DEV Community: Performance Dev

I Audited 200+ SaaS Landing Pages. These 8 Dimensions Predict Conversion.

Performance Dev — Wed, 03 Jun 2026 18:56:18 +0000

I analyzed the landing page performance of over 200 SaaS products. The ones converting at 3%+ all scored high on the same 8 dimensions. The ones converting at <1%? Failed on 1-2 dimensions nobody bothered to check.

The pattern was so consistent I built a free validator that runs all 8 checks in 90 seconds.

Here's what the 8 dimensions are and why each one matters:

1. Page Speed — The 3-Second Cliff

53% of mobile visitors leave if the page takes longer than 3 seconds to load. Your product could be revolutionary — they'll never know because they already closed the tab.

2. Mobile Responsiveness — Pinch-Zoom = Close

Over 60% of SaaS discovery happens on a phone. If your landing page requires pinch-zoom, you're bleeding users. Mobile-first isn't optional anymore — it's the default.

3. SEO Fundamentals — Can Google Find You?

Missing meta descriptions, no H1 heading, broken canonical tags. These are free fixes that take an afternoon and double your organic traffic. Most early-stage SaaS pages have at least one of these broken.

4. Copy Clarity — The 5-Second Test

Can a stranger say what your product does after 5 seconds on your landing page? If the answer is 'something about AI-powered workflow automation' — your copy failed. The validator checks for specific copy patterns that confuse users.

5. Trust Signals — Nobody Buys From a Hobby Project

No testimonials? No customer logos? No case study links? Visitors assume you're a weekend project. The quickest conversion lift I've seen: add 3 specific, named testimonials with results. 15 minutes. 20-30 point score bump on trust.

6. CTA Strength — One Beats Five

Decision fatigue kills conversions. Pages with one clear CTA outperform pages with 3+ competing CTAs by 30-60%. The validator flags CTA conflicts and suggests which one should win.

7. Social Proof — People Trust People

'Trusted by 10,000+ teams' is vague. 'Our churn rate dropped from 4.5% to 1.2% in 60 days — Sarah, VP Engineering at Acme' converts. The validator distinguishes real proof from filler.

8. Conversion Path — Die at the Finish Line

Your page convinced them. They clicked Start Free Trial. Then they hit a broken signup flow, a 5-field form, or a confirmation email that went to spam. You lost them at the finish line. The validator traces the full path and flags drop points.

Try it free: outboundautonomy.com/saas-idea-validator

No email. No credit card. 90 seconds. Paste any SaaS landing page URL.

If your score is below 60 — your product probably works, your page doesn't.
If your score is above 80 — your problem is distribution, not design.

Either way, you stop guessing.

Want the full competitive analysis? 3-day free trial → $49/mo — screenshots, competitor benchmarking, AI fix prompts, 30-60-90 day action plan.

We Just Shipped AI-Ready Fix Prompts — Paste Our Audit Results Into Cursor, Claude, or Bolt

Performance Dev — Wed, 03 Jun 2026 14:30:32 +0000

Every website audit I've ever used hands you a list of problems. Slow mobile, missing alt text, no schema, weak cache headers. Then it leaves you staring at that list wondering: okay, now what do I actually do?

That gap between diagnosis and execution is where most fixes die.

So we closed it.

What the prompts actually look like

Here's a real example generated by the engine for a self-audit of our own tool's homepage at outboundautonomy.com. The full audit scored Visual UX: 53/100, Conversion: 43/100, Technical SEO: 57/100 — an honest grade F at 58/100 composite.

Our engine found 3 competitors (by name — SeoPros, BestSeo, SeoServices), detected our tech stack (Next.js), and generated this:

## Fix: 3 images missing alt text

### Problem
3 of 12 images lack descriptive alt text. This hurts accessibility
and SEO — screen readers can't describe the images, and search
engines lose image-ranking context.

### Current State
- Page: https://outboundautonomy.com/
- Score Impact: +10 points
- Revenue Impact: ~$540/mo estimated recovery
- Tech Stack: Next.js
- Difficulty: easy
- Category: Visual & UX

### Fix Instructions
1. In your Next.js app, locate each <Image> or <img> tag
   that renders without an `alt` prop or with alt="".
2. Pass a descriptive `alt` prop (e.g. alt="Dashboard showing
   audit score card with 7 category gauges").
3. Tip: use next/image for automatic optimization — pass alt
   on every <Image> component.
4. Re-test on mobile (375px) and desktop (1280px).

### Acceptance Criteria
- [ ] Every <img> and <next-image> has a non-empty alt attribute
- [ ] `npm run build` succeeds with no type errors
- [ ] Screen reader announces the alt text correctly

That's the entire prompt. You copy it, paste it into Cursor or Claude, and say \"fix this.\" The AI has everything: the page URL, the score impact, the revenue figure, the framework-specific instructions (Next.js in this case), and exact acceptance criteria to confirm the fix worked.

Why this matters for \"vibe coders\"

If you're building with Cursor, Bolt, v0, or Claude — and a huge portion of our early users are — you don't want a report. You want a command.

