If you're shopping for audit software, close the tab. This isn't a tool pitch. It's not a 100-point landing page audit checklist either.
It's the 15-minute read I do on every page before I touch any data. One page, one timer, your eyes. Same section order I use on paid audits: hero, offer, proof, CTA, friction, structure.
I analyzed 521 SaaS landing pages building pages.report. The pages that fail don't fail in exotic ways. They fail the same five checks, in roughly the same order. So that's the audit.
Here's the definition worth stealing: a landing page audit is a structured review of one page against one question: does anything here stop a stranger from converting? Not "is it pretty." Not "is it on-brand." Can a cold visitor understand it, believe it, and act on it. You can answer that in 15 minutes with no tools, because the biggest conversion killers are visible to the naked eye.
This is the quick version. The full multi-week process, with analytics and testing, lives in the pillar: how to do a CRO audit. Start here first. Most pages don't survive the 15 minutes.
TL;DR
- A landing page audit checks one thing: what stops a stranger from converting on this page.
- You need 15 minutes, a timer, and the page. No tools.
- Minute budget: 0-3 the stranger test, 3-6 offer clarity, 6-9 proof, 9-12 CTA and friction, 12-15 section order.
- Each section gets one pass/fail question. Any fail is your first fix.
- Benchmark for context: median landing page conversion is 6.6%, and SaaS sits at 3.8% (Unbounce Conversion Benchmark Report, 41K+ pages, Q4 2024).
- What 15 minutes can't see: your actual visitors' behavior. That needs analytics.
Set a timer. Open the page in an incognito window so you see what a stranger sees. Go.
Minutes 0-3: The five-second stranger test
Look at the hero only. Nothing below the fold. Give yourself five seconds, then look away and answer three questions out loud:
- What is this product?
- Who is it for?
- What happens if I click the button?
Pass/fail question: could a stranger answer all three without scrolling?
If you built the page, you can't grade this yourself. You know too much. Grab anyone who's never seen the page, give them five seconds, and ask. Their blank stare is the audit result.
The most common failure I see: heroes that describe a category instead of a product. "The modern platform for teams." That headline fits 400 other companies. It's the miss I run into most often, and everything below a vague hero is wasted scroll.
Fail here and stop the audit. Fix the hero first. Nothing else matters until a stranger knows what they're looking at. If you want to see what passing looks like, the annotated pages in our teardowns all clear this in one sentence.
Minutes 3-6: Offer and pricing clarity
Now scroll until you can answer: what exactly am I getting, and what does it cost me? Money, time, or data, whatever the page is asking for.
Pass/fail question: after one scroll, do I know the offer and the price of saying yes?
For context on what's normal, Unbounce's Conversion Benchmark Report (41K+ pages, Q4 2024) puts median landing page conversion at 6.6% across industries, with SaaS at 3.8%. Most pages I audit that sit well under those numbers share one trait: the visitor can't tell what saying yes actually commits them to.
The most common failure: hedged offers. "Get started" with no price anywhere. "Book a demo" when the visitor wanted a trial. A free plan buried in a pricing footnote. When the cost of yes is unclear, the visitor assumes the worst and leaves.
Ask one more thing here. Does the offer match where the traffic comes from? If the ad promised a template and the page sells a subscription, no design fixes that.
Minutes 6-9: Proof and trust
Scroll the whole page and count the evidence. Not the claims, the evidence. Named customers. Numbers with sources. Testimonials with a face, a company, and an outcome. Screenshots of the real product.
Pass/fail question: does the page prove anything, or does it only assert?
Trust isn't decoration; it's measurable money. In Baymard Institute's running cart abandonment study, 19% of users abandoned a checkout because they didn't trust the site with card information. That's checkout data, but the mechanism is identical on a landing page. A stranger who doesn't trust you doesn't convert, and they don't file a complaint on the way out.
The most common failure: fake-feeling proof. "Loved by 10,000+ users" with no logos. "Game changing! - J.S." with no face and no company. That's worse than nothing, because it signals you couldn't get anyone real to vouch for you. Pull it.
Grade honestly. One specific case study beats a wall of anonymous quotes. The pages in our high-converting landing page examples post win this section with numbers, not adjectives.
Minutes 9-12: CTA and friction
Find every button and link on the page. Count them. Then walk the primary path: click the main CTA and note every single thing you're asked for before value shows up.
Pass/fail question: is there one clear action, and does it ask only for what it's earned?
This check is personal. When I launched pages.report, conversion was bad. I spent four weeks rewriting the hero, headlines, colors, screenshots. Nothing moved. Then I changed one button: magic-link signup became "Continue with Google." Login rate tripled. The page was fine. The ask was too expensive.
The data backs the pattern. In Baymard's study, 18% of users abandoned because the site forced account creation. Every field, every step, every premature ask is a toll booth.
The most common failure: competing CTAs. "Start free trial" next to "Book a demo" next to "Watch video" next to a nav bar with nine links. Split intent converts nobody.
One friction check your eyes can do without software: load the page on your phone, on cellular, and count seconds. Portent's 2022 research found sites converting at 3.05% at a 1-second load versus 0.67% at 4 seconds. Older study, but if you're counting past three, you already know.
Minutes 12-15: Section order and message match
Last pass. Scroll top to bottom in one motion and read only the headlines of each section. That sequence is the argument your page makes. Does it build?
The order I keep coming back to: what it is, what I get, why I should believe you, what to do next. Problem, outcome, proof, action.
Pass/fail question: do the section headlines alone tell a story that ends at the CTA?
The most common failure: pages assembled instead of argued. A feature grid before anyone knows what the product is. Pricing before proof. An FAQ answering questions nobody asked yet. Each section fine alone, in an order that persuades no one.
Also recheck message match here, end to end. The headline made a promise in minute zero. Does the closing CTA pay it off in the same words? If the hero says "audit," the button shouldn't say "demo."
Time's up. Count your fails. Fix them in walkthrough order, hero first, because each section only works if the one above it did its job. If you want the exhaustive version with all 30 checks, that's the CRO audit checklist.
When 15 minutes isn't enough
This audit finds what a stranger can see. It cannot find what your visitors actually do. Where they drop, what they scroll past, which traffic source bounces, whether mobile behaves differently. That's a data problem, and eyes don't solve data problems. When you hit that wall, you need analytics and session recordings: here's the CRO audit tool stack I'd start with.
One warning before you change anything based on gut feel. Microsoft's experimentation team reported in 2013 that less than a third of tested ideas moved their target metrics. Old number, still humbling. The 15-minute audit tells you where to look. It doesn't guarantee your fix is the right one.
And if you'd rather have a second set of trained eyes on your page, that's literally what I sell: a $99 landing page audit, delivered as a video teardown in 48 hours. Same walkthrough as above, plus 521 pages of pattern memory behind it.
Set the timer. Fifteen minutes from now you'll know your first fix.
– Luke
