The SaaS SEO strategy that actually produces pipeline

A workflow automation company came to us in early 2025 with 41,000 monthly organic sessions and eleven inbound demo requests a quarter. Their previous agency had published 240 blog posts in two years. The traffic chart looked like a hockey stick. The pipeline chart looked like a heart monitor on a corpse.

That gap is the whole problem with how SaaS SEO usually gets done, and it comes from a strategy decision made so early that nobody remembers making it.

The decision nobody remembers making

Somewhere in the first month of most SaaS content programs, someone opens a keyword tool, sorts by volume, and filters for terms with a difficulty score under 40. What comes back is a list of educational queries: what is X, how does Y work, X best practices, the ultimate guide to Z.

Those terms are winnable. They also belong almost entirely to people who are eight to eighteen months from buying anything, if they buy at all. A lot of them are students, job seekers, and competitors’ interns.

Then the program runs for two years, produces the traffic it promised, and gets cancelled because nobody in sales can name a deal it created.

The alternative is to start from the opposite end. Not “what can we rank for” but “what does someone type into a search box three weeks before they sign a contract.” Those queries are smaller, uglier, and harder to find in a keyword tool. They are also where nearly all of the money is.

What high-intent actually looks like in SaaS

For most software categories, the searches that precede a purchase fall into a handful of shapes. Comparison queries where your name sits next to a competitor’s. Alternative queries, which are people actively trying to leave a product they already pay for. Category-plus-qualifier queries, where the qualifier is an industry, a company size, a compliance requirement, or an integration. Pricing queries. Migration queries. And the unglamorous ones about whether your product does one specific thing the buyer needs.

Here is a real comparison from a client program, numbers rounded slightly:

Query	Monthly searches	Demo conversion	Est. annual pipeline
“what is revenue operations”	14,800	0.3%	$84,000
“clari alternative for mid-market”	90	7.1%	$310,000

The second keyword gets 0.6% of the traffic and produces almost four times the pipeline. Once you have seen that table for your own category, most content calendars start looking indefensible.

The catch is that the 90-search keyword took eleven weeks to find. It does not appear in Ahrefs at a useful volume, nobody has written about it, and the only reason we knew it mattered was that three separate prospects had said something close to it on recorded sales calls.

Where the good keywords are hiding

Keyword tools are built on clickstream and search-partner data that skews heavily toward high-volume consumer queries. In narrow B2B categories they are wrong constantly, usually in the direction of underreporting. A term the tool says has zero volume can have forty real searches a month, all from people with budget.

So we get keywords from four places, in this order of usefulness.

Recorded sales calls come first. The exact phrasing a prospect uses when they describe their problem is, with minor grammatical cleanup, the phrasing they typed into Google two months earlier. We listen to twenty calls in the first fortnight of every engagement, and the transcript search is the single highest-yield research activity in the whole process.

Search Console comes second, and specifically the queries with impressions and no clicks. Those are searches where you already appear somewhere on page two or three, which means the algorithm already associates you with the topic. They are the cheapest wins available and almost nobody mines them systematically.

Third, competitor SERPs. Not their keyword list, which the tools will give you badly, but the actual pages they have built. If a competitor has spent engineering time building 300 integration pages, they know something about demand in that pattern.

Fourth, and least reliably, the tools. Useful for sizing and for finding modifier patterns. Bad at telling you what is real in a small market.

Sequencing matters more than the list

Having the right keyword map does not help if you build it in the wrong order. We have watched companies with a perfectly good strategy fail because they started with the pillar content and got to the commercial pages in month nine, by which point the budget conversation had already gone badly.

The order that works, roughly:

Technical foundation first, because everything after it multiplies. This is four to six weeks of rendering fixes, indexation decisions, template changes, and internal link architecture. It produces no visible wins and it is the part clients most often ask to skip. On one recent engagement, a canonical tag misconfigured in the CMS had been pointing 180 use-case pages at the homepage for a year. Fixing that took an afternoon and moved more traffic than the next four months of content.

Commercial pages second. Comparison, alternatives, use cases, verticals, integrations. Low volume, fast to rank, high conversion. These usually start showing up between week eight and week fourteen, which conveniently lands right around the first serious “is this working” conversation.

Supporting editorial third, and only where it serves the commercial pages. Every article should have a job written into its brief: which money page it links to, which cluster it adds weight to, what it is meant to do. Content with no job is content nobody will maintain.

Authority building runs alongside from month three. In competitive categories this is the binding constraint, and no amount of content quality fixes a 40-versus-400 referring domain gap.

The traffic will go down, and that is fine

The uncomfortable part of doing this properly is that a real strategy usually starts by removing things.

On the automation client, we deleted or merged 118 of their 240 posts. Sessions dropped 12% over the engagement. Demo requests from organic went up 290%. The CMO had to explain a declining traffic number to her board twice before the pipeline number caught up, and she has since told us that was the hardest part of the year.

I do not have a clean way to make that easier. What helps a little is agreeing on the reporting metric before you start, in writing, and putting the traffic number somewhere below the fold in the monthly deck. What helps more is having one or two commercial pages ranking by month three so there is something concrete to point at.

If your board is going to fire you for a traffic decline regardless of pipeline, do not run this strategy. Run the other one and update your CV.

Measuring it without lying to yourself

Last-click attribution will tell you organic contributes almost nothing. In a five-month sales cycle with six touchpoints, the last click is nearly always a branded search or a direct visit, because by then the buyer knows your name.

The version that survives a CFO conversation has three parts. Session-level data pushed into the CRM so you can see influenced pipeline across the whole opportunity, not just the final touch. A self-reported attribution field on the demo form, which is unscientific and catches a surprising amount of what the tracking misses. And page-level analysis of which URLs appear in the journey of closed-won deals, which is the number that actually changes how sales treats your content.

On the Cadence HQ program, 68% of closed-won deals had at least one organic session in the opportunity history. Last-click attribution credited organic with 9%. Both numbers are technically true. Only one of them is useful for deciding next year’s budget.

What this looks like twelve months in

A functioning SaaS SEO program at the one-year mark has maybe 25 to 60 commercial pages ranking in the top three for terms with genuine purchase intent, a body of editorial content that exists to support them, a link profile that grew by nine to fourteen referring domains a month, and a monthly report that connects specific URLs to specific opportunities.

Total organic traffic might be up 80%. It might be flat. In one case it was down and the program was the best-performing marketing investment the company made that year.

The number to watch is how many opportunities in your CRM touched an organic page. If that is going up and the traffic is going down, you are doing it right and your reporting is finally honest.