SaaS keyword research works best when you start from the problems your buyers are trying to solve and work toward your product, not the other way around. Build a seed list from your own product facts, sort it by what the searcher wants to do next, expand it with Google's free tools, and check it against the queries your site already appears for. Then give each keyword group one page with one job.
This guide walks through that process step by step. The framework is the author's recommendation, not an industry standard. Every step uses data you can pull yourself, mostly from free Google tools.
Why SaaS keyword research is different
Most keyword advice assumes people search for the thing they want to buy. With software, some buyers may not know your product category exists. They may search for the pain ("why do invoices get paid late"), the manual workaround ("spreadsheet template for tracking trials"), or a job title plus a task. Check whether the category term your marketing team uses is one your buyers actually type.
Three consequences follow:
- Your category name is only one part of the list. If you only research terms that describe your product, you miss earlier-stage searches from people who would buy once they understood the solution.
- Specific B2B terms may not appear in keyword tools. Google's Keyword Planner states that keywords with very low search volumes, or keywords it considers sensitive, aren't discoverable or forecastable (Google Ads Help). A niche term missing from the tool doesn't prove nobody searches for it. Treat the tool as one signal.
- One keyword often maps to one buying situation. A comparison search, a how-to search and an integration search come from different people at different moments. They need different pages.
Step 1: Write a seed list from facts you already know
Before opening any tool, write down what you know for certain about the business. This keeps research tied to what you can truthfully offer, which also matters later when articles are written.
Collect these inputs:
- Problems you solve, in the words customers use in sales calls, support tickets and onboarding emails.
- Jobs and workflows the product supports, each phrased as a task ("reconcile payments," "onboard a new client").
- Roles that use or approve the product (finance lead, operations manager, agency owner).
- Tools your customers already use, including the ones you integrate with and the ones you replace.
- Manual alternatives such as spreadsheets, email threads or templates.
- Constraints that decide deals: team size, industry, compliance needs, pricing model.
Aim for a raw list you can sort, not a polished one. Customer wording is usually more useful than your own marketing copy, because search queries are typed in the searcher's vocabulary.
Hypothetical example: a SaaS product that automates invoice reminders for freelancers might list "late client payments," "invoice follow-up email," "spreadsheet to track unpaid invoices," "invoice reminder software," and "accounting tool integration." The product category term appears once; the rest are problems and workarounds.
Step 2: Sort keywords by buyer situation
Group the seed list by what the searcher is trying to do. The table below is the author's proposed working structure, not a fixed taxonomy. Adjust the buckets to your market.
| Bucket | What the searcher wants | Typical page | Hypothetical example |
|---|---|---|---|
| Problem | Understand or fix a pain | Educational article | "how to chase late invoices politely" |
| Workaround | Do it manually, often with a template | How-to guide or template page | "unpaid invoice tracker spreadsheet" |
| Category | Find software that does the job | Product or solution page | "invoice reminder software" |
| Comparison | Choose between named options | Comparison or alternatives page | "[tool A] vs [tool B]" |
| Integration and use case | Confirm it fits their stack, role or industry | Integration or use-case page | "invoice reminders for [accounting tool]" |
Label each keyword by bucket in a spreadsheet column. Then use this recommended framework for the page type: for problem and workaround keywords, plan a page that teaches first and connects to the product honestly. For category, comparison and integration keywords, plan a page that answers practical questions: what the product does, who it suits, how it connects, how it is priced. Test the framework against your own analytics. If visitors from problem-stage pages never reach your product pages, change the page's next step rather than assuming the bucket is wrong.
Be careful with comparison pages. Only make claims about another product that you can verify from that vendor's own current documentation, and avoid blanket statements that your product is better.
Step 3: Expand the list with Keyword Planner
Google's Keyword Planner has two entry points that map well to this process (Google Ads Help):
- Discover new keywords lets you search for ideas from a keyword, a URL, or both. Google notes that specifying a keyword and a URL can produce a larger volume of ideas than a URL alone.
- Get search volume and forecasts lets you upload keywords you already have and see volume and forecast data for them.
A practical way to use both:
- Run each bucket's strongest seed terms through Discover new keywords, with and without your own URL, and note ideas you hadn't considered.
- Paste your sorted list into Get search volume and forecasts to compare roughly how much demand each term shows.
- Keep terms that Planner doesn't show if customers genuinely use them. Mark them "no tool data" rather than deleting them. Google says very low-volume and sensitive keywords aren't discoverable or forecastable, so a blank result is not a verdict.
If Planner returns few ideas for your problem-stage or workaround buckets, fill the gaps with vocabulary from your sales and support conversations.