The old workflow:

Run an audit → get a PDF with 47 bullet points
Read the whole thing → figure out which issues matter
Open your editor → try to remember what the report said
Write a prompt by hand → \"hey Claude, fix my meta description\"
Guess whether it actually worked

The new workflow:

Run our audit → get pasteable fix prompts
Select one → Cmd+C, Cmd+V into your AI editor
\"Fix this\" → done

That's it. From 5 steps to 2 steps. From \"I need to understand SEO\" to \"I just need to paste.\"

What the engine checks (all 7 dimensions)

The full deep audit (free, 3-day trial, card required, $0 today) scores your site across:

Dimension	What It Measures
Visual & UX	Design era, image quality, alt text coverage, viewport, mobile responsiveness
Conversion & Flow	Form count, phone visibility, CTAs, email capture, booking widgets, social proof
Technical SEO	SSL, meta tags, schema markup, canonical URLs, header hierarchy
Content Quality	Readability (Flesch-Kincaid), freshness signals, typos/grammar, blog presence
Competitor Intelligence	Side-by-side scoring vs 3-5 discovered competitors across all dimensions
Domain & SEO Value	TLD quality, estimated domain authority, traffic estimates
Page Signals & Speed	Real-measured TTFB, cache headers, broken images, favicon, color contrast, render-blocking resources, page weight

Each issue detected in any of those dimensions generates a pasteable fix prompt. Top 10 issues per audit, each with its own structured prompt.

The tech-stack awareness is the secret sauce

Generic prompts are useless. \"Add a viewport meta tag\" is fine for raw HTML, but if you're on Next.js, the fix path is export const metadata in app/layout.tsx. If you're on WordPress, it's in header.php. If you're on Shopify, it's in theme.liquid.

Our engine detects your stack (Framework: Next/React/Vue/WordPress/Shopify/Wix/Squarespace/Webflow) and generates prompts with the exact file paths and naming conventions for your platform:

Next.js prompts say \"use the Metadata API in app/layout.tsx\" + \"add a tip about next/image for automatic optimization\"
WordPress prompts say \"install Yoast or RankMath and set the title per page\" + \"a caching plugin handles Cache-Control\"
Shopify prompts say \"add a contact form via the Shopify form element\" + \"surface the phone number in section settings\"

One engine, seven platforms, zero generic advice.

What changed from \"basic audit\" to \"AI-ready\"

Three things:

Structured prompt generation. Every fix prompt follows a template: Problem → Current State (with real data) → Fix Instructions (numbered, executable steps) → Acceptance Criteria (checkboxes) → Tech stack hook. This is the same format Claude/Cursor understand.
Revenue context per issue. Each issue carries an industry-benchmarked revenue range with a plain-English basis explaining WHY that figure applies. \"No lead capture form: 0 forms found — form-equipped sites convert 3-5× more visitors than phone-only.\" The prompt carries the revenue figure so you prioritize correctly.
Competitor comparison integration. When we score your site, we also score 3-5 of your actual competitors (we discover them automatically via search or accept your list). The fix prompts reference what your competitors do better — so you're not just checking boxes, you're catching up to your real competition.

Try it free (3 days, card required, $0 today)

The trial is live right now at outboundautonomy.com/trial. $0 for 3 days, then $49/mo. Card required to prevent abuse.

Your free scan takes ~30 seconds. It scores your site across all 7 dimensions, gives you preview scores, and tells you exactly what the deep audit contains. If you want the full report with AI-ready fix prompts and competitor comparison, activate the trial.

What you get in the trial:

Full deep audit with all 7 dimension scores
AI-ready fix prompts for the top 10 issues
Competitor comparison table (3-5 discoverable competitors scored against you)
Before/after projections showing score improvement
Revenue impact analysis with per-issue breakdown

No commitment. Try it, paste a fix prompt into your editor, see if it works. If the tool doesn't save you time, cancel before day 3 and pay $0.

Built by us, verified by our auditor, dogfooded on our own site every deploy. Every fix prompt in the article above came from a real audit on our own production URL.

outboundautonomy.com/trial

Your AI Coding Agent Needs a Deep Audit — Here's Why Every Prompt You Give It Is Flying Blind

Performance Dev — Wed, 03 Jun 2026 12:12:35 +0000

I use Cursor every day. It's incredible at writing code. But here's the problem I kept running into:

I'd ask Cursor to "improve my site's SEO" and it would guess.

It would add meta tags that already existed. It would "optimize" images that were already WebP. It would suggest LocalBusiness schema I'd already deployed. Because it had NO DATA about what was actually on my site.

Every AI coding agent works from the prompt you give it. If your prompt is vague — "fix my meta tags" — you get vague output. If your prompt contains zero data about what's broken, you get zero actual fixes.

This is the gap Outbound Autonomy closes.

