SaaS keyword research that maps to pipeline: a 7-step method

Most SaaS keyword research starts in a tool and ends with a spreadsheet sorted by search volume. That's how teams end up ranking for terms that bring traffic and no demos.

The problem isn't the tools; it's the starting point. Volume tells you how many people search for something. It doesn't tell you whether they'll ever buy software like yours.

This is the method we use in the Map stage of The Cited Pipeline Method. It starts with your revenue, uses your buyers' own words, adds the prompts they type into AI assistants, and scores every keyword by how likely it is to create pipeline. It takes a few days the first time and a couple of hours a month after that.

Step 1: Start from revenue, not from a tool

Before you open a keyword tool, write down three things:

• Who you win. The segments, company sizes and roles in your best closed-won deals over the last year.

• What a deal is worth. Your average contract value by segment. A keyword that attracts enterprise buyers can be worth more than one with ten times the volume that attracts small teams you can't serve profitably.

• How buyers decide. Your typical sales cycle and the questions that come up at each stage.

Done when: you can name the two or three segments where better search visibility would create the most pipeline, and why.

Step 2: Mine your buyers' own words

Your best keywords are already in your company. They're just not in a keyword tool yet. Collect the exact language buyers use from:

• Sales call notes and recordings: how prospects describe the problem in the first call.

• Your CRM: closed-won and closed-lost reasons, and which competitors were in the deal.

• Support tickets and onboarding questions: what new customers were trying to do.

• Review sites: what reviewers praise and complain about on your G2 or Capterra pages, and your competitors'.

• Communities: Reddit, Slack groups and forums where your buyers ask each other for recommendations.

Write phrases down verbatim. "Our month-end close takes two weeks" is a better seed than "accounting automation", because it's how a buyer actually thinks.

Done when: you have 50 or more real phrases, each tagged with where it came from.

Step 3: Expand with tools, and with your own data

Now turn those phrases into keywords. Run your best seeds through a keyword tool such as Semrush, Ahrefs or Google Keyword Planner to find variations, questions and related terms. Note search volume, keyword difficulty and cost per click for each.

Don't skip your own data. In Google Search Console, the Performance report shows queries you already get impressions for but rank poorly on. These are often the fastest wins, because Google already associates you with them.

A note on volume: keyword tools often underestimate niche B2B terms, and show "0" for searches that do happen. A specific, high-intent phrase with tiny reported volume can still be worth a page.

Done when: you have a long list, often a few hundred keywords, with volume, difficulty and CPC for each.

Step 4: Add the AI layer: prompts, not just keywords

Buyers now also ask AI assistants for shortlists, and they ask differently. A search might be "payroll software for startups". A prompt might be "What's the best payroll software for a 40-person startup with employees in the US and Canada?"

For each segment from Step 1, write 20 to 50 realistic buying prompts. Use the language from Step 2: the problem, the company size, the tools they already use and the constraints they mention. Then run them through ChatGPT, Perplexity, Gemini and Google's AI Overviews, and record which brands get named and which pages get cited.

These prompts become your baseline for Answer Share, the share of high-intent AI answers and search results in your category that cite or rank your brand. They also reveal topics no keyword tool shows, because prompts rarely match a keyword exactly. Our AI SEO service is built on this layer.

Done when: you have a prompt list per segment, and a record of who gets cited for each.

Step 5: Classify every keyword by type and funnel stage

SaaS keywords fall into a small number of types. Labeling them tells you what page each one needs and how close it is to revenue:

1. Problem: "how to reduce churn". Early stage; best answered with guides.

2. Category: "customer success software". Evaluating; needs a strong category or product page.

3. "Best" and listicle: "best customer success software". Evaluating; often won by review sites, so aim to be listed in the pages that rank, as well as ranking yourself.

4. Comparison: "[your brand] vs [competitor]". Deciding; needs a fair comparison page.

5. Alternatives: "[competitor] alternatives". Deciding; some of the highest-intent searches in SaaS.

6. Integration: "[product] Salesforce integration". Deciding; strong signal of fit, and often very low competition.

7. Use case and industry: "customer success software for fintech". Evaluating; needs dedicated pages.

8. Pricing: "[brand] pricing" or "[category] pricing". Deciding; buyers and AI assistants both look for it.

9. Branded: your own name. Protect it, but don't count it as SEO growth.

Done when: every keyword has a type and a funnel stage: early, evaluating or deciding.

