Open a keyword tool, set the difficulty filter to "easy," sort by volume, and you get a list that looks like a goldmine. Two hundred terms, thousands of searches each, competition scores so low they seem broken.
I did exactly that about four years ago. Picked twelve keywords, wrote twelve articles, waited. Six months later, three of them ranked. The other nine sat on page four while a forum thread from 2019 and a Reddit post with fourteen upvotes outranked a 2,000-word article I'd rewritten twice.
That was the moment I stopped trusting the difficulty score as a standalone answer. Learning how to find low competition keywords isn't a filter problem. It's a judgment problem—and the tools only give you half the inputs.
Key Takeaways
- A "low difficulty" score means nothing without checking who actually ranks on page one.
- Forum threads and Reddit posts ranking well are often a better signal than any numeric score.
- Search volume between 100 and 1,000 per month is where most realistic wins live for a new site.
- Free methods (Google autocomplete, Keyword Planner, Trends) get you 80% of the way.
- AI-generated answers in search results are thinning out informational clicks—buyer-intent terms hold up better.
- You need around 20 to 30 candidate keywords before you find 5 worth writing.
What "low competition" actually means in 2026
Most people define low competition as a number under 20 in Ahrefs or under 30 in Semrush. That's a proxy, not a definition. The real question is simpler: can a page like yours realistically outrank what's already there?
Here's what I check before writing a single word. I search the keyword in an incognito window, no personalization. Then I look at who's on page one and ask three questions.
- Are there forums, Quora threads, or Reddit posts in the top five? That usually means nobody has written a proper answer yet.
- Are the ranking pages thin—under 800 words, no images, no real structure?
- Do the domains look like household names, or are they small sites similar to mine?
If two of those three lean your way, it's a candidate. If it's Forbes, a government site, and three e-commerce giants, the difficulty score can say 8 and I'll still walk away.
Why the metric lies
Difficulty scores are calculated from backlink profiles. But plenty of pages rank with almost no links because the query is specific enough that Google has nothing better to show. A keyword tool sees "low competition" and assumes you'll waltz in. Reality: you're competing with a 400-word blog post that's been sitting there for six years and Google trusts.
A mistake I made early on was treating that number as permission. It isn't. It's a hint that you should go look with your own eyes.
The free manual method that works without paid tools
You don't need Ahrefs to find low competition keywords. You need about forty minutes and a browser.
Step 1: mine autocomplete and related searches
Type your seed topic into Google and let the dropdown do the work. Every suggestion is a real query typed by real people. Then scroll to the bottom of the results page and collect the related searches—those are the queries Google already considers semantically linked.
Do this for your main keyword plus four variations. I usually end up with 30 to 50 raw phrases in a spreadsheet inside twenty minutes.
Step 2: filter in Google Keyword Planner
Keyword Planner is free with a Google Ads account, and you don't need to run a campaign to use it. Paste your list, set the location, and look at two columns: average monthly searches and competition.
Competition here is about ad bidding, not organic, so it's an imperfect proxy. Still useful. A phrase with 100 to 1,000 searches and low ad competition is worth investigating further, especially if it's a longer phrase—four words or more.
Step 3: sanity-check with Google Trends
Trends won't give you volumes, but it tells you direction. A keyword that's flat for five years is safe. A keyword that spiked nine months ago and is now collapsing is a trap—you'll write, rank, and watch the traffic evaporate.
I check the five-year view, not the twelve-month one. Long flat lines are my favorite thing to see.
Free vs. paid: what each tool actually gives you
Here's a straight comparison based on what I've used across three sites.
| Tool | Cost | Best signal it gives you | Where it falls short |
|---|---|---|---|
| Google Keyword Planner | Free | Real search volume ranges | Ad-focused competition metric |
| Google Trends | Free | Direction over time | No absolute numbers |
| Ahrefs | Paid | Keyword Difficulty + SERP overview | Score ignores content quality |
| Semrush | Paid | Intent breakdown, PAA-style clusters | Volume estimates run high for niche terms |
| Manual SERP inspection | Free | Ground truth | Slow, doesn't scale past ~30 terms |
Paid tools save time, not thinking. If you're starting out with no budget, Keyword Planner plus Trends plus a careful SERP check will get you to the same shortlist—just slower.
Finding low competition keywords on YouTube and Reddit
The YouTube angle
YouTube search behaves differently from Google. Video results compete with fewer pages, and small channels regularly rank on terms that would be impossible on the web. The method is the same: type your topic into the YouTube search bar, note the autocomplete, then look at the top results.
What you want to see: videos with under 20,000 views ranking for a phrase that gets searched. That's a soft signal the topic isn't saturated on that platform. I built a small channel around three such phrases and got consistent views within two months—nothing spectacular, but it proved the pattern.
The Reddit angle
Reddit posts ranking on Google are one of the best low-competition signals there is. When a forum thread outranks every dedicated article on a topic, it means nobody has written the definitive piece yet.
Search site:reddit.com [your topic] in Google. Look for threads with high engagement and a clear question in the title. That question is often your article title, almost verbatim.
- A thread titled "how do I X without Y" is a ready-made H1.
- Comments with the most upvotes reveal the sub-questions to answer.
- If the thread is over two years old and still ranking, the demand is stable.
How AI-generated answers change the game
This is the part most guides haven't caught up with. When someone searches a simple factual question, they increasingly get an answer without clicking anything. That kills a whole category of low competition keywords—the ones that were easy precisely because they were shallow.
What survives? Terms where the reader needs your specific experience, opinion, or process. "What is X" is dying. "How I did X and what broke" is not.
So when I'm screening candidates now, I ask one extra question: could this be answered in three sentences? If yes, I skip it. Not because it won't rank, but because ranking won't matter.
Building a shortlist you can actually execute
Gather 20 to 30 candidates. Then run each through this filter, in order:
- Search it in incognito. Note the weakest result on page one.
- Check Trends over five years. Reject anything declining.
- Confirm volume is 100+ per month in Keyword Planner.
- Ask whether the answer needs experience or just definition.
From 25 candidates, I usually keep 4 or 5. That ratio sounds brutal, and it is—but writing five articles that rank beats writing twenty that don't. I learned that the expensive way.
The last thing worth saying: this whole process gets faster with reps. The first time I did it, screening ten keywords took an afternoon. Now it's twenty minutes. Not because I found a better tool, but because my eye for "who's actually ranking here" got sharper.
That judgment is the actual skill. The tools just speed it up.
