Guide 09 · Evidence review
Flat URL structure: should URLs be shallow for AI crawlers?
It makes no measurable difference. URL path depth correlates with citation at r = +0.002 — statistically indistinguishable from zero. /guides/topic/subtopic/ is not penalised against /subtopic/.
One URL property does show an effect, and it is not depth.
What about query strings?
URLs carrying a query string are associated with roughly 24% fewer citations. So the practical rule is not “keep it shallow”, it is “keep it a clean path”.
The likely mechanism is mundane rather than mysterious: query strings mark URLs that are often duplicates, filtered views, tracking variants or session-specific, and a crawler has ordinary reasons to treat those as lower-value. That is a reason to eliminate unnecessary parameters, not to fear the character.
| URL shape | Evidence | Do |
|---|---|---|
/a/b/c/ nested path | r = +0.002 | Nothing |
/a/ flat path | Same | Nothing |
/a/?utm_source=… | Associated with 24% fewer citations | Canonicalise |
/?p=123 unrewritten | Same class | Turn on pretty permalinks |
What does that mean for a WordPress site?
Almost nothing, if your permalinks are already set to post name. That is the default advice for other reasons and it happens to cover this one.
Three things are worth a glance:
Custom post type prefixes are free. If your docs live at /docs/getting-started/installation/, leave them there. Depth costs nothing and a legible hierarchy is worth more to a human reader than a saved slash.
Canonical tags do the work on parameters. Filtered archives, paginated views and campaign URLs will exist. Point their canonical at the clean path and the duplicate versions stop competing with it. This is what canonical is for and it is one setting in any SEO plugin.
Do not restructure an existing site for this. Changing URLs means redirects, and a redirect chain is a real cost against a benefit measured at r = +0.002. Rewriting a working URL structure to save a directory level would be spending something to gain nothing.
“Keep URLs shallow” is inherited from classic SEO, where the reasoning was about crawl depth and internal link equity. It was never a strong effect there either, and it does not transfer. This is the pattern behind most GEO advice worth checking: a technique from one system, assumed to work in another, never retested.
What this page cannot tell you
Both figures are correlational, from observational analysis of cited URLs. The 24% query-string association in particular is very likely explained by what kind of page tends to carry a parameter rather than by the parameter itself — which is why the recommendation is to canonicalise duplicates, not to treat the character as harmful.
An r of +0.002 is the clearest kind of result on this whole list: not a small effect that might matter at scale, but a number consistent with no relationship at all. Of everything in the GEO evidence review, this is the item I would most confidently tell you to stop thinking about.
How this was made: figures come from our own runs in the engines’ public interfaces or from the studies named above. Drafted with an AI assistant (Claude) and reviewed by Alauddin Chowdhury. Method and AI use.
Worth checking what actually is costing you
Send me your wordpress.org slug. I will check the things that are binary rather than marginal — certificate, redirect chain, robots rules, whether your server actually answers each retrieval crawler, and whether your content is in the HTML they get. Free, no call, usually within a day.