SaaS keyword research for startups,

SaaS keyword research for a startup starts with the problems your buyers type, not with a keyword tool. Check each term in Google autocomplete and the live results, give each term one page, and test the same questions in ChatGPT and Perplexity.

You don’t need a paid tool to do this well. Free tools give you estimates, and at seed stage your sales calls and support inbox are better data than any volume column.

In short

• Write down the problems, jobs and comparisons buyers bring to sales calls. Those are your seed terms.
• Expand them with free tools: Google autocomplete, the live results, Search Console, Google Trends and Keyword Planner.
• Treat every volume number as an estimate. Google calls Keyword Planner numbers estimates, and Trends shows relative interest, not counts.
• Read the live results for each term to see what kind of page wins and whether you can compete.
• Sort terms by stage: problem, solution, comparison and alternative.
• Map one term to one page, then turn the same terms into AI prompts and test them weekly.

Start from buyer problems, not seed keywords

A keyword tool only knows words people already type in volume. A new SaaS category often has words nobody has put into a tool yet. So start from your own evidence:

• Sales calls and demos. What did the buyer say just before “that is why we are looking”? Write it down in their words.
• Support tickets and onboarding. The “how do I” questions your users ask are questions other people search for.
• Lost deals. Which tools did they compare you with? Each one is a comparison or alternative term.
• Your product’s jobs. For each feature, the job it does in plain words, such as “send invoices from Stripe” rather than “billing automation”.

Aim for a list of problems and jobs, not a list of single words. Our SEO for SaaS startups guide has a short section on picking queries with buyer intent. This post goes deeper on the research itself.

Expand the list with free tools

Google autocomplete. Type each seed term into Google and note the suggestions. Google says autocomplete predictions “come from real searches”, but also that autocomplete “should not be taken as always reflecting the most popular search terms related to a topic” (Google). Use it for wording, not for ranking terms by size.

The live results. The results page itself suggests related questions and searches. Note the questions that match your buyer’s problem.

Search Console. Once your site has any impressions, the Performance report shows “what search queries are most likely to show your site” (Google). Sort by impressions and look for queries where you show up but have no page made for that question. These are often the best terms you have, because Google already connects them to you.

Google Trends. Trends compares interest over time and between places. Google says the data is normalized and “scaled on a range of 0 to 100”, and that search terms with low volume appear as zero (Google). That makes it useful for picking between two phrasings of the same idea, or spotting a seasonal pattern, and much less useful for niche B2B terms, where a zero does not mean nobody searches.

Keyword Planner. Google’s ads tool gives “estimates of the searches” a keyword receives (Google Ads). The number covers a keyword “and its close variants” over the months and places you choose (Google Ads), so two similar phrasings can share one number. Use it to compare terms against each other, not to forecast traffic.

Read the live results before you pick a term

Search each term in a private window and look at the first page. Three questions matter:

1. What type of page wins? Product pages, comparison posts, how-to guides or forum threads. If every result is a how-to guide, a feature page is unlikely to rank there.
2. Who ranks? If the top results are from your direct competitors and large review sites, note it. If they are from small blogs or old pages, that is a gap.
3. How fresh and complete are the pages? Old pages, pages that don’t answer the question, or forum threads in the top results usually mean the question is not well served yet.

This is the intent check. A term with the right buyer and the wrong page type for your site is a term to skip, or to answer with a different kind of page.

Sort terms by stage

Put every term into one of four groups. Each group needs a different kind of page.

Comparison and alternative terms are the closest to a purchase, which is why people call them bottom of funnel. They are also where you need to be fair to the other tools, or readers will not trust the page.

Map one term to one page

Give every term or close group of terms exactly one page. If two pages target the same question, they compete with each other and neither is clearly the answer. Google’s guidance on AI search makes a related point: AI systems understand synonyms, so “you don’t have to worry that you don’t have enough ‘long-tail’ keywords or haven’t captured every variation” (Google). One strong page per question beats five near copies.

Before you write, preview the title and description you plan to use with our SERP preview tool, so the main term is in the part that shows.

For the page types themselves, see our guides to SaaS comparison pages and programmatic SEO for SaaS. Programmatic pages only make sense when many terms share one shape and you have real data for each.

Google’s helpful content guidance asks whether content is “made primarily to gain search engine rankings” (Google). Keyword research tells you what to write about. It should not decide that a page exists only to catch a phrase.

Turn terms into AI prompts and test them

People ask ChatGPT and Perplexity longer questions than they type into Google. Turn each problem, comparison and alternative term into the question a buyer would ask, for example “what is the best dunning tool for a small SaaS on Stripe?”.

Keep a fixed set of these prompts and run them on the same day each week. Note which tools get named and which pages get cited. Our guide to tracking AI visibility covers how to set up and score the test.

Keep the list in one sheet and re-check it

One sheet is enough. Use these columns: term, stage, buyer problem it maps to, page type that wins, target URL, status, and the matching AI prompt. Re-check it monthly against Search Console, because new queries show up there before they show up anywhere else.

FAQ

How do you do keyword research for a SaaS with no budget?

Start from sales calls, support tickets and lost deals. Expand with Google autocomplete, the live results, Search Console, Google Trends and Keyword Planner, all free. Then check the live results for each term before you commit a page to it.

Should a new SaaS target high-volume keywords?

Usually not first. Broad terms tend to be held by large sites, and the searchers are often not buyers yet. Comparison, alternative and specific problem terms are smaller but much closer to a purchase.

What are bottom-of-funnel keywords for SaaS?

Terms people use when they are close to choosing a tool, such as “tool A vs tool B”, “tool A alternatives” or “best tool for a specific job”. They need comparison and alternatives pages.

How is keyword research different for AI search?

The questions are longer and written as full sentences. Use the same buyer problems, phrase them as questions, and test them in ChatGPT and Perplexity every week instead of looking for a volume number.

How often should I update my keyword list?

Monthly is enough for most startups. Check Search Console for new queries, re-run the AI prompt set, and add terms from recent sales calls.

Want the research and the pages done for you?

Ignura automates SEO and GEO, and you approve what goes out. Keyword research and the pages that follow are part of our content work. See the pricing page for how we price it.

Book your free Growth Check