What your site is actually doing.
One page: what is working, what is costing you visitors, and what to fix next. A phone-and-desktop teardown of the live site: 4 phone loads + 2 desktop loads. No login, no insider access, no Harvv pixel needed. The full evidence is at the bottom.
Working
What's already working. These held up across every load.
Speed is good.
The page is about 0.1 MB across 42 requests. That keeps it quick on mobile data and cheap to load repeatedly.
No JavaScript errors on load.
Lighthouse scores SEO 100/100. The fundamentals Google looks for are present.
Nothing spilled past the edge at either 390px (phone) or 1366px (desktop), so the structure is responsive.
Costing you
Ranked by what hurts conversion most. Full evidence below.
Tiny buttons are hard to tap on mobile
A floating widget is covering a clickable element
Add quotations so AI engines cite this page
Broken internal links
1 form field has no label
1 input uses the wrong type attribute
3 small changes, on the pages people actually land on: Tiny buttons are hard to tap on mobile; A floating widget is covering a clickable element; Images have no width or height set (layout is stable for now). We re-measure the same samples after.
This is one of several AI-built sites we tore down the same way. The same friction repeats on Reversely, and we pulled the cross-site pattern together in the vibe-code friction tax.
01Findings, ranked by what hurts conversion most
| Severity | Finding | How we know |
|---|---|---|
| High | Tiny buttons are hard to tap on mobileMobileAccessibility (WCAG)Conversion 16 of 25 tappable items on this page (64%) measure under 24 pixels on their shorter side, below the 24px floor WCAG 2.5.8 sets for accessibility and well under the platform minimums (Apple recommends 44pt, Android 48dp) for reliable tapping, and the same ones came up small on every test load. When visitors can't hit what they expect to, they get frustrated and many of them leave instead of trying again. The exact elements we found: paste into Claude, Cursor, or ChatGPT | |
| High | A floating widget is covering a clickable elementBothConversionAccessibility (WCAG) A clickable element is sitting underneath a fixed overlay (nav.container over a "Products"), so a visitor who tries to tap it hits the widget on top instead. The element is there and looks fine in a preview, it just can't be reached. Move the floating widget to a clear corner, lower its z-index, or add spacing so it never overlaps your links and buttons. The exact elements we found: paste into Claude, Cursor, or ChatGPT | |
| Low | Images have no width or height set (layout is stable for now)BothPerformanceSEO 6 of 7 images don't declare width and height. Your layout currently stays stable (measured Cumulative Layout Shift is 0, which is fine), most likely because space is reserved another way (CSS), so this is not causing visible jumps today. It is still worth setting explicit dimensions or a CSS aspect-ratio so a slow connection or a future style change cannot reintroduce shift. The exact images we found: paste into Claude, Cursor, or ChatGPT | |
| Low | Unused JavaScript is being downloadedBothPerformance Code that never runs on this page still costs download and parse time on every visit. Splitting or removing it speeds up load. Lighthouse measured: Est savings of 139 KiB. paste into Claude, Cursor, or ChatGPT | |
| Medium | 1 form field has no labelBothAccessibility (WCAG)ConversionTracking Screen readers can't announce these fields, and a sighted user who clears the placeholder can't recover the prompt. Wrap each input in <label>…</label> or add aria-label. The exact elements we found: paste into Claude, Cursor, or ChatGPT | |
| High | Add quotations so AI engines cite this pageBothAI SearchSEO Generative engines (ChatGPT, Perplexity, AI Overviews) lift sourced, attributed quotes almost verbatim, and quotations are the single strongest citation lever (studies measure roughly +41%). Add 1-2 attributed expert quotes or blockquotes to the pages below. (Found across a sample of 8 pages from your sitemap, a partial crawl rather than your full site.) paste into Claude, Cursor, or ChatGPT | |
| Low | No advertising pixel detectedBothTracking Analytics is present but no Meta/Google/TikTok ad pixel was found. If you run paid ads, conversion tracking should be installed before the next campaign so the platforms can optimize toward buyers, not clicks. paste into Claude, Cursor, or ChatGPT | |
| High | Broken internal linksBothSEOConversion 1 internal link target returned an error (404/410/5xx) or did not respond. Broken links waste crawl budget, frustrate visitors, and leak ranking signal. (Found across a sample of 8 pages from your sitemap, a partial crawl rather than your full site.) The exact examples we found:
paste into Claude, Cursor, or ChatGPT | |
| Medium | 1 input uses the wrong type attributeBothConversion Mobile keyboards adapt to the input type (email shows @ key, tel shows numeric pad). Generic type="text" gets a generic keyboard and visitors mis-type. Switch to the correct type, or add a matching inputmode attribute if the native validation is too strict for what people type (bare domains in a URL field, for example). The exact elements we found: paste into Claude, Cursor, or ChatGPT | |
| Low | No llms.txt fileBothSEO No /llms.txt. Optional: some AI tools read this file (Anthropic, Vercel, Stripe, Cloudflare and Hugging Face publish one, and AI coding assistants use it to find docs), so adding one is a cheap courtesy to them. Google has said in writing that Google Search and its AI features do not use llms.txt, so do not expect a search-ranking or AI Overviews benefit from it. paste into Claude, Cursor, or ChatGPT | |
| Medium | Security headers are missing or weakBothSecurity The server response is missing browser-hardening headers that protect visitors and are a standard security and agency checklist item. Missing or weak here: Content-Security-Policy (the main defense against injected and cross-site scripts); clickjacking protection (X-Frame-Options or a CSP frame-ancestors rule); a stronger HSTS policy (max-age at least 180 days plus includeSubDomains); X-Content-Type-Options: nosniff (stops MIME-type sniffing attacks). These are set at the server, CDN, or host level (most platforms expose them in settings or a config file) and do not change how the site looks or performs. paste into Claude, Cursor, or ChatGPT |
Accessibility findings are automated checks against Web Content Accessibility Guidelines (WCAG) 2.1 and 2.2. They flag potential barriers and legal risk, not a certification or a determination of compliance with the ADA, Section 508, or EN 301 549. Automated testing catches only a subset of issues; a full conformance review needs manual and assistive-technology testing by a qualified reviewer.