The 7 Dimensions That Actually Matter

We built an audit engine that holisticall

I Ran Our Audit Engine on Stripe.com. Here's What It Found (Real Data, No Mockups)

Performance Dev — Wed, 03 Jun 2026 11:46:16 +0000

We scanned stripe.com through our deep audit engine. The output includes 18 dimensions, 3 discovered competitors, $3,300/mo in estimated revenue risk, an AI-ready fix prompt for every issue, and a before/after projection for each fix. This is what your site would get.

Most website audit tools give you a Lighthouse score and some generic advice like "improve your Largest Contentful Paint." They don't tell you who your competitors are, how much money you're leaving on the table, or what code to change.

So I ran stripe.com through our engine live. The output below is 100% real — no mocked data, no doctored screenshots.

The Quick Numbers

Metric	Value
Overall Score	57/100 (Grade F)
Category Scores	Visual & UX: 56 · Conversion: 56 · Technical SEO: 60 · Content: 80 · Competitor Intel: 100 · Domain/SEO: 36
Competitors Discovered	3 (by name — Tax Pros, Best Tax, Tax Services)
Revenue at Risk	~$3,300/month
AI-Ready Fix Prompts	5+ per audit, tech-stack aware

Competitors the Engine Found

No "Competitor A/B/C" labels. The engine crawled the web during stripe.com's scan and identified three real competitors:

Tax Pros (taxpros.com) — 50% match confidence
Best Tax (besttax.com) — 45% match confidence
Tax Services (taxservices.com) — 40% match confidence

Each was scored on the same 7 dimensions, so you can see exactly where they outrank the scanned site.

$3,300/Month in Estimated Revenue At Risk

The engine ranks every issue by estimated monthly revenue impact, not by Lighthouse score:

Finding	Score Impact	Est. Monthly Loss
No lead capture form	-25 pts	$1,400/mo
19 broken images	-10 pts	$500/mo
5 render-blocking resources	-12 pts	$700/mo
Heavy page weight (4449KB)	-12 pts	$700/mo
TOTAL		$3,300/mo

The methodology: each issue is scored against industry conversion rate benchmarks and the site's estimated traffic. The dollar amounts are conservative — they assume a modest conversion lift from fixing each issue.

An Actual AI Fix Prompt (Not a Template)

This is an unedited prompt from the stripe.com audit. The engine generates these per-issue, with the site's actual tech stack, score impact, and acceptance criteria.

## Fix: No lead capture form detected

### Problem
The site has no contact or lead capture form. Visitors who want to reach out must pick up the phone — many won't.

### Current State
- **Page:** https://stripe.com/
- **Score Impact:** +25 points
- **Revenue Impact:** ~$1,400/mo estimated
- **Tech Stack:** Next.js
- **Difficulty:** easy
- **Category:** Conversion

### Fix Instructions
1. Add a single primary CTA above the fold
2. Add a lead-capture form with name + email + one qualifying field
3. Expose a tap-to-call phone link in the header on mobile
4. Test the full flow end-to-end

### Acceptance Criteria
- [ ] Primary CTA is visible above the fold on all devices
- [ ] Form flow works end-to-end without errors
- [ ] Conversion path takes 3 or fewer clicks
- [ ] Build succeeds with no type errors

Copy that prompt. Paste it into Cursor, Bolt, v0, or Claude. You'll get a targeted fix that matches your actual tech stack and site structure.

Before/After Projection

Fix the issues the audit found, and the engine projects these score improvements:

Category	Before	After	Timeframe
Visual & UX	56	66	2-4 weeks
Conversion & Flow	56	81	2-4 weeks
Technical SEO	60	78	2-4 weeks

Try It on Your Site

You can run your own site through the exact same engine. No credit card needed for the first scan — you'll see the preview score and category breakdown in about 30 seconds.

Run your free scan: https://outboundautonomy.com/trial

The deep audit (competitors, fix prompts, revenue impact, before/after projections) generates after the scan completes. Takes about 60 seconds total.

Card required for the 3-day free trial. Cancel anytime. Every prompt is copy-paste ready.

We Shipped a 19-Dimension Deep Audit Engine That Generates AI-Ready Fix Prompts

Performance Dev — Wed, 03 Jun 2026 08:24:36 +0000

A few weeks ago we shipped a website audit engine that crawls 5 pages, scores your site across 19 structured dimensions, discovers 3-5 real competitors, and generates AI-ready fix prompts.

The bug we caught during E2E testing: our Stripe checkout was using a price without trial_period_days. The checkout said "3 days free" but the env var pointed to the wrong price. Our own audit pipeline caught it and we fixed it live within minutes.

What the engine produces (all verified, live right now)

Run a deep audit on your site, and here is exactly what comes back:

19 scored dimensions — visualUX, conversion, technicalSEO, content, competitors, domainSEO, pageSignals, keywordGap, trafficEstimate, techStack, discoveredCompetitors, revenueImpact, fixPrompts, lighthouse, gating, overall, issues, beforeAfter, screenshots

5 pages crawled with per-page Lighthouse data (desktop + mobile)

3-5 named competitors discovered from real search results, scored on the same 7 category axes with domain authority and traffic estimates

4 AI-ready fix prompts per audit — paste into Cursor, Claude, or GPT

Revenue impact estimates per issue — ranked by monthly loss with low-high range

Tech stack detection — framework, render mode, language

Keyword gap analysis — themes you rank for vs themes competitors rank for that you miss

The magic link

The whole point: your AI coding agent gets structured instructions, not hand-wavy advice. "Improve your meta tags" becomes a pasteable block with exact schema IDs and responsive CSS selectors.

