Keyword Research for SaaS and New Products

Why keyword research is different for a new product

Most keyword research advice assumes a site that already gets traffic. A new SaaS starts with none, and with a disadvantage that established products do not have: nobody types its name into Google yet.

That changes where to look. The searches that matter are the ones people make before they know your product exists: the problem they are trying to solve, the cost they are trying to estimate, the alternatives they are trying to compare. Those searches are also where a new site has the best chance, because the most specific of them are often answered badly or not at all.

Google’s starter guide puts the starting point simply: “Think about the words that a user might search for to find a piece of your content.” The research is about finding those words and deciding which ones are worth a page.

Step 1: Start from the problem, not the product

Write down, without any tool, the questions a buyer asks on the way to a product like yours. They tend to fall into five groups:

Your support inbox, sales calls and community threads are the best source for this list. They contain the questions in the words buyers actually use, which are often different from the words on your website.

Step 2: Check demand with a keyword tool

A keyword tool turns the list into numbers. Three are worth knowing, and two are free.

Google Keyword Planner is free inside a Google Ads account. Its “Avg. monthly searches” column is “the average number of times people have searched for a keyword and its close variants,” averaged by default over twelve months. Google notes that the figures are rounded. Its Competition column is easy to misread: it shows “the number of advertisers that showed on each keyword relative to all keywords across Google.” High competition there means many advertisers, not that ranking organically is hard.

Google Search Console is free once your site has impressions. Its Performance report shows the queries for which your pages already appeared, with impressions, clicks and average position. For a new site it is the best source of keywords you did not think of, because Google has already decided your page is relevant to them.

Paid tools such as Ahrefs add organic difficulty, the pages that currently rank and their links, and the related questions people search for. They are most useful once you know which subject you are researching.

Step 3: Read difficulty for what it measures

Keyword difficulty scores look precise, but each tool calculates them differently. Ahrefs explains its own: it pulls “the top 10 ranking pages for your keyword,” looks up how many websites link to each of them, and turns that into a score from 0 to 100.

Ahrefs is candid about the limits. The score “doesn’t account for your specific website’s authority.” It can “hide specific weaknesses in the search results,” because it averages ten pages. And it does not say whether the pages that rank actually satisfy the searcher. Its conclusion is that no metric can accurately predict your chances of ranking.

For a new site, the useful reading is simple: a low score means the current top pages do not have many links, so a good page without many links has a chance. A high score means the opposite, and those queries are better left for later.

Step 4: Look at the results page yourself

No metric replaces searching the query and looking at what comes back. Three questions decide whether a keyword is worth a page:

• What kind of page ranks? If the results are comparison articles, a product page will struggle. If they are product pages, an article will.
• Are the top results good? Old, thin or off-topic pages are an opening, whatever the difficulty score says.
• Is there an AI answer? If Google shows an AI Overview, the page you write should open with an answer clear enough to be quoted. Our guide to what AEO is explains why.

Step 5: Prioritize and map one query to each page

With demand, difficulty and the results page in front of you, sort the list. For a new product, the order that usually works is:

1. Specific questions with low difficulty and clear relevance to your product. They rank first and bring the right readers.
2. Cost and comparison questions. They bring buyers close to a decision.
3. Broader category terms. They take longer, and they get easier once the specific pages have built some authority.

Then map each chosen query to one page. A query that needs its own answer gets its own article. Close variants go on the same page. Related articles are grouped under a guide that links to all of them, which helps readers and search engines see how the subject fits together. Our guide to landing page SEO shows how those articles support the product page.

Step 6: Keep the subject fixed

Keyword tools are good at finding phrasing and terrible at deciding what your site should be about. If you follow volume alone, you end up writing about whatever is popular, which Google warns against: it lists “producing lots of content on many different topics in hopes that some of it might perform well” among the signs of search-engine-first content.

Decide the subject first, from what your product does and what you know. Then use the data to choose the exact words, the order of publication and the questions to cover. A query with modest volume and perfect relevance is worth more than a popular one your product has nothing to do with.

A worked example: how we chose the articles in this guide

This blog is part of the UpRank experiment, which tests whether useful content and answer engine optimization can bring a pay-to-rank board the attention that hype once did. When we added SEO articles to our guide to organic visibility, the subject was fixed in advance: what a small product with no budget needs to know to be found. We then used Ahrefs to choose the questions within that subject.

These are the figures we worked from, as shown in Ahrefs on October 12, 2026:

Three decisions came out of the numbers:

• Close variants share a page. “seo for startups” and “startup seo” ask the same question, so one article answers both. The same goes for landing page and product page SEO.
• Low difficulty first. Every query in the table has a difficulty under 10, which means the pages ranking today do not have many sites linking to them. A new blog has a chance there and almost none on broad terms.
• Relevance over volume. “keyword research for saas” has the smallest volume in the table, but it is exactly the question a founder asks at this stage, and the article could use real data that nobody else has. That made it worth writing.

We will check these queries in Search Console after launch and add the results, including the ones that do not work, to the experiment log.