Step 6: Score keywords by pipeline value

Sorting by volume rewards the wrong keywords. Instead, give each keyword a simple pipeline score:

• Intent (1 to 5): how close the searcher is to buying. Problem keywords score low; comparison, alternatives, integration and pricing score high.

• Fit (1 to 3): how well the searcher matches the segments you named in Step 1.

• Winnability (1 to 3): whether you can realistically rank or get cited, based on difficulty, what currently ranks, and whether you have something genuinely better to say.

Pipeline score = Intent × Fit × Winnability, from 1 to 45. Use volume as a tiebreaker, not the main ranking. A high cost per click is a useful hint, too: it means advertisers find those searches commercially valuable.

An illustrative example, for a hypothetical customer-success software company (scores are made up for the example):

• "[competitor] alternatives": intent 5 × fit 3 × winnability 2 = 30

• "customer success software for fintech": intent 4 × fit 3 × winnability 2 = 24

• "what is customer success": intent 1 × fit 2 × winnability 1 = 2

The definition keyword may have far more volume. It's still the last of the three to build.

Done when: your list is sorted by pipeline score, and the top 20 are keywords your sales team would recognize as buying signals.

Step 7: Map every keyword to one page

The final step turns a spreadsheet into a site plan:

• Group synonyms. Keywords that mean the same thing ("customer success platform" and "customer success software") belong to one page, not two.

• Give each page one primary keyword. Pages that target the same keyword compete with each other and confuse both Google and AI models about which page is the answer.

• Pick the page type. Comparison keywords need comparison pages, integration keywords need integration pages, and problem keywords need guides. Match the format that already ranks.

• Decide: existing page or new page. If a page already exists for a keyword, improve it before creating another.

Keep it all in one keyword map, with columns for keyword, type, stage, pipeline score, target URL, status and owner. That map becomes your content and technical roadmap. Our SaaS SEO strategy service starts from exactly this document.

Done when: every top-scoring keyword has exactly one target URL, and no URL has two primary keywords.

Keep it alive every month

Keyword research isn't a one-off project. Once a month:

• Check Search Console for new queries you're getting impressions for.

• Add new phrases from the month's sales calls and lost-deal reasons.

• Re-run your AI prompts and update your Answer Share baseline.

• Re-score anything that changed, and move effort to what's creating pipeline.

If you're putting this into practice, our SaaS SEO checklist covers what comes next: earning the rankings, getting cited, and measuring the pipeline.

Common mistakes to avoid

• Chasing volume. High-volume definition keywords rarely create demos for B2B SaaS.

• Skipping comparison and alternatives pages because they feel awkward. If you don't publish them, a review site or competitor answers for you.

• Counting branded growth as SEO success. Brand searches grow when other marketing works.

• Ignoring zero-volume keywords that match exactly what your buyers say on sales calls.

• Researching once. Buyer language and AI answers change; your keyword map should too.

Frequently asked questions

How many keywords does a SaaS company need?

Fewer than you think. A focused list of the 50 to 100 highest-scoring keywords, each mapped to one page, will usually create more pipeline than a list of thousands. Start with the top 20 and build outward.

Are zero-volume keywords worth targeting?

Often, yes, for B2B SaaS. Keyword tools estimate volume from limited data and often underreport niche, high-intent searches. If a phrase comes straight from your sales calls and signals buying intent, it can be worth a page even when a tool shows little or no volume.

How is keyword research different for AI search?

AI assistants answer conversational prompts, not short keywords, and they name a few sources rather than listing ten links. Research the prompts your buyers ask, record which brands get cited, and make sure your pages answer those questions clearly and consistently.

How often should you redo SaaS keyword research?

Do a full pass once a year or when you launch a new product or segment, and a light monthly update from Search Console, sales calls and your AI prompt baseline.

Want a second opinion on your keyword map? Book a strategy call. We'll look at which searches and AI answers should be creating pipeline for you, and which pages should win them.