The short version
Our internal rule is 60/20/20: 60% bottom of funnel, 20% middle, 20% top.
The 15 sites averaged roughly 15/20/65. Two thirds of everything was top of funnel.
None of the 15 had a link path from their blog content to the page that sells.
The ratio takes an hour to audit. The architecture almost never gets audited, and it is the part that decides revenue.
Why we run a 60/20/20 rule
Most B2B content calendars are built around what is easy to write. "What is X." "Five trends in Y." "The ultimate guide to Z." These pieces are useful, they rank, and they are almost never the reason anyone buys.
The pages that earn revenue are the commercial ones. A buyer two weeks from a decision reads a comparison or a cost page, not the ultimate guide. So we weight the calendar toward them.
60% bottom of funnel Where demand is captured. Comparison, cost, use-case, integration and alternatives pages.
20% middle of funnel Where the buyer works out how to solve the problem and which approach fits.
20% top of funnel Enough to build the authority that feeds the rest, and to give commercial pages something to link from.
The 60 is not arbitrary. For a $2M to $20M services or SaaS firm, commercial coverage is usually the gap. There are twenty or thirty commercial queries that matter, and most sites have written two of them. Until those are covered, more top-of-funnel content is effort spent on readers who are not ready.
What each funnel stage looks like in practice
Stage | What the buyer is asking | Page types | Example: bookkeeping firm | Example: QA automation SaaS |
|---|---|---|---|---|
Bottom of funnel | "Which one, and what will it cost me?" | Comparison, pricing and cost, use-case, integration, "best X for Y", alternatives |
|
|
Middle of funnel | "How do I solve this, and which approach fits?" | Problem-specific how-tos, frameworks, checklists, templates, "signs you need X" |
|
|
Top of funnel | "What is this, and why should I care?" | Definitions, trends, broad guides, industry data, opinion |
|
|
What 15 B2B sites actually publish
We classified every published article on 15 B2B sites by funnel stage. Here is the average against our rule.
Two thirds was top of funnel Definitions, trends, guides, listicles. Content a competitor could have written and a buyer reads once, eighteen months before they are in market.
Three of the 15 came close to a healthy mix Real comparison pages, cost pages, use-case and integration pages. Someone had mapped the commercial queries and written to them.
That is the content test, and three out of fifteen passed it. This is where most audits stop: count the articles, bucket them, recommend more bottom-of-funnel content. It is a spreadsheet column. It takes about an hour.
Then we looked at where the commercial pages pointed.
The architecture test, which nobody passed
A commercial article only earns revenue if a reader can get from it to the page that sells, and if the page that sells gets authority from it. That needs a link path in both directions. Of the three sites that passed the content test, here is what we found.
Service page to articles: zero links The service pages linked to none of their own commercial articles.
Articles to service page: zero in-body links The comparison page existed, ranked, and sat in a dead end. The cost page answered the question and offered nowhere to go.
Articles to each other: a closed loop Comparison to use-case to integration and round again, never touching the money page.
Related posts: wrong industry Modules pulled by recency or tag. A reader on a fintech comparison page was offered a healthcare guide.
The ratio is a spreadsheet column. Whether your content can reach your money page is the question that decides if any of it earns revenue.
Why architecture decides revenue
Three reasons, and they compound.
Readers follow links, not intentions A buyer on your "X vs Y" page who decides you are the right fit clicks whatever is in front of them. If every link goes to another article, they read it and leave. If one goes to the service page, a meaningful share go there.
Search engines move authority through links Most backlinks land on blog content, not the service page. Internal links are how that authority reaches the page competing for commercial queries. Without them, the service page is starved however good the articles are.
AI answer engines cite what is connected A service page cited by five of its own comparison and use-case pages reads as the hub of a cluster. One with no internal citations reads as an orphan. We see this directly in our AEO work: connected clusters get cited, isolated pages do not.
The three sites with good content were paying for articles that could not pass readers, authority or citations to the page that converts. The content was fine. The plumbing was missing.
