Reading the link graph

By Bill Hartzer · Hartzer Consulting · 25+ years in search, domains and expert witness work

Almost everything published about links stops at tier 1

Open a backlink tool, type in a domain, and you get a list. Each row is a page that links to that domain, with a couple of authority numbers beside it. Sort the list, filter it, export it, and you have done what most link building education teaches. You have read tier 1.

The problem is that a list is a flattened version of something that is not a list. Links form a network. The page linking to you is itself linked to, by pages that are themselves linked to, and the value that reaches you depends on that upstream structure at least as much as on the page you can see. A link from a page that nothing links to passes close to nothing, whatever the domain-level number on the row says.

This is why people buy DR 60 links that do nothing. The metric is a property of the domain. The link sits on a page published an hour ago, absent from the navigation, absent from the sitemap, with no inbound links of its own. Every number on the row is accurate. The row is still worthless.

A tier 1 link means nothing until you know what links to it. That is the argument of this module, and every lesson in it serves that one point.

What reading at depth actually gives you

Depth is not a curiosity. Four things become visible at tier 2 and beyond that are invisible at tier 1:

• Whether a profile was earned or assembled. Earned profiles are irregular. Assembled ones repeat — the same link count per domain, the same metric band, the same publication rhythm.
• Whether a cluster of apparently independent sites is one operation. Seven domains with the same Trust Flow to within a point or two are not seven decisions. They are one.
• Whether a linking page is alive in the graph at all. A page with no inbound links is a dead end, and a dead end passes almost nothing forward.
• Where the risk in your own profile actually sits — and, just as usefully, where it does not, because most of what looks alarming at depth is not yours and never was.

The last point is worth stating early, because it upends how most people react to a deep audit. You control your tier 1. You choose, earn or place those links. From tier 2 outward you control nothing. Anyone can point a network at a site that links to you, and there is no mechanism by which you would know or could stop it. Junk at depth is a fact about the internet, not a symptom of something you did.

The two profiles used throughout this module

Every worked example here comes from real Majestic data captured in August 2026, on two domains chosen because they sit at opposite ends of the same scale.

billhartzer.com is my own site. Twenty years of publishing, 22,260 referring domains in the Historic index, and — this is the part that makes it useful — not one link on it was ever bought. It is a rare specimen: a large profile with no manipulation in it, which makes it a usable control for what unmanipulated actually looks like.

1-website-promotion-internet-marketing-services.com is an abandoned exact-match domain from the era when that kind of name was a strategy. Its profile was manufactured, and it has since decayed almost completely. It shows you the shape of bought links both while they work and long after they stop.

One thing I want to say plainly before Lesson 3.4, because it is the reason that lesson is credible. There is an identifiable link network sitting in my own site's tier 2. I did not put it there, I cannot remove it, and it is entirely normal for any domain with a long history. I am going to show it to you by name rather than pick a stranger's site and pretend my own is spotless.

What you need, and what transfers

The worked examples use Majestic, for a specific reason: its Link Graph tiers a profile natively. You move a slider and the whole graph is re-selected at the depth you asked for, rather than walking a chain of lookups by hand. Nothing else I use does that as directly, and it changes the method — Lesson 3.2 is built on it.

The tools that appear across the six lessons are Site Explorer, the Link Graph, the Topics tab, Anchor Text, Related Sites, Clique Hunter, Mutual Links, Neighbourhood Checker, Compare Sites, Flow Metric History and Backlink History. Several are metered, so plan a session rather than clicking around.

The reasoning transfers even where the tooling does not. If your tool only gives you tier 1, you can still reconstruct a tier 2 by exporting the referring domains and running the strongest twenty through the same tool one at a time. It is slower and it samples rather than surveys, but the questions you ask of the result are identical. Disclosure: I have been a Majestic Brand Ambassador for over fifteen years. I also tell you in Lesson 3.3 where its metrics mislead, which is the only version of that disclosure worth anything.

The lessons

1. 3.1What the link graph isNodes, edges, tiers, and why Fresh and Historic are two indexes rather than two eras.
2. 3.2Tiers two through fiveThe tier slider, links per domain by tier, and why you do not control your own tier 2.
3. 3.3Trust Flow, Citation Flow and Topical Trust FlowThe flow metrics, the TF/CF ratio, and sorting Topical Trust Flow by value not volume.
4. 3.4Detecting link networks and footprintsMetric signatures, shared infrastructure, cadence, and the CDN problem with IP checks.
5. 3.5Related sites and co-citationReading the company a site keeps, with Related Sites, Clique Hunter and mutual links.
6. 3.6Running a tiered link auditThe full sequence, a scoring approach, and what to put in the report — and leave out.