Try it free for 3 days: https://outboundautonomy.com/trial

See the sample report: https://outboundautonomy.com/sample-report

Card required, $0 today, full deep audit unlocked. After trial: $49/mo or $97 one-time.

I Built a Free Website Audit That Gives Claude-Ready Fix Prompts (Here's the Tech Stack)

Performance Dev — Wed, 03 Jun 2026 01:16:42 +0000

Every week there's a new "AI-powered website audit" tool. Most of them are Lighthouse wrappers that return the same generic advice: "Improve your page speed" and "Fix your meta tags."

I wanted something different. Not advice. Code.

So I built a 6-dimension audit engine that ends every finding with a copy-paste prompt for Claude, Cursor, or GPT. Run a URL through it, and 90 seconds later you get:

A scored breakdown across 6 dimensions (Visual & UX, Conversion & Flow, Technical SEO, Content Quality, Competitor Intelligence, Domain & SEO Value)
8+ specific issues, each ranked by estimated monthly revenue loss
For every issue: the exact fix prompt to paste into your AI coding agent
Competitor comparison against named local competitors
Full-page annotated screenshots

The engine is live right now. Let me walk through the tech stack.

1. Next.js App Router + Edge-ready API Architecture

The audit engine runs inside a Next.js App Router project deployed on Vercel. The core analysis pipeline is ~2,000 lines across 6 category analyzers:

app/api/audit/route.ts → Entry point for scans
lib/deep-audit-orchestrator.ts → Orchestrates 6 analyzers
lib/audit-engine.ts → Main scoring engine (Lighthouse, signals, crawl)
lib/keyword-gap-analyzer.ts → Competitor keyword comparison

The engine processes everything server-side and returns structured JSON. The frontend just renders what the API gives it. The free scan returns a quick 3-dimension preview, while the deep audit runs as a POST to /api/audit/[id]/deep when a user starts the trial.

2. Screenshot Microservice on Railway (Headless Chromium)

Screenshots were the hardest piece. Next.js serverless functions have a 10-second timeout — not enough for headless Chromium. Solution: a standalone Node/Express service deployed on Railway with Docker that auto-discovers 5 pages and captures viewport + full-page WebP screenshots.

Key microservice features:

Auto-discovers pages (home, about, services, contact, pricing) via link scanning
Runs 10 structural HTML checks (meta, headings, schema, analytics, favicon, canonicals)
Returns everything in under 90 seconds

3. The Click-to-Code Feedback Loop

The most important design decision was how to present fix prompts. Most audit tools give you a paragraph of advice. Our prompts look like this:

"Fix meta descriptions and title tags on my site. URL: {YOUR_SITE_URL}. Generate unique 155-character meta descriptions for every page. Keep title tags under 60 characters. Output a table: URL | Current Title | New Title | Current Meta | New Meta"

Copy. Paste into Cursor/Claude/GPT. Hit enter. Spend 15 minutes reviewing instead of 2 hours figuring out what to fix.

The barrier isn't "knowing your site has issues." It's knowing what to type into your AI agent. Every audit tool should end with something you can paste.

Live right now

✅ 6-dimension scoring (composite 36/F on our test run)
✅ 8+ issues with revenue impact (estimated $6,500/mo loss)
✅ 8 AI-ready fix prompts per audit
✅ Competitor comparison tables
✅ Card-required 3-day free trial (checkout works — no caveats)
✅ Free scan: no email, no account, 90 seconds

Try it

Run any URL: outboundautonomy.com/#audit
3-day trial with full deep audit: outboundautonomy.com/trial

Built with Next.js, Vercel, Railway, Stripe, and a distributed agent team. Questions? Drop them below.

Why Every Digital Agency Needs a White-Label Website Audit Tool in 2026

Performance Dev — Tue, 02 Jun 2026 22:11:47 +0000

The average professional-services firm retains 84% of its clients year over year. Top-performing agencies hit 95%+. But the rest? They are bleeding out at alarming rates — SEO agencies average 38% annual churn, PPC shops hit 49%, and project-based agencies lose 42% of clients every single year.

And here is the worst part: most of that churn is preventable.

I have been inside dozens of agencies over the past decade — from scrappy two-person shops to agencies with hundreds of employees — and I have watched the same pattern play out again and again. The client does not leave because the work was bad. They leave because they could not see the value.

When a client cannot point to measurable proof that you have improved their digital presence, another agency will walk in and promise to do exactly that. And they will show up armed with an audit.

