38 / 50 (76%)
Sites with no product visible in the hero	31 / 50 (62%)
Sites with 3+ competing CTAs above the fold	27 / 50 (54%)
Sites that hid pricing entirely	34 / 50 (68%)
Sites with mobile load time > 3 seconds	19 / 50 (38%)
Sites with copy at college reading level or above	42 / 50 (84%)
Sites with no named customer logos in first scroll	29 / 50 (58%)
Sites with a team page using stock photos or no photos	23 / 50 (46%)
Sites with a typo or broken capitalization on the homepage	18 / 50 (36%)
Sites I would describe as "investor-ready" without changes	6 / 50 (12%)

The top-line takeaway: only 12% of funded startup sites are ready for investor scrutiny without changes. These are companies with millions in the bank and engineering teams shipping product. The website is the most-neglected asset on the balance sheet.

Pattern 1: The vague headline epidemic (76% of sites)

Three in four sites led with a headline that could belong to any of the others. Examples from the audit (lightly anonymized):

• "The intelligent platform for modern teams."
• "Workflow automation, reimagined."
• "Where data meets decisions."
• "Smarter collaboration for the AI era."
• "The unified solution for enterprise growth."

Each one fails the basic test: a person from outside the industry can't tell what the company does. The audience tells you nothing. The mechanism tells you nothing. The outcome tells you nothing.

What the 12 sites that did this well had in common. They named either the audience or the mechanism in the first sentence. Linear's "The product development system for teams and agents" names the category and audience. Loom's "One video is worth a thousand words" names the outcome. The good ones all passed the 5-second test on someone outside the industry.

Pattern 2: No product, just vibes (62% of sites)

Almost two-thirds of sites had no actual product visible above the fold. Stock illustrations of diverse people at laptops, abstract gradients with floating UI elements, hero videos of generic office footage, or pure-text heroes with nothing visual at all.

This is the biggest miss across the audit. The product is the most compelling asset most startups have, and most are hiding it.

What the 19 sites that did this well did. Real screenshots of the actual product doing its actual main thing. Linear shows actual Linear issues. Notion shows actual Notion pages. Vercel renders live AI model telemetry in the hero. The pattern: the homepage is the product demo.

Pattern 3: Decision-fatigue CTAs (54% of sites)

A typical hero in the audit had: "Start free trial," "Book a demo," "Watch video," "Talk to sales," and "Learn more", five CTAs competing for one click.

Decision fatigue triggers immediately. The visitor's path of least resistance becomes the back button. CXL's research on this is unambiguous: multiple CTAs reduce action across every tested vertical.

What the 23 sites that did this well did. One primary CTA in the filled-button style. One secondary CTA in outline or text-link style. That's it. Stripe: "Sign up" and "Sign up with Google" (same action, two methods). Linear: "Start building" with a small "Pricing" link. Maximum two CTAs, with one obviously the primary.

Pattern 4: Hidden pricing (68% of sites)

This is the one that surprised me most. Funded startups with mature products and real ARR are still hiding pricing on 68% of homepages.

The justification I hear from founders: "Our pricing is custom" or "We don't want competitors seeing it." The cost: 62% of B2B buyers disqualify hidden-pricing vendors (TrustRadius 2025). For every founder protecting against competitor visibility, two-thirds of their qualified buyers are walking away.

What the 16 sites that did this well did. Three patterns. Some published full pricing tiers (Linear, Stripe, Vercel). Some published a starting-at anchor with custom enterprise pricing above ("Starting at $499/mo, contact sales for enterprise"). Some put pricing entirely on a separate page but linked it from the nav, the key was that the visitor could find a number in under 30 seconds.

Pattern 5: Copy that reads like a 10-K filing (84% of sites)

The worst pattern by frequency. Forty-two of 50 sites had homepage copy at college reading level or above. Hemingway scored most heroes at grade 11–14.

Real samples (lightly anonymized):

"Our enterprise-grade platform leverages proprietary machine learning to optimize workflow efficiency across distributed organizational structures, enabling seamless cross-functional collaboration and accelerated time-to-value."

"A robust, scalable solution that empowers data-driven decision-making through unified observability and intelligent automation."

Unbounce's data: pages at 5th–7th grade copy convert at 11.1% vs 5.3% at professional level. Difficult words alone correlate with a 24.3% conversion drop. This is the single biggest mechanical lift available to most SaaS sites.

