You can treat cocon sémantique as a checkbox, or as a design decision. The first path is fast and ineffective; the second takes an hour of thought and lasts for years.
The mistakes we see most often
Three mistakes recur with surprising regularity: applying a recommendation without checking it fits the context, changing several parameters at once which makes measurement impossible, and giving up after six weeks because nothing moved.
The third is the most expensive. Effects are rarely visible before a full recrawl and re-evaluation cycle, which often takes two to three months on a mid-sized site.
The real cost of doing nothing
An untreated technical problem does not stay stable: it spreads. Every new piece published on a faulty structure inherits the defect, and remediation cost grows with volume.
That is why fixing a structural flaw on a hundred-page site beats waiting until you have a thousand. The work is the same; the cleanup is not.
Keeping the structure alive
A cocoon built once then forgotten degrades with every publication. The habit to install is simple: before publishing, every new piece must know which set it belongs to and who will link to it.
Asking that upfront avoids accumulating isolated content that will need attaching later, at a far higher cost.
Automate what should be automated
Anything that must be repeated on every publication will eventually be forgotten. The rule is simple: if a check depends on human vigilance, it will fail on a deadline day.
Move those controls into the tool — validation on save, automatic alert, a correct default — rather than into a procedure nobody re-reads.
Spotting orphan pages
A page no internal link points to technically exists but has almost no chance of performing. On older sites they often represent a fifth of published content.
Identifying them is easy; deciding their fate is not. Each orphan must be attached, merged or removed — leaving it as is, is not a neutral option.
Check before moving on
- Automate every repetitive check
- Measure before, not after
- Handle what affects the most pages first
- Schedule a review in six months
- Remove before adding
A concrete example
On an editorial site of roughly four hundred pages we supported, methodically applying this principle produced a measurable gain after eleven weeks: more pages actually crawled, and progress on mid-tail queries.
No spectacular action was taken. Most of the work consisted of removing what was in the way, before adding anything at all.
A silo is not a folder
The most frequent confusion is believing a cocoon is built in the URL tree. It is built in contextual links, inside the text, where they actually carry meaning.
A clean URL structure helps readability, but it is internal linking that transmits authority and signals thematic kinship.
The three-link rule
Each secondary piece points to its main page, to a sibling of the same level, and receives at least one link from the main page. That minimal reciprocity is enough to form a coherent set.
Beyond that, extra links quickly show diminishing returns, and end up diluting the very signal you meant to concentrate.
What it changes for the reader
It is tempting to treat this as a purely technical constraint. In practice, each of these optimisations translates into a clearer, faster or more predictable experience for the person reading.
That is precisely why engines care about these criteria: they approximate, imperfectly, genuine user satisfaction.
Read next
- Cannibalisation : quand vos pages se battent entre elles
- La page pilier : ni fourre-tout, ni sommaire
- Maillage interne : la méthode des ancres utiles
Key takeaways
None of these measures produces a spectacular effect on its own. It is their accumulation, sustained over months, that durably moves rankings. Start with whichever point is closest to your current situation, measure, then move on.