"How we know": unlabeled = a deterministic fact, identical on every load (e.g. element sizes). Most findings are this kind, so we only mark the exceptions: median across loads = a noisy lab metric, reported as a median. real-user field data = Google CrUX, actual Chrome visitors.
Structural and AI-search checks crawl up to 8 pages from your sitemap (a sample, not your full site). "Broken" means a link returned 404, 410, or 5xx, or did not respond; access-controlled pages (401, 403) are not counted.
02Performance: phone, desktop, and real visitors
| Metric | Mobile | Desktop | Read |
|---|---|---|---|
| TTFB (lab median) | 110 ms | 80 ms | Lab |
| FCP (lab median) | 230 ms | 184 ms | Lab |
| LCP (lab median) | 230 ms | 184 ms | Good |
| Page weight (median) | 0.1 MB | 0.1 MB | OK |
Google Lighthouse (lab): Performance 84 mobile / 98 desktop, SEO 100, Accessibility 100, Best Practices 100.
Lab numbers are from a headless mobile browser on an unthrottled connection: treat them as a floor, not a typical experience. Add a Google API key to light up real-user field data (CrUX) and Lighthouse scores.
03Tiny buttons are hard to tap on mobile
16 of 25 tappable items on this page come in below the platform minimums for reliable tapping on a phone (Apple recommends 44pt, Android 48dp; WCAG 2.5.8 sets 24px as the hard accessibility floor). The same ones came up small on every one of the 4 test loads, so this is the page itself, not a fluke.
The buttons measuring below the minimum on this scan:
- a 308x20 "hello@contractive.app"
- a 38x18 "About"
- a 57x18 "Products"
- a 51x18 "Contact"
- a 91x18 "Local Analysis"
- a 78x18 "Risk Review"
- a 113x18 "Clause Extraction"
- a 99x18 "Private Legal AI"
The fix is CSS-only on most sites: add padding around the icon (don't just change the icon size) so the actual tap area is at least 44×44 pixels. No redesign, no new assets.
04Technical SEO & structured data
| Check | Result |
|---|---|
| Title | Contractive | Local AI Contract Analysis Software (49 chars) |
| Meta description | 125 chars |
| H1 | 1 on page |
| Canonical | Present |
| Structured data (JSON-LD) | Organization, WebSite, SoftwareApplication, BreadcrumbList, FAQPage |
| Open Graph | Title + image |
05The fix checklist
Everything to fix, priority first, each tagged with the screen it affects and a rough effort. Work top to bottom.
- Tiny buttons are hard to tap on mobileMobileCSS only
- A floating widget is covering a clickable elementBothCSS only
- Images have no width or height set (layout is stable for now)BothCSS only
- Unused JavaScript is being downloadedBothVaries
- 1 form field has no labelBothVaries
- Add quotations so AI engines cite this pageBothVaries
- No advertising pixel detectedBothDev afternoon
- Broken internal linksBothVaries
- 1 input uses the wrong type attributeBothVaries
- No llms.txt fileBothVaries
- Security headers are missing or weakBothVaries
Effort is a rough read from the outside: "CSS only" means no new assets or backend work, "1 line" means a single tag, "Dev afternoon" means a developer needs to touch tracking or scripts.
06What this report cannot tell you
Everything above is from the outside, looking at the page on a simulated phone and desktop. The questions that actually decide revenue need real visitors. Install the Harvv pixel (one script tag, 16 KB, zero personal data, no engineering project) and within about 72 hours you'd know which buttons real customers tapped and missed, how often Google Analytics is missing visits, and exactly where mobile shoppers stalled and left. This report shows you where to look. The pixel shows you how often it happens, and to whom.
Drop the Harvv pixel on contractive.app and we turn this one-off scan into ongoing measured behavior: which taps miss, where sessions stall, and the real drop rates. Free to start, no card needed.
Add the pixel free07How we did this, and what it can't prove
- 4 mobile + 2 desktop loads of one URL from headless Chrome (iPhone viewport at 390px, desktop at 1366px), August 30, 2026. Enough loads to separate real defects from random noise, not a full-site crawl.
- Lab numbers, not real-user numbers (no field data was available for this run). Real devices on real networks run slower.
- Friction is inferred, not counted. We can prove a button is small. We can't, from the outside, count how often it causes a missed tap. That requires the pixel on a live page.
About Harvv, the source of this teardown
Harvv is a behavioral UX analytics platform (harvv.com). A lightweight JavaScript pixel captures how real visitors behave on a site (dead clicks, rage clicks, scroll depth, Core Web Vitals, JavaScript errors, and 50+ other signals) and the engine turns them into prioritized, plain-English findings. This teardown is the outside-in version of that: the same detectors run against a public page, with no pixel installed.
How to read it. Every finding here is a reproducible, automated measurement, not an opinion: element sizes, contrast ratios, load metrics, and structured-data checks that anyone can re-run against the same URL. The method is stated in full above. Automated testing catches a subset of issues, so this is a starting point, not a certification.
Full disclosure. Harvv makes the pixel that would measure the friction these findings imply, so we have a commercial interest. That is exactly why the findings are kept to things a reader can verify independently, and why nothing here is inflated: an unreproducible claim would undermine the tool it is meant to demonstrate.
Prepared by Harvv (harvv.com), a behavioral UX analytics platform. Last updated August 30, 2026.