PageFriction

See where your landing page loses people.

PageFriction measures your page section by section, so you can see where visitors stop and decide what to test first.

Find out what to test first on each landing page, and keep a record of what you already tried.

No sign-up for the URL analysis. Measuring real visitors starts with 14 days free, card required.

This is the view you get

Your own page, with your own sections, and one mark where people leave. Switch the device to see how the reading changes.

example-dental.example/free-assessment
  1. A new smile starts with one visit

    Free assessment with our dentists this week.

    Hero 100% reached this sectionNothing unusual here 100% reached this sectionNothing unusual here
  2. Why patients choose us

    Same-day care, clear prices, parking on site.

    Benefits 74% reached this sectionNothing unusual here 86% reached this sectionNothing unusual here
  3. Before and after

    Treatments explained step by step.

    Results 61% reached this sectionNothing unusual here 77% reached this sectionNothing unusual here
  4. What patients say

    Twelve long reviews, in a row.

    Testimonials 52% reached this sectionMost exits on this page31% leave here, against 18% in nearby sections. 71% reached this sectionNothing unusual here
  5. Plans and payment

    Three ways to pay for your treatment.

    Prices 36% reached this sectionNothing unusual here 58% reached this sectionNothing unusual here
  6. Book your free assessment

    Name, phone and preferred time.

    Form 29% reached this sectionCollecting data 49% reached this sectionCollecting data

Fictional page and numbers. In the product, each row shows the real image of your section.

One loop, for every page

  1. Find

    Paste a URL for first hypotheses from how the page looks. Install the tag, and real visitors show which ones match what happens.

  2. Test

    Take one suggested change, with its evidence and what it does not prove. Make it on your page and log the date.

  3. See

    Compare before and after, with the rest of the page beside it, and keep a record of everything you tried.

Everything in the product

All of it is in every plan.

Measurement

What the tag records on your page.

  • One script tag

    One line added to your site, on any page builder or custom code.

  • Automatic sections

    Finds up to 40 blocks per page, including Elementor and Shopify layouts.

  • Reach by section

    Which sections each visit saw, and how long it took to get there.

  • Attention by section

    Time on each section, counted only while the tab is in focus and the person is active.

  • Clicks by section

    Button, WhatsApp and phone clicks, tied to the section they happened in. A WhatsApp click counts as intent, not as a lead.

  • Form starts and sends

    Who began a form and who pressed send. Search and login forms stay out.

  • Sources and devices

    UTMs, click IDs from the main ad platforms, referring sites, and phone, tablet or desktop.

  • Goals and purchases

    Send your own goals from the page. Purchases stay in the page record, never credited to a section.

Diagnosis

What turns measurement into a decision.

  • Instant URL analysis

    Paste a link and get up to three hypotheses marked on the real sections, in about a minute, before any traffic.

  • Section map

    Confirm, merge, split or rename sections. Corrections apply to past data too.

  • Section reading

    Green, red or gray for each section, by device. Gray until there is enough data.

  • Do this first

    One action at a time, instead of a wall of charts.

  • See why

    The number, the sample, the hypothesis, what it does not prove and how to test it.

  • Filters

    Read any page by source, device and period.

  • Form field drop-off

    Which field people give up on, by field name and never by what they typed.

  • Technical friction

    Slow loads, page errors and problems by browser, including in-app browsers.

  • Ad-to-page reading

    How far the visitors of each campaign and ad read.

Testing

What keeps the work from being lost.

  • Test log

    What changed, when, and on which page.

  • Automatic change detection

    Notices when a page changed and marks the date for you.

  • Before and after

    Two full weeks on each side, labeled as not a controlled test, with untouched sections beside it. Inconclusive is a normal answer.

  • Page goal

    Leads, clicks in a region or reading to a point. The goal decides what counts first.

For agencies

What makes it work across clients.

  • Accounts and agencies

    Create an agency, add client sites and get a tag for each one.

  • Client overview

    The state of every site and the next action, on one screen.

  • Client report link

    A page to share with your client, without giving them access.

  • Weekly summary

    One short message with what to do next.

  • Active alerts

    Page down, form broken or tag missing.

Privacy and safety

How it behaves on your client’s site.

  • Consent first

    Nothing is stored or sent until the visitor accepts. Refusing clears the identifiers.

  • Never reads what people type

    It sees that a form started, not what was written.

  • Leaves your page alone

    It does not change the page, the forms or the dataLayer.

  • Light

    About 5.3 KB compressed, with a test that fails above 6 KB.

  • Cleaner data

    Removes IDs from URLs, filters bots and accepts only your approved domains.

  • 12-month retention

    Raw events are deleted after 12 months.

  • Site ownership check

    Proves a site is yours before its visitor data appears.

What it will never tell you

One version of a page cannot prove what caused a result. So the product says what the data supports, and stops there.

We will not sayWe will say
“This section sells”“This is where most people stop reading”
“This section converts 9%”“42% of visitors reached this section”
“The test worked”“Before and after. Traffic changed too, so this is not a controlled test.”
“Our AI found what sells”“A suggestion based on how the page looks. Test it first.”

What it runs on

The product runs on Cloudflare. The tag sends events to its own collector, and visitor data never mixes between accounts.

Site, dashboard and collector
Cloudflare Workers
Accounts and sites
Cloudflare D1
Visitor events
Cloudflare R2
Page analysis
Cloudflare Browser Run and Workers AI
Payments
Stripe

Priced by client site

One table in US dollars. All plans start with 14 days free, and a card is required.

  • Solo

    For an independent media buyer

    $29/month

    1 site

  • Agency Plus

    For a larger agency

    $249/month

    30 sites

Every plan includes all the features. Each one has an event allowance, with a warning before you reach it. Your past data is never locked.

Questions

What can I do without an account?

Analyze a public page. We open it on a phone and on a computer and mark up to three hypotheses on its sections. Without an account the result is kept for 30 days.

Is this a heatmap?

No. It works per section and ends in one suggested test, instead of asking you to read colors on a screen. There are no session recordings.

Does it replace Google Analytics?

Not on its own. It answers a narrower question: what is happening inside this page, and what to test next.

How much traffic do I need?

The URL analysis needs none. For behavior, a section stays gray until about 300 visits reach it on each device. That threshold may change.

Does it work on my site?

The tag is one script you add to your site. We tested section detection on 87 public pages across Elementor, WordPress, Astro, React, plain HTML, Shopify, GreatPages, Leadlovers, Hotmart Pages and RD Station, and read 83 of them on both devices. Sites behind a bot wall cannot be read by URL, and the screen says so. The tag works there all the same.

Will the tag slow my page?

It is about 5.3 KB compressed and loads without blocking the page. It records nothing until the visitor accepts.

Who proves the site is mine?

You do, with a meta tag or a DNS record that our server checks. Behavior data only appears after that.

Start with one page.

Paste a URL and see the first hypotheses in about a minute.

Analyze a page for free