How do you Measure Website Structure and Authority
A high page count does not guarantee that high-value URLs hold strong structural authority. Through internal link networks, authority and topic relationships flow across your architecture. Analyzing this flow reveals whether your site design actually supports your core conversion pages or diverts value elsewhere.
Three websites, one pattern
In each of these, the pages the business cared about most were the pages its own links treated as least important. Here’s what the analysis found.
Used cars: the site grew, its structure didn’t
A used car dealer was ranking well on page one, then slowly started slipping. Nothing was broken. But the main Used Cars page, the reason the business exists, held under 4% of the site’s structural weight, less than its contact and utility pages. Every new listing had quietly made it worse.
Pharmacy: 30,000 links pointing nowhere in particular
A UK pharmacy website had 374 pages and over 30,000 internal links, around 80 per page. On paper, that’s a strong site. Yet it sat on page 2 for the search it wanted most. The problem wasn’t a lack of authority. It was that the site treated almost every page as equally important.
Holiday travel: the most important page was the Privacy Policy
According to its own internal links, the most important page on this holiday website wasn’t a destination or a package. It was the Privacy Policy. The pages that actually sell holidays sat right at the edges of the site, and the route to booking was one of its weakest paths.
How Your Website Is Being Interpreted on Search Systems
Modeling the website as a directed graph and uses graph theory, probability and Markov chains to examine where structural authority accumulates, how it moves between pages and whether commercially important areas are being reinforced by the wider website. This provides a mathematical way of examining how search systems interpret website structure.
These calculations do not attempt to reproduce the private algorithms used by Google or other search systems. They measure something we can observe and calculate: the structure those systems encounter when they move through a website. This distinction is central to the mathematics that explain website search rankings.
PageRank and Structural Authority
PageRank introduced one of the most important mathematical ideas in search: links can be treated as pathways through which importance is distributed across a network. A page receiving links from structurally important pages occupies a different position from a page receiving few links, or links primarily from peripheral parts of the website. This means that simply creating an important commercial or informational page does not make it structurally important.
The rest of the website must reinforce it.
This movement of importance through internal links can be examined as structural authority flow: where authority originates, where it travels and which pages ultimately accumulate it.
How the calcululation works
The website is converted into a directed graph:
Pages = nodes
Internal links = directed edges
The resulting network can then be analysed using PageRank-style probability calculations to determine where structural authority naturally accumulates.
Rather than asking: Which pages do we want to be important?
the calculation asks: Which pages does the structure of the website actually make important?
The difference between those two answers can reveal one of the most significant structural weaknesses within a website.
What This Can Reveal on a Real Website
The mathematics becomes particularly useful when the calculated structure is compared with the commercial purpose of the website. The following examples demonstrate three different structural problems uncovered through this approach.
The Mathematics We Use to Measure Website Structure
No single calculation explains a website’s structure. Different mathematical measures reveal different parts of the same network — how probability moves, where it accumulates, how concentrated it becomes and whether it reaches the pages that matter.
Ranking Probability
Ranking probability measures how likely individual pages are to be encountered as probability moves through the internal link network. It can reveal an important difference between the pages a business considers important and the pages its website actually makes structurally prominent.
The Random Surfer Model
The Random Surfer model provides a simple way of understanding that probability. Imagine repeatedly following links from page to page. Some destinations will naturally be encountered more frequently because of their position and connections within the network. This is one of the foundations for understanding Markov chains and the mathematics of ranking.
The Damping Factor
The original PageRank model used a damping factor, conventionally illustrated as 0.85, to account for the probability that movement does not continue indefinitely along links. We do not assume that Google uses this value today. Its usefulness is mathematical: it demonstrates how distance and network structure affect the distribution of probability and why deeply positioned pages can become structurally weak.
Stationary Distribution
When probability is repeatedly passed through a sufficiently connected network, it can eventually settle into a relatively stable distribution. This stationary distribution shows which pages or areas the structure naturally favours. If the underlying network remains largely unchanged, its distribution may also remain similar — one structural explanation for why website rankings plateau.
Structural Entropy
Entropy measures how concentrated or dispersed a probability distribution is. Applied to a website, it can help distinguish a network with clear structural direction from one where probability is spread across many competing destinations. The issue is not simply how many pages exist, but how clearly the structure distinguishes what matters most.
The Authority Core
The Authority Core is the smaller group of pages representing the website’s principal subjects, services or commercial purpose. Once these pages are identified, the wider network can be measured against them: does the website actually reinforce the pages it considers strategically important?
Authority Core Capture
Authority Core Capture measures how much structural movement reaches or remains within those important pages. A website may contain thousands of internal links while relatively little of its structural probability reaches its intended core.
Supporting Content → Authority Transfer
This measures whether supporting pages reinforce the Authority Core or primarily circulate probability amongst themselves. Comparing Supporting → Supporting with Supporting → Authority shows whether supporting content is helping move structural strength towards the intended destinations.
A falling Supporting → Supporting rate, for example, can be positive when Supporting → Authority is increasing. The relationship between the measurements matters more than either percentage alone.
Observed Search-System Movement
The internal link graph tells us what the structure makes possible or probable. Server-side measurement adds a different form of evidence: what verified search systems actually do when they move through the website.
Page-to-page requests are recorded as observed transitions and kept separate from the theoretical structural model. Crawler identity and verification establish who made the request; they do not alter the mathematics of how that system moved.
The Website Search Behaviour Matrix
Observed transitions can then be grouped according to the structural role of the pages involved, producing movements such as:
- Authority → Authority
- Supporting → Authority
- Supporting → Supporting
- Entry → Exit
This turns individual crawler requests into a behavioural model and allows us to ask a more useful question: what pattern of movement is developing across the website?
Measuring Change Over Time
A single calculation provides a snapshot. Repeated measurement shows direction. Authority Core Capture, Authority retention, Supporting → Authority transfer, Supporting → Supporting retention and equilibrium can be compared over time to see whether the structure is actually changing.
This can also provide useful evidence when investigating why website rankings change.
The individual numbers are not the conclusion. What matters is the developing relationship between them.
What All These Measurements Tell Us
These calculations do not reveal the private algorithms used by Google, Bing or AI search systems. They measure something we can examine directly: the website those systems encounter.
Graph theory maps the relationships. Probability shows how structural importance is distributed. Markov chains show where that probability settles. Entropy measures how concentrated it is. Authority Core measurements show whether it reaches the intended pages. Server-side observations show how verified systems actually move through the network.
Together they replace an assumption:
“These are our most important pages.”
with a question that can be measured:
“Does the structure of the website actually support that conclusion?”
This is the basis of how the structural analysis works in practice.