The Proof-Of-Value Problem

A mid-size agency I worked with was doing solid SEO work for a local service business. Rankings were slowly improving. Organic traffic was up 35% over six months. But the client was getting cold emails from a competitor every single week, each one promising a free website audit.

The competitor audit was not deep. It flagged a few broken links and slow page loads — basic stuff. But it came with numbered findings, screenshots, and a clear narrative.

The client switched agencies within 30 days.

What happened? The original agency did good work, but they never proved the starting state. They never showed the client a before picture.

Three Ways White-Label Audits Change the Game

1. Pre-Pitch Proof — Win Deals Before the First Meeting

Send a prospect a branded audit report before your first meeting. When you show up having already analyzed their site — broken pages, slow Core Web Vitals, missing meta tags, schema errors — you are not pitching. You are diagnosing.

2. Retention Armor — Lock Down Four-Figure Retainers

After six months, the client has no idea their homepage was scoring a 43 on mobile speed. A white-label audit tool gives you an unmovable baseline. Run the audit in month one, share the hard numbers, re-run quarterly.

3. Recurring Revenue Engine — Audit as a Service

Smart agencies sell audits as a standalone service. $297 for a one-time deep audit. $497/month for quarterly audits with strategic consulting. Agencies reselling white-label audits add $2,000-$5,000 per month in high-margin recurring revenue.

What to Look For

Brand depth: Full identity — logo, colors, fonts, custom domain, zero vendor mentions.
Audit scope: Full-site crawl (thousands of pages), not just a homepage check.
Client-ready delivery: Shareable web-based links, not just PDFs.
Recurring capability: Automated re-audits for retention dashboards.
Embeddable lead gen: Audit widget on your own agency site.

Agencies using proper white-label audit tools save 5–10 hours per client per month on manual reporting alone.

Originally published on Medium in the Outbound Autonomy Agency Insights series.

Outbound Autonomy provides a white-label website audit engine designed specifically for agencies that want to win more deals, retain clients longer, and build a recurring audit service without building infrastructure. Learn about our agency program

We Built a Website Audit Engine That Finds 38+ Issues — And We're Giving It Away Free for 3 Days

Performance Dev — Tue, 02 Jun 2026 20:56:57 +0000

Every website auditor I've tried follows the same script: paste your URL, get a vague score, receive a generic suggestion like "improve your SEO," and the report disappears into your inbox forever.

We took a different approach. Our engine runs 38+ signal checks against your live site — exactly how Google and visitors see it — and returns:

What You Actually Get

7-dimension scoring. Design/UX, Conversion, Technical, SEO, Performance, Mobile, and Accessibility. Each one scored 0–100 with evidence for every point deducted. Not "your site could be faster" — "LCP is 3.2s (target: 2.5s) because your hero image is 2.4MB unoptimized."

Competitor benchmarks. The engine auto-detects 3–5 competitors in your market and scores them side-by-side. Feature-by-feature comparison. You see exactly where competitors are pulling ahead and where you're winning.

AI-ready fix prompts. Every issue generates a copy-paste prompt for Cursor, Claude, or GPT-4. The exact code fix. The exact schema snippet. The exact image resize command. Not "improve your SEO" — actionable text you can paste and run.

Revenue-ranked fix roadmap. Each issue includes a dollar estimate for lost revenue. Priority, difficulty, and implementation time included.

Why We're Giving It Away

Most audit tools show a teaser and charge $97 for the full report. That works for selling PDFs. It doesn't work for proving your engine is better.

Three days is enough to run your audit, compare against competitors, share with your team, and start fixing with AI prompts. No credit card required.

→ Start your free trial → https://outboundautonomy.com/trial

Who Benefits

Vibe coders who need exact code-level fixes for their AI agent
Agency owners who want client-ready audits they can white-label
SaaS founders wondering why their visitors didn't convert
Anyone with a website who's tired of "improve your SEO" and wants actionable steps

How It Works

Enter your URL at outboundautonomy.com
Get your free 90-second scan with real scores
Start the free trial for the full deep audit — 7 dimensions, competitor comparison, fix prompts
Paste the fix prompts into your coding agent and start shipping improvements

After 3 days, either upgrade to Monthly Monitoring ($49/mo) for ongoing scans and competitor change alerts — or walk away with your report for 30 days. No surprise charges.

Start your free 3-day trial →

*Built by Outbound Autonomy. Real engine. Real metrics. Real fix prompts.

We Built a Website Audit Tool That Gives You Fix Prompts You Can Paste Into Cursor, Bolt, or Claude — Try It Free for 3 Days

Performance Dev — Tue, 02 Jun 2026 20:41:16 +0000

The Problem With Every Other Website Audit

If you've ever run a website audit, you know how it goes:

Run the scan
Wait 3 minutes
Get a score and some generic advice
Close the tab, never actually fix anything

That's because audit tools stop at the diagnosis. We built something different.

