Internal vs external link equity: what each one can do
External links bring authority in. Internal links decide where it goes. Fixing one never replaces the other, and the order matters when you triage.
ArchiLinks research team ·
Side by side
| Internal links | External links | |
|---|---|---|
| Who controls them | You, completely | Other sites, not you |
| What they do | Distribute authority, show structure, help discovery | Add authority to the site as a whole |
| Fastest way to change | One template edit, live in a day | Weeks or months of outreach |
| Typical failure | Orphans, deep pages, links hidden in JavaScript | Good links pointing at URLs that return 404 |
| How to check | Crawl the site, count inlinks per URL | Pull referring domains and linked pages from an index |
The order we would triage in
Start with the leak. Before you ask for a single new link in 2026, check links to my site that already exist and see where they land.
If the best-linked URL redirects through two hops, or returns a 404, outreach is premature. The same applies when the strongest page links to nothing important. Equity flows in by external links and then spreads along internal ones, much like water through pipes, so a blockage upstream wastes the whole effort. Simple.
Only once those are clean does it make sense to build more of either kind.
Our redirect map recipe and JavaScript audit cover the two commonest leaks. Moz and Ahrefs call the shared idea link equity; Google's own term is PageRank, first described in the 1998 paper by Page and Brin.
Myth vs reality
Myth: a flat site means no equity problem
Reality: flat structure shortens click depth, but it also dilutes whatever the hubs hold. A page with 400 links on it passes a sliver to each. Google's published guidance on crawling does not set a hard cap on links per page. The practical limit is usefulness for readers.
- Reality check: at DA 42 or DR 42 levels, a siloed structure can lift a section, but it will not replace the authority of a strong linked page.
- Mega menus are fine when they are real navigation; they are noise when they list every page.
What to measure for each
Internal: inlinks per URL
Export from your crawler. Look for money pages with fewer than three internal links, and for pages deeper than four clicks.
Internal: anchor variety
Generic anchors such as 'click here' tell search engines nothing. Descriptive, varied anchors help, but do not stuff them.
External: linking domains per URL
Pull toppages, then compare against the internal inlink counts. Pages rich in outside links and poor in inside ones are your quick wins.
External: anchor profile
anchors shows which phrases other sites use. A profile of only exact-match commercial anchors is a flag.
A small join: external value against internal links
Export inlinks from your crawler as a CSV, then join. This is an example; adapt the column names.
import pandas as pd
ext = pd.read_csv("toppages-2026-10-29.csv") # url, ref_domains
int_ = pd.read_csv("crawl-inlinks.csv") # url, inlinks
m = ext.merge(int_, on="url", how="left").fillna({"inlinks": 0})
# external value, few internal links: link these pages from your hubs
print(m[(m.ref_domains >= 5) & (m.inlinks <= 3)].sort_values("ref_domains", ascending=False).head(25))Internal linking questions
- Does an author box linking to a bio page add internal links?
- Yes, it adds a link on every post by that author, which can be useful for discovery. It does little for relevance when the anchor is only a name, so treat it as navigation and not as a ranking tactic.
- When does internal linking become a governance problem?
- Once more than one team edits templates and content. Then nobody owns hubs, anchor rules or orphan clean-up. Give one owner the link map and review it each quarter, like any other shared dependency.
- How do I tell circular linking from related content?
- Circular means pages link to each other with no route to a deeper or higher-value page. Related content is fine when the cluster also links up to its hub and out to commercial pages.