Audit your own site in 20 minutes
No tools needed for the first pass. Tick each check your site fails.
Open your main service pageCount in-body links to specific commercial articles. Not the nav, not the footer, not "read our blog". Tick if you count zero.
Open your three best commercial articlesFind a contextual, in-body link to the service page, not the template CTA button. Tick if any of the three is missing one.
Check the related-posts moduleOn a commercial article, see what "related" serves. Tick if it shows a different industry or a top-of-funnel piece.
Count the orphansList your commercial articles and check whether any other page links to each one. Tick if any has zero inbound internal links.
0 of 4 checks failed. Fail more than one and your content is not reaching your money page.
To get your ratio as well, bucket your last 20 posts against the table above.
How to run the full audit with Claude
The manual pass covers four pages. The full audit covers every page, and Claude can do most of it. You need Claude with web access turned on.
Part 1: The ratio
About 10 minutes
Open yoursite.com/sitemap.xml and copy every blog and service URL, or export the list from your CMS.
Paste the list into Claude with this prompt. It reads the slugs and titles, opens the ambiguous pages, and returns your split against 60/20/20.
Prompt
Here is a list of URLs from my website. Classify each one as bottom of funnel, middle of funnel, or top of funnel using these definitions:
bottom = comparison, pricing, cost, use-case, integration, alternatives, or "best X for Y" pages
middle = how-to guides for a specific problem, frameworks, checklists, "signs you need X"
top = definitions, trends, broad guides, opinion
Where the URL is ambiguous, open the page and read the title and first paragraph. Return a table with URL, stage, and a one-line reason, then give me the percentage split.
[paste URLs]Part 2: The architecture
About 15 minutes
Give Claude your service page URL and the bottom-of-funnel URLs from Part 1.
Use this prompt to get a pass or fail table for every commercial page.
Prompt
Open my service page at [URL] and list every in-body link to a page on my own domain, ignoring nav and footer.
Then open each of these commercial articles [paste list] and for each one tell me:
1. Does it contain an in-body link to the service page?
2. Which other pages on my domain does it link to?
3. What does the related-posts module serve?
Finally, list any commercial article that none of the other pages link to.
Give me a table and a pass or fail against these rules: every commercial article links to the service page in the body, the service page links to every commercial article, no commercial article is an orphan.This works well for 20 to 40 pages in a normal chat. For a few hundred pages, export the internal links report from Screaming Frog or Ahrefs and upload the CSV with the same prompt, or ask Claude Code to write a small crawler that extracts internal links across the site.
Part 3: The fix plan
About 5 minutes
In the same chat, ask for a page-by-page edit list your writer or developer can work through in an afternoon.
Prompt
Based on the audit, give me a link map. For each commercial article, tell me which paragraph should link to the service page and suggest the anchor text. For the service page, write a short section that links back to each article, framed as evidence: how we compare, what it costs, who it is for.The whole thing takes under an hour. It is the same process we run on every client site before we write a single new piece.
Fixing it: the hub and spoke that actually works
Make the service page the hub Every commercial article in the cluster links to it in context. The service page links back to each article in a section that frames them as evidence.
Make the articles spokes, not a chain Spokes link to the hub and to one or two adjacent spokes. No closed loop that bypasses the hub.
Rebuild the related-posts logic Serve by cluster and funnel stage, not by date. A bottom-of-funnel reader sees other bottom-of-funnel pages in the same cluster, plus the service page.
Fix the ratio second, not first Ten more commercial pages on a site with no link paths is ten more dead ends. Connect what exists, then expand.
Re-audit every quarter Put the link map in every content brief, so nothing is published without its hub link and spoke links specified.
Ran the audit and didn't like the result?
Content ratio is a planning problem. Content architecture is a structural one. Inside the Leads Generator we fix both: content planned to a commercial ratio, built into clusters with the service page as the hub, and measured on pipeline rather than rankings.