Most keyword research starts in a tool and ends with a spreadsheet of popular phrases nobody on the team can connect to revenue. For SaaS it works better the other way round: start from what customers say, then check whether people search for it.
The result is a shorter list, and every line on it has a reason to be there.
Short on time? Comppound does this work for SaaS companies. AI-native, and a person reviews every page.
Have us do itCollect the language first
Your best keyword source is the words buyers use before they know yours. Those live in sales call recordings, demo request forms, support tickets, churn surveys, review sites and community threads. Copy phrases verbatim. “Billable hours” and “time tracking” may mean the same thing to you and very different things to a searcher.
- Sales and demo calls: objections, competitors named, “can it do X” questions
- Support tickets: how people describe the job they hired the product for
- Review sites: what reviewers praise and complain about in your category
- Communities: the questions people ask peers before asking a vendor
Then check the demand
Take each phrase into a keyword tool or Google Search Console. You are checking three things: whether anyone searches it, which wording is most common, and what other searches sit around it. Search volume figures from tools are estimates, so use them to compare options, not to forecast traffic.
Low or zero reported volume is not a reason to drop a decision-stage search. “Tallyroom vs Timewell” may show no volume and still be searched by every serious prospect.
Comppound can do this for you
Want your SEO foundations handled?
Comppound finds the searches your buyers make, fixes what stops Google reading your site and builds the pages that win them. Our system is AI-native, and a person reviews every page before it ships.
- Buyer search map from your sales calls
- Technical fixes on the pages that matter
- Pages published to your CMS
Score each search on three things
A simple score keeps the list honest. Give each search a mark from one to three on:
- Business value: would someone searching this plausibly buy from you?
- Ability to win: are the pages ranking today beatable by a better answer?
- Effort: can you write a strong page with what you know today?
Group searches into pages
Several searches often belong on one page. If Google shows mostly the same results for two searches, write one page for both. If the results differ, they need separate pages. This is keyword clustering, and it prevents two of your own pages competing for the same search.
Worked example · an illustration, not a client result
Scoring Tallyroom's first ten searches
Tallyroom scored its list and wrote pages for the top four first. Two high-volume problem searches went to the bottom because the ranking pages were strong and the buyers were early.
| Search | Value | Can win | Effort | Decision |
|---|---|---|---|---|
| timewell alternatives for agencies | 3 | 3 | 2 | Write now |
| time tracking with xero invoicing | 3 | 2 | 3 | Write now |
| agency timesheet template | 2 | 2 | 3 | Next month |
| what is time tracking | 1 | 1 | 3 | Skip |
Do it yourself
Make your first scored keyword list
- 1Collect thirty phrases from calls, tickets and reviews, verbatim.
- 2Check each in a keyword tool or Search Console and note the most common wording.
- 3Score each on value, ability to win and effort, one to three.
- 4Search the top ten and merge any that show the same results.
- 5Pick the five highest totals as this month's pages.
Free tool for this: Keyword traffic potential calculator
Common mistakes
Sorting by volume
The biggest numbers are usually the broadest and least likely to convert. Sort by your score instead.
Trusting tool numbers as forecasts
Volume and difficulty are estimates from clickstream samples. Treat them as relative, and check results pages yourself.
One page per keyword
Writing separate pages for near-identical searches splits your signals. Merge searches that share results.
Rather not do it yourself?
Want your SEO foundations handled?
Comppound finds the searches your buyers make, fixes what stops Google reading your site and builds the pages that win them. Our system is AI-native, and a person reviews every page before it ships.
- Buyer search map from your sales calls
- Technical fixes on the pages that matter
- Pages published to your CMS
Questions
Any of the main ones will do for checking demand. Google Search Console is free and shows the searches you already appear for, which is often the best list of all.
One main search and the close variations that show the same results. If a page tries to answer unrelated searches it usually answers none of them well.
Keep them when they are late-stage and specific to your product or competitors. Tools often miss these searches entirely.
Key terms
- Keyword ResearchKeyword research is the work of finding the searches your buyers make, judging what each is worth, and deciding which ones to build pages for.
- Search VolumeSearch volume is the estimated number of times a keyword is searched per month. Tools model it, so figures vary and small B2B searches are often under-counted.
- Keyword DifficultyKeyword difficulty is a tool's estimate of how hard it would be to reach the first page of Google for a search. Each tool calculates it differently.
- Keyword ClusteringKeyword clustering groups searches one page can serve, usually by comparing how much their Google results overlap, so each page targets a full cluster.
- Zero-Volume KeywordsZero-volume keywords show no search volume in keyword tools, yet often get real searches. In B2B SaaS many valuable, specific queries fall into this group.
- Keyword CannibalizationKeyword cannibalization is when several pages on one site target the same search intent, so they split signals and compete with each other in results.
- Seed KeywordA seed keyword is a short starting phrase, usually one to three words, that you expand into a full keyword list using tools, autocomplete and customer language.
