Planet Social Marketing Planet Social Marketing
Planet Social Marketing

Site Map Analysis: Field Guide

Your website is a planet. Pods are the probes that map it. This guide covers how to run them, how to read what they bring back, and where the honest limits are, on the web version and in the coming app.

What the tool does

Site Map Analysis builds a visual map of a website’s structure and grades its health. It surfaces three things most site owners never see:

  1. Orphan pages. Pages that exist but that no internal link reaches. Visitors cannot find them. Search engines struggle to. They are the quiet failure of most site migrations.
  2. Broken and redirected pages. Links pointing at pages that error or bounce elsewhere.
  3. Buried pages. Pages four or more clicks from home, technically reachable but practically invisible.

Everything runs in your browser. Nothing you upload leaves your device. There is no account, no database, and no server storing your data.

Three ways to feed the Pods

One: Send Pods to a URL (live crawl)

Enter a URL and press Deploy Pods. The Pods fetch the page, follow every internal link, and build the map as they travel. They crawl politely: they respect your robots.txt, pace their requests, and skip system pages (login screens, admin paths, feeds) and media attachment pages automatically, so the map shows real landing pages only.

Where this works: on the website the tool is hosted on. If you are using the tool at planetsocialmarketing.com, entering planetsocialmarketing.com works immediately.

Where it does not: other domains. Browsers block a web page from fetching content across domains for security reasons. This is a browser rule, not a tool choice. When it happens, the tool tells you, and points you to the app, which does not share this limit.

Two: Upload crawl exports (the richest data)

If you run Screaming Frog or SEMrush, their exports carry more detail than any live crawl:

If your Screaming Frog crawl was run with the GA4 integration connected, the tool picks up session counts, and you can size the map’s nodes by traffic instead of links. Big traffic on thin link structure is instantly visible.

Three: Paste link data (the fallback)

Under Advanced, a console snippet captures the links from any single page you are viewing. Paste the result and the tool maps it. This is a partial view by nature: it covers only the pages you captured, statuses are unknown, and orphan detection is incomplete. Use it when the other two paths are unavailable.

Reading the map

The default view is a traditional sitemap: home at the top, each level of click depth in its own row, connections drawn top down. Level markers (L0, L1, L2) run down the left side.

The disconnected band. Orphans sit in their own band at the bottom, under the line marked “Disconnected, no path from home.” They are literally outside the org chart, because that is where they are on your site.

Hierarchy first, detail on demand. The map draws parent to child connections only, because drawing every nav and footer link produces an unreadable tangle. Click any node and its real inbound and outbound links light up in mint. The full link graph is always one click away.

Controls. Drag to pan. Scroll or pinch to zoom. Double click or double tap to zoom toward a spot. Drag nodes to rearrange. Click a node for the inspector: full URL, status, links in and out, click depth, and traffic if available. The “layout: galaxy” toggle switches to an organic force view that draws every link.

The grade. A letter from A+ to F with a structure score out of one hundred. Orphans weigh heaviest, then broken pages, then redirects and buried pages. The scoring is scaled so one orphan on a small site reads as a warning, not a failing grade. The grade also rides along in the exported findings CSV.

Export findings. One button produces a CSV of every orphan, broken page, redirect chain, and buried page, with the grade stamped at the top. That file is a work order.

Pod tiers

A capped crawl grades what it saw. On larger sites, higher tiers are not just bigger numbers, they change how complete the diagnosis is. Higher tiers unlock with a client access code, and full Pods access is included in every client engagement.

Honest limitations

Web version:

  1. Cross-domain live crawling is blocked by the browser. The web tool live-crawls the site it is hosted on. Other domains require either the app or a fetch relay (a technical setup available to clients).
  2. JavaScript-rendered sites show less. The Pods read the HTML a server sends. If a site builds its navigation entirely in JavaScript after load (some single page applications), links created that way are invisible to the crawl. Screaming Frog with JavaScript rendering enabled, exported and uploaded, is the reliable path for those sites.
  3. Page caps by tier. Crawls stop at the tier limit and grade what they mapped.
  4. Long page names are shortened on the map. The full URL is always in the inspector and the export.
  5. The grade is a structural measure. It reads architecture, not content quality, keywords, or speed. An A+ structure with weak content is still weak, it is just findable.

The app (in development):

The Site Map Analysis app removes limitation one entirely. As a native application, its Pods fetch any site directly, no relay, no configuration. It also carries the tool offline. Limitations two through five apply to the app as well, they are properties of crawling itself, not of the browser.

The app will launch on Google Play following a closed testing period. Joining the waitlist makes you part of that early group.

Why this matters more in the age of AI answers

Search is changing shape. AI assistants now answer many of the questions your buyers used to type into Google, and they build those answers by crawling the web: following links, reading sitemaps, and assembling what they find into a knowledge graph. Getting cited in those answers has a name now, answer engine optimization, and it runs on the same plumbing this tool inspects.

Three things follow from that:

  1. Orphans are invisible to AI. A human might reach a disconnected page through an old bookmark or a shared link. An AI scraper never will. It follows paths, and a page with no path in does not exist to it. If your proprietary science lives on an orphan page, it will not be cited in the answer, and increasingly the answer is all anyone reads.
  2. Crawlers verify what AI generates. Generative tools are good at producing pages, plans, and site structures, and they produce them faster than anyone can check. But a generator works on what should exist; a crawler reports what actually does. When a site expands quickly, the map is the quality check that catches what the generation missed: the page that never got linked, the redirect that never got mapped, the section four clicks deeper than anyone intended.
  3. Client-side analysis is the private option. Most AI-powered site tools require sending your URLs and structure to someone else’s servers. For a staging site, an unannounced rebrand, or a regulated organization’s architecture, that is a real cost. This tool processes everything in your browser, so structure can be audited without a byte of it leaving the machine.

None of this replaces the classic reasons structure matters: visitors finding pages, search engines ranking them. It adds a newer one with a sharper edge, because in an AI answer there is no page two.

A note on what the grade is for

The tool tells you what is disconnected, broken, and buried. It does not tell you why it happened or what order to fix things in, and it cannot rebuild the message, the funnel, or the tracking around the structure. That is the work. If your map came back with a band of orphans and a grade you did not expect, that is usually the fingerprint of a past migration or redesign, and it is exactly the problem rebrand and site transfer marketing exists to prevent and repair.

Book a consult, bring your map, and we will read it together.