The Ahrefs internal links report: how to use it, and where it stops

Knowing which report answers which question saves most of the confusion:

• chevron_rightSite Audit → Internal pages — the crawl table. Per-URL inlink and outlink counts, depth, and status. This is your raw data.
• chevron_rightSite Audit → Link Opportunities — the suggestion engine. Finds phrases on your pages that could link to a page ranking for that phrase.
• chevron_rightSite Audit → Issues (Links) — the problem list. Orphan pages, broken internal links, redirect chains, nofollow internal links, pages with only one internal link.
• chevron_rightSite Explorer → Internal backlinks — per-URL. Which of your own pages link to one specific URL, and with what anchor text.

Site Audit reports run off your own crawl, so they need a completed crawl before they show anything. Site Explorer runs off Ahrefs' index, so it works immediately but reflects what Ahrefs has crawled rather than your site as it stands today. That difference explains most "why is this empty" and "why do these two numbers disagree" confusion.

This is the report most people are actually looking for, and it is the one most likely to come back empty for reasons that have nothing to do with your site.

1. The project must be set up and verified in Site Audit — a bare Site Explorer lookup will not do.
2. A crawl must have completed. Not started, completed. The report is built from crawl output.
3. Your crawl budget has to cover the pages you care about. If the crawl stopped short, suggestions for the uncrawled section simply do not exist.
4. The target pages need to rank for something. The engine matches phrases on your pages against keywords your other pages already rank for — a page that ranks for nothing is not a candidate target.

Each row proposes one link. The columns worth attention:

• chevron_rightKeyword context — the actual sentence on the source page containing the phrase. This is the most useful column, because it tells you whether a link there would read naturally or be shoehorned in.
• chevron_rightSource page — the page that would carry the new link.
• chevron_rightTarget page — the page that would receive it.
• chevron_rightTarget keyword — the term the target already ranks for, which is why it was matched.
• chevron_rightSearch volume — volume for that keyword, not a measure of how much the link would help.

Read Keyword context before anything else. The engine matches text, so it will happily propose a link from a sentence where the phrase appears incidentally — a passing mention in an unrelated paragraph, or the same phrase used in a different sense. A link that does not fit the sentence is worse than no link, because it reads as manipulation to a reader and passes a weak relevance signal.

1. Decide the target first. Pick the page you want to rank better — usually something on page 2 with impressions in Search Console. Do not start from the opportunities list.
2. Filter Link Opportunities to that target page.
3. Read the Keyword context for each suggestion and discard the ones that do not fit the sentence.
4. Check the source pages in Site Audit's Internal pages table. A suggestion from a page with real inbound links is worth more than one from a page nothing links to.
5. Place two or three good links rather than every suggestion. Diminishing returns set in quickly and a burst of links to one page in one day is a poor pattern.
6. Re-crawl, then check the target's position and impressions in Search Console a few weeks later.

Step 4 is the one the report will not do for you, and it matters more than the rest combined. The suggestion list treats every source page as equivalent — see the next section for why that is a real gap rather than a nitpick.

Site Audit has an orphan pages issue, and it carries the same constraint every crawler has: a page with no internal links cannot be reached by following links, so the crawl alone can never find one. Ahrefs fills the gap by comparing the crawl against external sources — your sitemap, and Google Search Console or Analytics if you connect them in the project settings.

Connect those sources before the crawl, not after. It is the same workflow as finding orphan pages in Screaming Frog, and the same caveat applies: a URL in your sitemap that the crawl never reached is a candidate, not a verdict. Paginated URLs, deliberately unlinked landing pages and stale sitemap entries all show up in that list.

When the question is about a single URL — why is this page not ranking, what links to it — Site Explorer is faster than Site Audit and needs no crawl.

1. Paste the exact URL into Site Explorer and set the mode to Exact URL, not Domain. Getting this wrong is the usual reason the numbers look absurd.
2. Open Backlinks → Internal backlinks.
3. Read the anchor text column, not just the count.

The anchor column is where the useful signal is. A page with thirty internal links whose anchors are all the site name, “read more”, or the page title repeated identically is being linked from templates rather than from content. That is a different situation from thirty links with varied, descriptive anchors, and only the anchor text distinguishes them.

Why the two reports disagree

Site Audit and Site Explorer routinely show different inlink counts for the same page, which looks like a bug and is not. Site Audit reflects your most recent crawl with your crawl settings and budget. Site Explorer reflects Ahrefs' own index, crawled on its own schedule, and only covers pages Ahrefs has found.

• chevron_rightSite Audit higher — Ahrefs' index is stale, or has not crawled a section of your site.
• chevron_rightSite Explorer higher — your crawl budget stopped short, or crawl settings excluded pages that do link to the target.
• chevron_rightBoth low, page ranks fine — check whether the links are rendered client-side, in which case neither crawler necessarily saw them.

For deciding what to fix, trust Site Audit: it reflects the site as it is now. For seeing what a search engine's index currently believes, Site Explorer is the closer proxy.

Both of these are consequences of how the suggestion engine works rather than oversights, and knowing them changes how much weight you put on the output.

1. Inlink count is not authority

The Internal pages table gives you an inlink count. Two pages showing 40 inlinks each can be in completely different positions: one has 40 links from a sitewide navigation, the other has 40 contextual links from relevant articles. The nav links pass a fraction of what the contextual ones do, and a count cannot distinguish them — the distinction is covered in contextual vs navigation links.

URL Rating is not the fix either. UR is dominated by external backlinks, so it tells you about a page's off-site position rather than how internal authority is distributed across your own site. For that you need a model over your own link graph — the idea behind PageRank for modern SEO.

2. Suggestions are matched, not ranked by impact

Link Opportunities finds where a phrase appears. It does not compute how much authority a given link would move, because that requires modelling the whole graph: how much the source page holds, how many other links it already passes through, and how starved the target is. Two consequences follow.

• chevron_rightA suggestion from a strong page and one from a page nothing links to appear side by side, ranked by keyword volume rather than by effect.
• chevron_rightThe best link is often one the report cannot see — a genuinely relevant page that happens not to contain the target's exact phrase. Text matching cannot propose a link where the wording does not already exist.

If you only take one thing from this: pick the target page yourself, using Search Console impressions, and use Link Opportunities to find placements for that target. Used that way — target first, suggestions second — the report is a good tool and the impact-ranking gap mostly stops mattering, because you supplied the prioritisation.

Where that breaks down is at scale. Choosing targets by hand across a few hundred pages is the part that does not survive a real site, and it is the part a graph model does well: which pages hold authority, which are starved, and which single link moves the most. That is what internal authority flow means in practice, and it is what RankForge computes — it crawls the site, models the graph, and returns ranked recommendations naming source page, target page, anchor text and paragraph.

The two are not substitutes. Ahrefs is a backlink and keyword platform with internal-link reporting attached; RankForge does not have a backlink index and is not trying to. If you want the direct feature-by-feature version, RankForge vs Ahrefs lays it out including where Ahrefs is clearly stronger.