Myth vs reality
Crawl depth and link equity: five myths, checked
Depth gets blamed for a lot. Some of it is deserved, and some of it is a symptom of something upstream that a crawl never shows.
Every audit report has a column for depth. It is easy to compute, so it gets treated as the cause of things.
In a link graph, though, depth is only a distance. Distance from where? That question decides whether the number means anything, and it is the thread through all five myths below.
We keep our own crawler's limits in view. It sees links between domains, not the clicks inside one site. For depth you still need Screaming Frog, Sitebulb or your own script, and what we add is the information about where equity starts. See entry points for how.
Five beliefs against what the data shows
| Belief | What holds up | How to test it |
|---|---|---|
| Anything past three clicks is invisible | Googlebot can follow deep paths; what thins out is attention and signals reaching them | Compare indexed share by depth in Search Console |
| Folder depth in the URL is click depth | They are different things; a page in /a/b/c/ can sit one click from home | Crawl and compare both columns |
| Flat architecture fixes everything | Flattening without hubs just makes a 2,000-link home page | Count outlinks per page before and after |
| Depth is measured from the home page only | External links make other pages the real starting points | Run depth from the entry pages instead |
| More internal links always help | Equity per link falls as the number of links on a page rises | Test one cluster, then wait one refresh |
Depth from the home page is the wrong origin if half your referring domains point at an article. Measure from the pages that actually receive links.
Rerunning depth from the inlets
Take the entry list from toppages and treat each URL as a source node. Then run a breadth-first walk over your crawl export and record the shortest click path from every entry page to each money page.
A page that is five clicks from the home page but one click from a heavily linked article is not deep at all. It is simply on a different branch of the delta.
Google's own public guidance, as we read it, treats clicks from a prominent page as the signal, and folder depth as cosmetic. The original PageRank paper of 1998 already modelled this as a walk, not as a directory tree.
Sources first, distances second. That is the method. Do it once and the depth column starts to explain things. The hub worked example puts it into practice.
A forty-minute test you can repeat
- 1
Export entry pages
Call
toppagesfor your domain and keep the top 50 URLs by referring domains. - 2
Export your crawl
Take the internal link edges from your crawler as a two-column list.
- 3
Walk the graph
From each entry page, compute shortest paths to the pages you want to rank.
- 4
Rank the gaps
Money pages with no path under four clicks from any entry page go on the fix list.
Related questions
Does crawl depth affect how quickly Google crawls a page?
It can. Pages far from any frequently crawled URL tend to be fetched less often. But a deep page that sits one click from a heavily linked entry page gets reached through that route, so measure from the inlets.
Should I flatten my URLs to fix depth?
Usually not. Changing folders creates redirects and risks the very entry pages that hold your links. Add links from existing hubs first, and only restructure URLs when you are migrating anyway.
How many internal links per page is too many?
There is no published threshold. Past a few hundred, each link carries less and the page gets hard to use. Group links into hubs instead of listing everything on the home page.
Can a backlink crawler show crawl depth?
Not for your internal pages. It records links between domains. Use it to find the starting nodes, then use your own crawler for the paths between them.