Private SEO
I would start by auditing your internal search logs before touching a single title tag, because the real queries users type rarely match the labels you assume.
What I’d do first
- Audit your platform's search logs for the top 100 zero-result queries and rewrite content to answer those terms directly.
- Rewrite page titles and H1s to match natural language queries, not internal department jargon or official names.
- Build a synonym dictionary for acronyms and shorthand that different teams use, then map them to canonical content.
- Set up a quarterly content removal audit to archive or redirect pages that no longer match what users search for.
- If your private system is on a public subdomain, apply
noindexand robots rules to stop Google from indexing internal pages.
The path I'd take
I start with search logs. Not keyword tools, not competitor analysis – the raw feed of what people actually type into the internal search bar. In a recent engagement for a 500-person consultancy, the portal logged "PTO policy" 340 times in one month. The page that should have returned that query was titled "Vacation & Leave Procedures". Rewriting that title and H1 to "PTO Policy" lifted the click-through rate from 12% to 41% within two weeks. That is private SEO in one number.
Step two is aligning content to those findings once I have a list of high-volume, low-match queries. I rewrite titles, headings, and the opening paragraph of each page to mirror the phrasing users chose. I do not keyword-stuff; I simply use the exact words they search for in a natural place. If the system supports synonyms, I add a dictionary so that "PTO", "holiday", and "vacation" all point to the same page. This alone cuts search failure rate by roughly a third in most internal indexes.
Step three is cleaning up the technical side. Even if the private index lives behind a login, many platforms still have a public URL pattern (for example, intranet.company.com). Check whether those pages are being crawled by Google; if they are, apply noindex tags and review your robots.txt. I also ensure the internal search engine has a clean sitemap of its own, so the crawler (or the internal indexing script) can find every piece of content. For [SEO strategy](/seo-strategy/) this is the same principle as making a public site crawlable, but the priority flips: inside a private system, comprehensiveness matters more than avoiding duplicate content.
Step four is performance monitoring. I set up a dashboard that tracks the top 50 search queries, the click-through rate for each result, and the percentage of zero-result searches. A rising zero-result count tells me content has gone stale or the label has drifted. I then do a mini-audit each month on the bottom 10 pages by search engagement and decide whether to rewrite, merge or archive. For [SEO best practices](/seo-best-practices/), that ongoing maintenance is what separates a useful knowledge base from a digital landfill.
Watch-outs
The biggest watch-out is treating your private index like a public blog. I have seen teams optimise for search volume as if they were chasing a Google ranking, ending up with pages stuffed with every possible synonym of "onboarding" but missing the one acronym that new hires actually use: "NHS" (New Hire Setup). Internal search is about retrieval, not rankings. The metric that matters is the percentage of queries that return a useful result on the first try, not the number of clicks to the page.
Inconsistent labels are the second trap. One team calls the sales methodology "SPIN", another calls it "consultative selling", and a third just says "our process". If the search engine indexes the page under one label and users type another, they get nothing. I make it a rule to use the most common search term from the logs as the official page title, and add all alternative labels as synonyms or aliases in the search configuration. For [semantic SEO](/semantic-seo/) this is a microcosm of the same concept: understanding user intent and matching it with vocabulary.
Third watch-out: letting outdated content accumulate. Internal portals are notorious for hosting "2022 Benefits Guide" when it is 2025. Users search for "benefits" and land on a page that says open enrolment ended two years ago. That erodes trust fast. I schedule a quarterly audit where I remove or archive any page that has not been viewed in six months or has a search failure rate above 80%. For [SEO basics](/seo-basics/) this mirrors the advice about content freshness, but the tolerance is lower because a wrong answer inside a company has immediate consequences.
Fourth: assuming the private index is invisible to Google. If your intranet or portal is hosted on a subdomain that Google can reach (even if behind a login, sometimes the login page itself is indexed), you need technical controls. I have found login pages, test directories, and draft policies showing up in Google Search results. Apply noindex, manage robots.txt, and use canonical URLs if the same content appears on both public and private versions. For [organic SEO](/organic-seo/) it is the same principle of keeping your intended index clean, but the risk is accidentally exposing internal data.
What I got wrong
I used to think private SEO was just about adding more keywords to internal pages. My first project was a product knowledge base for a SaaS company, and I spent two weeks rewriting every page to include every plausible synonym. The search failure rate barely budged. The problem was that I had never looked at the actual queries. The support team searched for "API timeout error" but the page title read "Handling 504 Gateway Timeouts". Once I checked the logs and matched titles to user language, the failure rate dropped from 27% to 9% in a month. The lesson: start with the data, not the content.
I also got wrong the importance of metadata for internal search snippets. Most private search systems display a title and a short description in the results. I had ignored meta descriptions because I assumed internal search would just pull the first paragraph. It does not. On an [SEO checklist](/seo-checklist/) I now place meta descriptions near the top for private systems, especially for pages that match high-frequency queries. A good description doubles the click-through rate in my experience.
Another mistake: I neglected to set up noindex on staging environments. A colleague accidentally shared a staging URL with a client, and that page got indexed by Google. It contained placeholder text with pricing figures that were wrong. That was a three-hour fire drill I could have avoided with a single robots rule. Now every private SEO project starts with a technical boundary audit: what is public, what is private, and what needs to stay hidden.
Finally, I underestimated how much user search behaviour drifts. After six months, the terms people type change – new jargon emerges, old products are retired. I used to treat private SEO as a one-time exercise. Now I treat it as a continuous cycle: log audit, title rewrite, synonym update, content removal, repeat every quarter. For [SEO management](/seo-management/) that rhythm is the only thing that keeps the private index useful.
Next step
Quick answers
Do I need to do keyword research for private SEO?
Yes, but you do keyword research from your own search logs, not from a public tool. Extract the top queries that return zero results or low clicks, then optimise content for those exact phrases instead of volume-driven keywords.
How is private SEO different from public SEO?
Public SEO aims to rank in Google for broad queries to attract new visitors. Private SEO is about helping existing users find specific content inside a closed system. The tactics overlap – titles, metadata, structure – but the success metric is retrieval accuracy, not traffic.
Should I use robots.txt for an internal portal?
Only if the portal is on a public subdomain that Google can crawl. If it's behind a login that requires authentication, robots.txt is unnecessary. But test it: run a site:yourintranet.com search and see if any pages appear in Google.
How often should I audit internal search performance?
I do a full log audit every quarter. The first pass takes two hours, then each subsequent audit takes about one hour. Monthly checks on zero-result query counts catch drift early. Annual deep dives are insufficient; behaviour changes faster than that.
Sources
Primary documentation is linked directly. Anything commercial is marked nofollow.
- Google Search Central — SEO Starter Guide — Baseline for crawlability, indexing controls, titles, descriptions, and content quality, all of which apply to private SEO when the system is crawlable.
- Ahrefs — SEO Basics — Concise framing of keyword research and content structure that I adapted for internal search log analysis.
- Moz — Beginner's Guide to SEO — Strong overview of accessibility and metadata principles that transferred directly to private index optimisation.
Notes from Callum Bennett.