Step 4: Check what your site already appears for
If your site has been live for a while, Google Search Console shows real queries before you write anything. The Performance report lets you view clicks and impressions, and group them by dimensions including queries, pages and countries (Search Console Help).
Search Console keeps data for the last 16 months (Analytics Help), so you can look back across months rather than just last week.
Three reads are particularly useful for a SaaS site:
- Queries with impressions but few clicks. If your site appears for them but gets few clicks, check whether the page that appears answers the query. Either improve it or write a dedicated page.
- The Pages view for a single query. If several of your pages appear for the same query, you may have overlapping targets. Decide which page should own it.
- Queries you never planned for. These can reveal the vocabulary buyers actually use, which can feed back into Step 1.
Search Console's Insights report groups similar queries together for sites with a significant amount of query data, which helps you analyze search trends. Google also warns that its query classification is powered by AI and may mislabel branded and non-branded queries (Search Console Help). If branded searches for your product name dominate your data, review the split by hand before drawing conclusions about non-branded demand.
A brand-new site won't have this data yet. In that case, rely on the seed list, buckets and Keyword Planner, then come back once the site has gathered impressions.
Step 5: Prioritize with a simple fit score
You will end up with more keywords than you can write for. This is the author's recommended scoring approach; rate each keyword group high, medium or low on four questions:
- Product fit. Could a reader who finds this page plausibly become a customer, or at least understand why your product exists? A keyword with large volume and no connection to what you sell isn't a good target.
- Intent clarity. Can you tell what the searcher wants? If you can't decide what page would satisfy them, the keyword is a poor target until you can.
- Realistic competitiveness. Look at the actual results page. If it's dominated by large, established publishers and software vendors covering the topic exhaustively, a newer site may do better with a narrower variant first.
- Your ability to be accurate. Can you describe the topic using facts you can confirm, such as features, pricing and integrations? If not, skip it until you can.
Start with groups that score high on product fit and accuracy and at least medium on the others. Narrow searches close to your product, such as a particular integration or a specific role's task, are worth testing as first targets. Check the live results page for your own market before committing.
Step 6: Give each keyword group one page and one job
Keyword research only helps if it ends in a decision about pages. For each group, record:
- The primary keyword and a few close variants that a single page can answer.
- The bucket (problem, workaround, category, comparison, integration or use case).
- The target page: an existing URL to improve, or a new page to write.
- The next step for the reader: which page or action the article should lead to, such as a trial signup, a related guide or an integration page.
- The facts the page depends on: pricing, plan limits, supported integrations, and any statistic you plan to cite. Confirm these before drafting.
If two keywords need different pages, separate them. If a reader would be equally well served by one page, combine them. Avoid creating a near-duplicate page for every phrase variation.
Common mistakes in SaaS keyword research
- Researching only your own category name. This leaves out the problem-stage searchers who may become customers later.
- Ignoring branded data in Search Console. Branded queries can mask how little non-branded traffic a site actually gets.
- Treating a blank in the tool as zero demand. Keyword Planner says very low-volume and sensitive keywords aren't discoverable or forecastable (Google Ads Help), so check those terms against Search Console and customer conversations.
- Writing for keywords you can't support with facts. An article about an integration you don't support, or a pricing claim that's out of date, costs trust.
- Researching once and stopping. Products, competitors and buyer vocabulary change. Revisit the list when you launch a feature or notice new queries in Search Console.
Where Pentra fits
Pentra is autopilot SEO software that takes over the research-to-publishing work once you have your inputs. It reads your website and drafts a business profile that you confirm, researches keywords people search for, and writes articles from those confirmed facts. Each article goes through an independent fact check before it's published to WordPress or a GitHub-based site. You choose a pace from 1 to 21 articles a week, or approve every article yourself. Pentra also improves published pages that sit close to page one and reports clicks and positions from your Google Search Console data.
It doesn't replace your judgment on the seed list, the buckets or the fit score: the confirmed business profile is where your knowledge of buyers and product facts goes in. A free plan includes 3 articles a month, with no credit card required. Try Pentra if you want keyword research, drafting and publishing to run from that confirmed profile.
A first-week plan
- Day 1: Write the seed list from sales calls, support tickets and your product's feature list.
- Day 2: Sort into buckets and add the manual workarounds buyers use today.
- Day 3: Run the strongest terms through Keyword Planner and note gaps where it shows no data.
- Day 4: Export Search Console queries and pages, remove branded queries by hand, and flag queries with impressions but few clicks.
- Day 5: Score the groups, choose a first set of pages, and record the facts each one depends on.
At the end you will have a keyword-to-page plan built from your own buyers' language, your own site data and verifiable product facts.