How to Research Keywords for Free: A Step-by-Step Workflow

Most "keyword research" guides are really tool roundups wearing a how-to headline. This one is not a list of tools - it is the actual sequence of steps for doing keyword research without paying for a tool, in the order that avoids wasted work. Skip the order and you end up polishing Google Trends charts for a topic Search Console already tells you is dead, or guessing at search volume before you've even written down what your own customers actually say. Follow it in order and every step feeds the next one with real data instead of a guess.

The actual free workflow for finding keywords step by step looks like this: write down real customer language first, expand it with Google's own autocomplete and People Also Ask, check what you already rank for in Google Search Console if the site is live, get a search-volume range from Google Keyword Planner, check momentum in Google Trends, run the shortlist through a free keyword tool for a difficulty score, cluster the survivors and map them to pages, then validate intent against the live search results page before a single word gets written. Eight steps, each free, each in an order that actually matters.

Step 1: Write down your seed topics in the words your customers actually use

Before touching any tool, write down 10-20 seed phrases using the exact words customers, support tickets, sales calls and reviews already use - not the internal jargon your team uses to describe the product. A SaaS company that calls its own feature a "workflow automation engine" internally might find customers searching "how to stop doing this by hand," which is a completely different keyword. This step costs nothing but attention, and it matters because every later step - autocomplete, Search Console, Keyword Planner - only expands on the seeds you hand it. Garbage seeds produce a polished list of the wrong keywords.

Step 2: Mine Google's autocomplete and People Also Ask for real long-tail phrasing

Type each seed phrase into Google's search box and note every autocomplete suggestion, then work through the alphabet after it ("keyword research a," "keyword research b," and so on) to surface long-tail variations you would not have guessed. Scroll to the People Also Ask box on the results page and expand every question - each click reveals more related questions, effectively giving a free, crowdsourced list of real sub-questions people ask around that topic. This step matters more than it looks: according to Backlinko's analysis of 306 million keywords, 91.8% of all search queries are long-tail, even though those long-tail terms collectively account for only 3.3% of total search volume. In other words, almost everything people actually type is a longer, more specific phrase than the seed keyword you started with - autocomplete and People Also Ask are how you find those phrases for free.

Step 3: Pull the queries you already rank for from Google Search Console

Can Google Search Console be used for keyword research? Yes, and if the site is already live it should be the first data source used, before Keyword Planner or Trends, because it shows demand that is already converting into clicks rather than demand you're guessing at. Open Performance > Search results, and look at the Queries tab. Per Google's own Search Console documentation, the report tracks clicks, impressions, click-through rate and average position for every query that surfaced your pages, and Google explicitly recommends using this data to "learn which queries are bringing traffic to your site." Filter for queries where your page ranks in positions 5-15 with reasonable impressions - those are keywords you're already being found for but not winning yet, which is a far cheaper opportunity than starting from zero on a brand-new keyword. A brand-new site has no Search Console history, so skip straight to Step 4.

Step 4: Get a real, if bucketed, search volume from Google Keyword Planner without running ads

Google Ads Keyword Planner is free and does not require running ads - during account setup, choose the option to set up an account without a campaign, then open Tools > Keyword Planner and use "Discover new keywords." What most free-keyword-research guides leave out is the actual limitation: an analysis of 60 million keywords by Authoritas found that accounts with no active ad spend see only around seven broad volume buckets (10, 100, 1K-10K, 10K-100K, and so on), while accounts running even a small active campaign unlock roughly 60 much narrower buckets and closer-to-exact monthly figures. A keyword showing "1K-10K" could be getting 1,200 searches a month or 9,800 - too wide a range to prioritize between two similar keywords with confidence. Use the free, bucketed numbers to eliminate keywords with genuinely low demand (the 0-10 and 10-100 buckets), not to rank-order keywords that both land in the same wide bucket.