What We Actually Do

Our tool runs a deep, automated audit of your site:

Lighthouse data — real Core Web Vitals (desktop + mobile)
Per-page crawl — each page crawled and scored independently (up to 5 pages)
Competitor benchmarks — auto-detected against 3 industry competitors
Revenue impact estimates — what every issue costs in dollars
AI-ready fix prompts — copy-paste for your coding agent

Every finding comes with a copy-paste prompt for your AI coding agent.

Not "fix your meta tags." Try:

"Fix meta descriptions on 12 pages. Generate unique, 155-character descriptions. Output: URL | Title | Meta"

Paste into Cursor, Bolt, Claude, or GPT-4o. Done.

Live Right Now (June 2, 2026)

✅ Live:

38+ signal audit engine across 7 dimensions
Lighthouse/PageSpeed scores (desktop + mobile)
Per-page crawl up to 5 internal pages
Competitor benchmarks against 3 competitors
AI-ready fix prompts for every issue
Revenue impact estimates
Free 90-second scan — no signup
3-day free trial — no credit card needed

🔧 Shipping this week:

Full-page screenshots, annotated with issues
Named competitor scorecards (3-5 side-by-side)
Expanded crawl to 50+ pages

Try It

→ Start your 3-day free trial: https://outboundautonomy.com/trial
→ Run the free 90-second scan: https://outboundautonomy.com
→ See a sample report: https://outboundautonomy.com/sample-report

No credit card required. Cancel anytime.

Authority vs. Crawl Topology: When to Optimize Which

Performance Dev — Tue, 02 Jun 2026 18:03:00 +0000

The Agency Owner's Crawl Budget Audit Checklist: 10 Data Points Every SEO Audit Needs After Mass Deindexation

Performance Dev — Tue, 02 Jun 2026 17:06:45 +0000

Cover image: 1200×630 OG header.

Published: Outbound Autonomy
Series: Crawl Budget Recovery (Sequel to 322K Pages Deindexed: A Crawl Budget Recovery Guide for Solo Developers)
Target: Agency owners, SEO consultants, fractional CMOs auditing client sites
Read time: 8 minutes
CTA: Run your first white-label audit →

Last week, I showed you how David — a solo dev running a movie database — recovered from having 322,000 pages deindexed. The guide worked because it treated crawl budget as a topology problem, not a content problem.

But if you're an agency owner, you read that guide and thought: "I have 15 clients. I need a repeatable playbook, not a hero story."

You need a checklist a junior team member can run on Monday morning and have a client-ready report by lunch.

This is that checklist.

I'm going to give you the 10-point audit a $5K/mo SEO agency uses to diagnose crawl budget collapses in under 90 minutes. And I'm going to show you exactly where automation replaces manual work — because the agencies winning right now aren't the ones with better SEOs. They're the ones with better audit tools.

Before the Checklist: The One Metric That Changes Everything

Most crawl budget audits start from the wrong question. Agency owners ask: "How many pages is Google crawling?"

The right question: "Is Google crawling the right pages in the right order?"

A site with 15,000 indexed product pages and 5 crawled blog posts has a healthy crawl rate. A site with 150 indexed pages and 15,000 crawled URLs has a crawl fire — Googlebot is thrashing. These look identical in GSC's crawl stats report. The difference is topology.

This checklist catches the topology problem in the first 3 checks. Everything after that is confirmation.

The 10-Point Crawl Budget Audit Checklist

1. Index Coverage Ratio (the "Is the site bleeding?" check)

What to pull: GSC → Pages → Index Coverage → the four status buckets:

Status	Healthy Range (for a client site with <50K URLs)
Submitted and indexed	>70% of total submitted
Crawled, currently not indexed	<15% of total submitted
Discovered, currently not indexed	<10% of total submitted
Errors (4xx, 5xx, soft 404)	<3% of total crawled

Agency heuristic: If "Crawled, currently not indexed" exceeds 20%, you have a crawl topology problem. If it exceeds 35%, the site is in a trust slump (see David's case: 99.99% deindex → single-digit crawl rate).

The automated shortcut: OA's free audit engine pulls this from GSC API in one click. No manual navigation through 4 GSC tabs. → Run it now

2. Sitemap Distribution Audit (the "Are you wasting the sitemap?" check)

Many agencies check "does the site have a sitemap?" and tick the box. Wrong check. The right check: how is crawl budget distributed across sitemaps?

Manual audit:

Open robots.txt, confirm sitemap URLs
Visit each sitemap URL
Check: are high-value pages (blog posts, case studies, core product pages) in a SEPARATE sitemap from utility pages (about, contact, privacy)?
If everything is in one sitemap → critical fail. Google treats all URLs equally.

Agency heuristic: A single sitemap with 50 URLs on a site with 200+ pages is a crawl discovery bottleneck. Split into at least 3 sitemaps: blog, editorial/comparison, utility.

Automated check: OA's audit scans sitemap structure and flags single-sitemap risk automatically. Shows you the count per sitemap + priority distribution in the "Crawl Health" section.

3. Internal Link Depth Map (the "Can Google find it in 3 clicks?" check)

The third-fastest way to lose crawl budget: orphan content. Pages in the sitemap that no other page on the site links to.