What the 8 sites that did this well did. Strip every adjective. Use short sentences. Lead with verbs. Use words from a 5,000-word working vocabulary. Linear's "Plan and build products" is grade 4. Loom's "One video is worth a thousand words" is grade 4. Stripe's "Accept payments" is grade 3.

Pattern 6: Social proof either absent or generic (58% of sites)

Twenty-nine sites had no named customer logos in the first scroll. Either there was no logo bar at all, or there was a generic "Trusted by industry leaders" without actual logos.

The most common pattern: the logo bar exists but is buried in the footer or in a "Customers" section midway down. CXL eye-tracking shows that social proof works near a CTA or at a point of friction, not at the bottom of the page where most visitors never reach.

What the 21 sites that did this well did. Named logos within the first scroll. Photo testimonials with full name, role, and company. Specific stats (Notion: "Over 100M users worldwide • 62% of Fortune 100"). Some used Vercel-style inline single-sentence social proof: "Runway build times went from 7m to 40s. LeonardoAi saw a 95% reduction in page load times."

The compounding pattern

The seven mistakes aren't independent. A site with a vague headline + no product visible + 5 competing CTAs + hidden pricing isn't seven times worse than a site with one of those issues, it's compoundingly worse. Each one erodes trust further, and visitors don't tolerate cumulative friction. Fix two of the seven and conversion roughly doubles in most cases.

Pattern 7: Mobile is an afterthought (38% of sites)

Nineteen of 50 sites had mobile load times over 3 seconds. Twelve had layout breaks on iPhone, overflow scrolling, unreadable text, broken nav. Three were essentially unusable on mobile.

This is genuinely surprising in 2026. Mobile is 79–82.9% of SaaS landing-page traffic (Unbounce). Every founder knows this. Most haven't actually opened their own site on a phone in the last six months.

What the 31 sites that did this well did. They tested mobile as the primary experience. The hero worked at iPhone width. The CTAs were thumb-sized. The page loaded in under 2 seconds on a throttled 4G connection. Most were built on Next.js + Vercel or Webflow, both of which default to good performance.

The top 5 sites I audited and what they did right

Without naming companies (some were Plinth pipeline), here's what the highest-scoring sites had in common.

1. Hero headline named the audience and outcome in 8 words or less. Every top-5 site passed the 5-second test cold.
2. Product visible in the hero. Either a real screenshot, an interactive demo, or an autoplaying product video.
3. Pricing visible from the nav. Every single one. Even the enterprise-focused sites had a "Starting at..." anchor.
4. Named customer logos in the first scroll. Average 5–7 logos, all recognizable.
5. One primary CTA + one secondary. Never more.
6. Mobile load under 2 seconds. Tested across iPhone, Android, and tablet.
7. Built on a modern stack. Three on Next.js + Vercel, two on Webflow. None on WordPress.

The fastest fixes you can ship this week

Rank-ordered by impact:

1. Rewrite the hero headline. 1 hour. Highest impact of any single change.
2. Replace stock illustration with a real product screenshot. 2 hours. Second highest.
3. Cut CTAs to one primary + one secondary. 30 minutes. Almost no cost, big lift.
4. Add a "Starting at..." price anchor in the hero or nav. 1 hour. Removes the 62%-disqualify problem.
5. Run copy through Hemingway, aim for grade 6 or below. 2 hours. Compounds with everything else.
6. Move customer logos above the fold. 30 minutes if they exist. Day one priority if they don't.
7. Open the site on your own phone. Fix what's broken. 1–4 hours depending on scope.

A focused founder with a half day can address 5 of the 7 patterns. That alone usually moves conversion from "below median" to "above median" for SaaS, i.e., past 3.8%.

What this audit told me about the market

Three meta-observations from looking at 50 sites in a row:

1. The funding gap doesn't predict website quality. Several Series B companies had worse sites than YC pre-seed startups. Money buys engineering. It doesn't buy good positioning or design taste.

2. AI startups have the worst homepages on average. The fastest-growing category in tech is also the category most often shipping placeholder sites with stock illustrations and "powered by AI" headlines. There's a window here for AI startups who actually invest in their homepage.

3. The same five patterns keep working. Real product visible. Clear positioning. Pricing visible. One CTA. Mobile-first. Every high-converting site does these. None of it is novel. Most founders just haven't applied it to their own homepage.