SaaS SEO: the four page engines, competitor queries, and where programmatic pages stop working

SaaS SEO is organic acquisition for a product people can sign up to themselves, so every page is judged on signups and activated accounts rather than on sessions. Four engines do the work, and they are worth building in this order: product pages for the way buyers describe the job, competitor queries such as alternatives and comparisons, a free tool or template that earns links on its own, and programmatic pages where each page carries genuinely different data. The two problems unique to SaaS are that nobody searches for a new category name, so you target the problem language and the incumbent instead, and that the marketing site and the application are different technical animals that need different indexing rules.

This page covers software specific strategy. For the wider committee and sales cycle questions that apply to any business selling to businesses, see a guide to B2B SEO.

The four engines

Engine	Pages	What success looks like
Job and use case	One page per job your product does, named the way a user would name it rather than the way your roadmap does	Signups from people who arrived with the problem already defined
Competitor and category	Alternatives to the incumbent, head to head comparisons, "best tools for X", your own pricing page	The highest intent traffic on the site, and the pages sales will quote in deals
Free tool or template	A calculator, generator, checker or template that solves a slice of the problem with no account	Links and repeat visits, and a natural step up into the product
Programmatic	One page per integration, per template, per data record, where each page has real content of its own	Long tail coverage at a volume no writer could produce by hand

Most SaaS companies start with a blog, which is the fourth best of these and the slowest to pay back. Build the use case and competitor pages first.

When nobody searches for your category

A new category name has no volume, and buying ads against it does not create demand either. Three things work instead.

• Target the problem, not the product. People search what hurts. "Spreadsheet keeps breaking when the team edits it" is a real search behaviour and your category name is not, yet.
• Target the thing you replace. The incumbent tool, the manual process, the agency, the spreadsheet. Alternatives and comparison pages for those capture people who already want a change.
• Build the definitional page anyway, and keep it the clearest explanation on the internet. It earns little traffic early and becomes the page everyone cites when the category does get searched.

Competitor pages, done properly

• Be accurate and keep it current. Check the competitor's live pricing and feature pages before publishing, and put a review date on the page. An out of date comparison is the fastest way to lose a technical buyer's trust.
• Say where they are better. Naming the cases where the other product is the right choice makes the rest of the page believable, and it stops wrong fit signups you would churn anyway.
• Never quote a price you have not verified on their own pricing page, and link to that page rather than restating the figure.
• Cover the full set: "alternative", "versus", "pricing", "reviews" and "competitors", for the two or three products your deals actually involve.

Programmatic pages, and the line not to cross

Generating pages from a database is legitimate when each page carries information a reader could not get elsewhere: a real integration with real setup steps, a real template, a real dataset, a real location with real inventory. It becomes a doorway when only one variable in a sentence changes, and doorway pages and scaled content abuse are both named in Google's spam policies.

• Set a quality bar before you generate anything: the page must answer the query on its own and have something unique on it beyond a swapped noun.
• Publish in batches and watch indexing. If Google indexes 200 of 2,000, the template is thin and more pages will not help.
• Give each page real internal links from somewhere a person would browse, not only from a sitemap.
• Prune. Pages that earn no impressions after six months are candidates to merge or remove.

The marketing site and the app are different jobs

• Keep marketing pages server rendered. Content that only appears after a client side fetch is content Google has to work harder to see. Google's JavaScript SEO basics covers the requirements.
• Keep the application out of the index. Logged in routes, share links, generated reports and password reset pages belong behind a noindex, and share URLs in particular can leak customer data into search results.
• Put the blog and the docs in subfolders on your main domain where you can. Google handles subdomains, and a subfolder keeps everything on one hostname, which is simpler to manage and to report on.
• Let the documentation be indexed. Docs rank for a huge amount of specific long tail and they are read by buyers during evaluation, not only by customers afterwards.
• Watch your status pages and changelogs. They get indexed, they get searched during outages, and they are frequently the first impression a prospect has of your reliability.

Measure signups and activation, not sessions

• Define the chain as key events in your analytics: signup, activation, paid conversion. Google's reference on key events covers the setup.
• Attribute by first landing page and keep it on the account record, so you can compare cohorts by acquisition page months later. This is the one piece of plumbing that makes SaaS SEO reporting possible.
• Report signups and activated accounts per page, then retention by acquisition page. A page can send plenty of signups that never activate, and that is a page targeting the wrong query.
• Keep query level data from Search Console split into branded and non branded. Branded growth often reflects your paid and product work rather than your SEO.

What people ask about SaaS SEO

How long does SaaS SEO take to pay back?

Position movement on new pages shows within weeks, first signups on commercial pages within two to three months, and a programme that compounds from about month six. Judge it on signups per page from the start so you are not waiting a year for a verdict.

Blog on a subfolder or a subdomain?

Use a subfolder on the main domain when your stack allows it. Google can handle both, and a single hostname is simpler to manage, link internally and report on.

Are competitor comparison pages a good idea?

Yes, when they are accurate, dated and honest about fit. Buyers search those comparisons regardless, so the only question is whether your page is in the set they read.

Should our documentation be indexed?

Yes. Docs cover an enormous range of specific questions and they are read during evaluation. Keep them on your main domain, keep them current, and link from the docs back to the relevant product pages.

Does programmatic SEO still work?

It works where each page has genuinely distinct and useful content. Integration pages, template libraries and real datasets qualify. A page per city with the city name swapped does not, and that pattern is named in Google's spam policies.

What should the call to action on a blog post be?

Something specific to the job the post describes, pointing at the part of the product that does it. A generic "book a demo" at the end of every article is the most common wasted conversion opportunity on SaaS sites.

A shortlist your buyers build in a chat window

A founder looking for a tool in your category asks an assistant for five options that fit a ten person team. The reply is a comparison table in words. Your alternatives page, your pricing page and third party reviews all feed it.

Test it with the prompts your buyers use: best tool for a use case, alternatives to a competitor, and which is cheapest. Ask each in three assistants and record the names. AI Knows Us runs the same prompts across ChatGPT, Gemini, Perplexity and Claude and lists the sources used.