Step 5: Check momentum and seasonality in Google Trends

What does Google Trends actually show that helps keyword research? Not absolute search volume - Trends only shows relative interest over time, scaled 0-100 against the peak in the date range selected. That is exactly what makes it useful as a check on the keywords that survived Step 4: it helps you see whether interest in a keyword is flat, seasonal, rising or fading, which a single volume number can't show. Enter up to four terms in the "Compare" box to see them side by side, and per Google's own Trends documentation, a related search will show a percentage increase, or the label "Breakout" when growth has exceeded 5,000% - a strong early signal for an emerging topic before it shows up with real volume anywhere else. Trends won't show a search-volume number at all, but it will show direction, and direction matters for deciding whether to prioritize a keyword now or file it for later.

Step 6: Run your shortlist through AI Rank Lab's free Keyword Planner for a difficulty score

Is AI Rank Lab's Keyword Planner actually free? Yes, with real limits worth knowing before you rely on it: the free tier gives up to 10 keyword searches per day with the two metrics that matter most at this stage of the workflow, search volume and keyword difficulty. Everything up to this point in the workflow - autocomplete, Search Console, Keyword Planner, Trends - tells you what people search for and roughly how much. None of it tells you how hard a given keyword is to actually rank for. That's the gap this step closes: feed your shortlist from Steps 1-5 into AI Rank Lab's Keyword Planner a handful at a time within the daily limit, and use the difficulty score to cut anything that's technically searched but unrealistic to rank for with your current site authority. Competitor-gap analysis, the AI Opportunity Score and historical trend data sit behind the paid tiers - the free tier is built for exactly this validation step, not for running your entire keyword universe through it in one sitting.

Step 7: Organize the survivors into clusters and map them to pages

How do you organize keywords into clusters and map them to pages? Group every keyword that survived Steps 1-6 by shared search intent and topic, not by exact wording - "best project management software" and "top project management tools" are different phrases chasing the same page. Each cluster becomes one page, not one keyword per page; trying to write a separate thin page for every close variant is what search engines treat as low-value, duplicated content. Put the cluster's primary keyword in the page title and first paragraph, and let the long-tail variants from Step 2 show up naturally in subheadings and body copy, where they'll get picked up without being forced. A spreadsheet with three columns - cluster name, target page, and the keyword list underneath it - is enough; you don't need dedicated software to do this step.

Step 8: Check search intent against the live SERP before targeting a keyword

How do you check search intent before targeting a keyword? Before writing anything, search the exact keyword yourself and look at what's actually ranking. If the top results are all product pages and yours would be a blog post, or vice versa, that mismatch means the keyword's intent doesn't match the content you're planning, no matter how good the volume and difficulty numbers looked. This is the step most free-keyword-research guides skip entirely - they treat "search intent" as a classification exercise (informational, commercial, transactional, navigational) rather than something you verify by looking at the actual page one results, which is the only source that reflects what Google's ranking systems currently believe the query wants. It costs nothing but two minutes per keyword, and it catches mismatches that volume and difficulty data alone cannot.

The real limits of doing keyword research for free

What are the real limits of free keyword research tools? Every source in this workflow trades one of three things for being free: precision (Keyword Planner's bucketed ranges instead of exact volume), scale (AI Rank Lab's and most free tools' daily search caps), or scope (Search Console only shows queries you already rank for, Trends only shows relative interest, not absolute demand). None of that makes the free workflow inferior for the actual decision you're making at this stage, which is choosing what to write about next - it just means the free path takes longer per keyword and caps how many keywords you can fully validate in a day. Semrush's own Keyword Magic Tool landing page is a useful comparison point here: it promises a database of "more than 26 billion keywords" and describes a six-step research process, but it starts at "expand your list with a keyword tool," skipping the seed-collection and intent-validation steps that come before and after it in the workflow above, and it discloses only that free users get "a limited number" of suggestions without stating what that number actually is. Paid tools remove the daily caps and widen the volume precision; they do not remove the need to do Steps 1, 7 and 8 by hand, because no tool, free or paid, can read your customers' support tickets or judge search intent against the live results page for you.

None of this changes because the destination is a traditional blue link or an AI-generated answer. Google's own guidance on generative AI features in Search states that these features are rooted in the same core ranking and quality systems as regular Search results, retrieving from the same index. A keyword workflow built on real customer language, validated demand and confirmed intent is the same workflow whether the page ends up as a top-10 organic result or a citation inside an AI-generated overview.