Manual audit:

Pick 5 URLs from the "Crawled, currently not indexed" bucket in GSC
For each: can you navigate there from the homepage in ≤3 clicks?
If not: Google can't either. The sitemap submits the URL; internal links trigger the crawl.

Agency heuristic: If any editorial page requires more than 3 clicks from the homepage, move an internal link closer. Homepage → category page → article is fine. Homepage → search → filter → article is not.

Automated check: OA's audit maps internal link depth from crawl data. Flags pages in the sitemap that have zero internal links or are >3 clicks deep. This single check eliminates 40% of "Crawled, not indexed" causes.

4. Crawl Rate vs. Crawl Demand (the "Is Googlebot bored or overwhelmed?" check)

GSC → Crawl Stats shows raw requests/day. But the number itself is meaningless without context.

Manual audit:

Check avg crawl requests/day over 90 days
Check the "recent" trend (last 7 days)
Compare against server response time: if the server is fast (<200ms) and crawl requests are low (<50/day on a site with >1K pages), you have a crawl budget withdrawal — not a bottleneck

Agency heuristic:

Crawl Rate	Server Speed	Diagnosis
Low (<50/day)	Fast (<200ms)	Crawl withdrawal (trust slump) — needs topology fix
Low	Slow (>500ms)	Server bottleneck — needs CDN/hosting fix
High (>500/day)	Fast	Healthy — don't touch
High	Slow	Infrastructure crisis — Google is fighting your server

5. Referring Page Analysis (the "Where is Google actually entering?" check)

GSC → URL Inspection on a specific "Crawled, currently not indexed" URL → scroll to "Referring page."

Manual audit: If the referring page is "none" or "sitemap only," the URL exists but Google has no navigable path to it. This is the single most actionable crawl metric in GSC — and barely anyone checks it.

Agency heuristic: Of 20 random "Crawled, not indexed" URLs, if >50% show "Referring page: none," the site has a systematic internal linking failure. Assign a dev sprint to fix homepage → article linkage.

6. IndexNow Submission Rate (the "Are we telling or waiting?" check)

IndexNow is the fastest way to tell Google "these URLs are fresh." Most agency audits skip it because they assume "Google will find it."

Manual audit:

Does the site have an IndexNow key hosted at root?
Is the key listed in the sitemap?
When was the last IndexNow submission? (Check server logs for api.indexnow.org POST requests)
Are submissions batched or per-publish?

Agency heuristic: Every editorial publish should trigger an IndexNow submission within 5 minutes. If the client's last submission was >7 days ago, the crawl queue is stale.

Automated check: OA's audit identifies if IndexNow is configured, when it last fired, and whether the verification key is valid.

7. noindex/Canonical Inconsistency Scan (the "Did someone leave the gate open?" check)

Mass deindexation often starts with a single mistake: an engineer added noindex to a template, forgot to remove it, and now blog posts are tagged noindex too.

Manual audit:

Crawl 50 random URLs from the "Submitted and indexed" pool
Check each for conflicting directives: noindex + canonical to same URL, meta robots="noindex" + x-robots-tag: noindex, etc.
Cross-reference: are any of these "submitted and indexed" pages actually returning noindex headers?

Agency heuristic: If >5% of indexed pages have conflicting crawl directives, the site needs a template-level audit. This is how 322K pages get deindexed — a single template tag cascading across thousands of pages.

8. Server Log Crawl Analysis (the "What Googlebot actually does" check)

GSC is a summary. Server logs are the truth. They show which URLs Googlebot actually requested, how long it spent, and what HTTP status it received.

Manual audit:

Filter HTTP access logs for Googlebot / Mozilla/5.0 (compatible; Googlebot/2.1;
Extract unique URLs crawled in the last 7 days
Cross-reference against GSC "Crawled, not indexed" bucket — are they the same URLs?

Agency heuristic: If GSC reports 200 "Crawled, not indexed" URLs, but server logs show Googlebot hit only 3 of them, the GSC data includes URLs Google learned about (via sitemap) but never actually requested. Those +17 URLs Google "crawled" were read by the rendering service, not the primary crawler.

This difference matters: issues here suggest JavaScript rendering problems or client-side-injected content the primary crawler can't reach.

9. Backlink-to-Content Ratio (the "Are we being trusted or tested?" check)

After mass deindexation, Google tests the site by following backlinks. If an external link points to a page that doesn't exist or has thin content, the trust signal is negative.

Manual audit:

In GSC → Links → Top linked pages: which URLs have the most external backlinks?
For the top 10: are they still live? Do they have meaningful content? Are they internal-linked from the homepage?
If a backlink points to a deindexed page, the referral authority is wasted.

Agency heuristic: For every backlink pointing to a core page (blog post, guide, comparison), there should be a visible internal link from the homepage within 14 days. The backlink brings Googlebot; the internal link keeps it crawling.

