Why your documentation eats your traffic and what to do about it
The help centre collects 41% of organic traffic and produces 4% of signups. The first instinct is to noindex it. That's the worst available option, and here's why.
Documentation versus product pages
What the numbers look like
Take a typical B2B service that's been around five years and has accumulated a decent help centre. Look at how organic traffic splits by section and you'll usually see something like this: documentation takes around 41% of visits, the blog another 34%, and product and pricing pages together collect under ten.
Now look at signups across the same sections. Documentation: 4%. Blog: 3%. Pricing and homepage: 38%. Comparison pages, which account for 6% of traffic, produce 34% of signups.
Four fifths of the traffic arrives where almost nobody signs up. This isn't a quirk of one project, it's the norm for a mature SaaS.
Why this isn't a bug
The first reaction is “docs are stealing our traffic, we need to do something”. That reaction is wrong, because it conflates two different problems.
Docs collecting lots of visits and few signups is normal and correct. The people arriving are existing users solving a specific problem. They're already customers; there's nothing for them to sign up to. Conversion there is low by definition and raising it is pointless.
The real problem is different. Documentation often ranks for commercial queries – the ones asked by somebody still choosing a vendor. They ask what integrating with their CRM costs, land in the help centre, read a technical description of the process and leave. Not because the product was wrong, but because the page answered a different question.
Three ways to separate intent
Rewrite headings and internal links
The cheapest option and usually sufficient. A help article's heading should answer “how do I set this up”, not “what is this”. “Salesforce integration” becomes “How to connect Salesforce: step-by-step setup”. The first competes with your product page, the second doesn't.
At the same time, the help article should link to the product page above the fold, not in the footer. A line like “still choosing a tool? here's what the integration does” sends a misrouted visitor where they meant to go.
Finish your product pages
Often the help centre wins simply because the product page has no answer on it. No prices, no plan limits, no list of supported systems. Search gives the query to whichever page matches it more closely – and that turns out to be the docs.
The fix is adding blocks: a plan comparison table, limits, what's included and what isn't, a full filterable list of integrations. Dull work, but it produces most of the signup increase.
Separate them with schema
Help pages get TechArticle or HowTo; product pages get SoftwareApplication with a price. This won't reorder rankings on its own, but it helps search understand it's looking at different page types, and it noticeably affects which one ends up quoted in an AI answer.
What not to do
Noindex the documentation. The most common suggestion and the most destructive. Docs earn links, branded traffic and trust: a developer who found your help centre through search later brings you into their company. Removing it throws all of that away to solve a problem that rewriting headings would have fixed.
Move the help centre to a subdomain. The idea looks logical – separate the entities physically. In practice you lose some link equity, end up maintaining two domains and still don't solve the intent problem: the pages will keep ranking for the same queries from the subdomain.
Cut the volume. Deleting help articles because “they don't convert” is like taking the phone off reception because nobody buys over it.
How to tell it worked
The metric isn't “traffic to product pages”, it's the share of commercial queries your help centre shows for. It should fall. Measured from the same Search Console export you started with.
The second metric is signups at unchanged traffic. If you did this right, total visits stay flat or dip slightly while signups rise. On the projects where we've done this, the increase was 20 to 40% within two or three months, and it's the cheapest part of any SaaS engagement.
Finally, timing. Rewriting headings and links shows up in four to six weeks – fast, because the pages are already indexed and already ranking. Finishing product pages takes longer, three to four months.
We'll check your case in an hour
We'll pull the queries your help centre shows for and give you the commercial share. Free, and the report is yours.
Let us check your case
The site address is enough. We'll look and tell you whether you have the same problem.