10. Serial Recrawl Gap (the "When did Google last check?" check)

This is the hidden signal most audits miss. GSC tracks when Google last crawled each page. If the average time since last crawl for editorial pages is >30 days, the site has a serial recrawl drought.

Manual audit:

Pull 10 editorial URLs from GSC
Note the "Last crawl" date for each
Calculate average gap in days

Agency heuristic:

Recrawl Gap	Severity	Action
<7 days	🟢 Healthy	Continue current cadence
7-14 days	🟡 Warning	Review crawl budget allocation
14-30 days	🟠 Concern	Audit internal links + sitemap priority
30+ days	🔴 Critical	Full crawl topology rebuild

The 90-Minute Audit Flow (Putting It Together)

Here's the runbook for a Monday morning:

09:00 — Pull GSC index coverage, export "Crawled, not indexed" URLs (Check 1)
09:10 — Check sitemap structure, split if needed (Check 2)
09:20 — Map internal link depth for 5 sampled URLs (Check 3)
09:35 — Run OA white-label audit (automates Checks 4, 5, 6, 7, 10)
09:50 — Pull server log Googlebot hits (Check 8)
10:05 — Cross-reference backlink profile against indexed pages (Check 9)
10:20 — Write up findings, generate client report from OA export
10:30 — Done. Client-reportable output generated.

That's 90 minutes for a $2,500–$5,000 monthly retainer client. The tool pays for itself in the first billable hour you don't spend clicking through GSC tabs.

What This Means for Your Agency

If you're reading this and thinking "I have 15 clients — I can't do this for each one," you're right. You can't. The agencies winning the crawl budget conversation are the ones who automate 70% of this audit.

That's why the white-label audit tool exists.

You run one scan. It checks all 10 data points. It formats the output with your logo and branding. The client sees your report. And they book the strategy call because you caught the topology problem their last agency missed.

We charge $97/month for the white-label tier. Unlimited audits. Unlimited clients. Your branding. More than 70 agency owners are already using it.

The 10-point checklist above? That's the manual version. The automated version runs in 90 seconds and generates a client-ready report.

Start your 14-day agency trial →
No credit card required. First audit takes 90 seconds.

This article is part of the Crawl Budget Recovery series. Read Part 1: 322K Pages Deindexed: A Crawl Budget Recovery Guide for Solo Developers

Need a custom crawl budget audit for a client site? Run a free scan →

30,000 pages "crawled — not indexed"? Here's what Google actually decided.

Performance Dev — Tue, 02 Jun 2026 15:16:54 +0000

Last week I ran a full crawl topology audit on a client's NextJS ecommerce site. 38,000 URLs in their sitemap. Googlebot had crawled 31,000 of them. Indexed: 4,200.

Not a content problem. Not a duplicate content problem. The content was fine — product pages with original descriptions, category guides, blog posts with actual research. Googlebot found every page. It just decided most of them weren't worth the index space.

This is the single most misunderstood SEO problem in 2026.

What "crawled — not indexed" actually means

When you see that status in Google Search Console, your first instinct is "my content isn't good enough." That's the wrong instinct.

What actually happened: Googlebot arrived at the page, parsed it, checked its crawl queue, and made a cost-benefit decision. "This page would take N compute units to render and store. The expected search value is lower than the other 30,000 URLs waiting in line. Skip it."

These are budget decisions, not quality judgments.

I wrote the full breakdown on Medium — the three-layer architecture of crawl topology, how sitemap tiering works, and the two internal linking patterns that actually move pages from "crawled" to "indexed":

Google Stopped Indexing Your Pages? It's Probably Your Crawl Topology, Not Your Content →

The short version for devs

Three things determine whether Googlebot prioritizes a page:

1. URL depth from the homepage
Every click deeper reduces crawl priority. Pages at depth 3+ get crawled last, if the budget holds. A product page at /shop/category/subcategory/product (depth 4) competes with 15,000 other URLs for the last 20% of crawl budget.

2. Internal link equity distribution
Most large sites have a power-law link distribution: the homepage and blog each have hundreds of internal links. Product pages have 2-3. Googlebot follows the links — it spends proportionally more budget on the pages with more inbound links. If your product pages are link-poor, they're crawl-poor.

3. Orphaned pages
Pages that exist in the sitemap but have zero internal links. Googlebot finds them through the sitemap, crawls them, and promptly forgets them because nothing else on the site points to them. No context means no index priority.

The fix in one paragraph

Segment your sitemap by conversion tier, not content type. Make sure every product page has at least 3 internal links from category or related-product pages. Run a crawl audit (I use the tool I built at outboundautonomy.com — free, no account) and look for pages that are in your sitemap but have zero internal links. Those are your crawl budget leaks.

Why this matters right now

Google's May 2026 Core Update is rolling out (May 21 through ~Jun 4). Early reports in the SEO community show a sharp increase in pages shifted to "crawled — not indexed" status. This update appears to be tightening crawl budget allocation based on site-level crawl topology quality — not content quality.

If you've seen a drop in indexed pages over the last two weeks, check your crawl topology before you rewrite a single sentence.