Someone asked me last week why their blog post — 2,400 words, well-researched, decent images — had collected eleven visits in two months. I opened their keyword sheet. Every single target had a difficulty score above 45. They'd written good content aimed directly at HubSpot, NerdWallet and a handful of legacy forums that have held those positions since before ChatGPT existed.
That's the whole problem in one screenshot. Not writing quality. Not domain age. They were aiming at keywords that were never winnable in the first place.
Finding low competition keywords isn't about hunting for a magic metric. It's about reading the search results page like a scout reads terrain — and being honest about which hills you can actually take. I've built three niche sites to five figures in monthly sessions using almost nothing but free tools and a spreadsheet. I've also wasted four months on a keyword I was sure I could rank for. That one still stings.
Here's what actually works, including the parts nobody puts in the tool tutorials.
Key Takeaways
- "Low competition" is not a single number — it's a judgment about who is currently ranking and whether you can realistically beat them.
- Keyword difficulty scores are estimates, not verdicts. A KD of 12 can still be a wall if the top results are exact-match pages.
- The fastest signal of a soft keyword is a SERP filled with forums, thin affiliate pages, or content that hasn't been updated in years.
- Free tools — Google Keyword Planner, Google Trends, and the search operators built into Google itself — cover 80% of what paid tools do for this specific job.
- Volume matters far less than intent. A 200-search term that converts beats a 5,000-search term that nobody clicks a result for.
- The best low competition keywords are usually hiding in the questions your own audience already asks you.
What "low competition" actually means in 2026
Pull up any keyword tool and you'll get a number between 0 and 100. It feels objective. It isn't.
That number is a prediction, calculated from the backlink profiles of the pages currently ranking. It says nothing about whether those pages are any good, whether they match search intent, or whether Google has simply run out of better options. Which happens more often than you'd think.
The thresholds I actually use
After enough trial and error, I settled on numbers that keep me out of trouble. These are mine, not gospel — adjust to your site's authority.
- Keyword difficulty under 15 for a brand-new site with no meaningful backlinks
- Under 30 once you have a few dozen referring domains and at least one page ranking in the top five
- Domain authority of every ranking competitor below 25, with at least two of them under 15
- Total results under roughly 50,000 when you search the exact phrase in quotes
- And the one I weigh heaviest: at least two of the top ten results are not professionally optimized pages — a forum thread, a Reddit post, a Quora answer, a product page pretending to be an article
That last criterion is the one that changed everything for me. A keyword tool might score a term at 18 because the ranking pages have weak links. But if all ten results are purpose-built, intent-matched articles from sites that clearly know what they're doing, you're not walking into an opening. You're walking into a wall with a friendly paint job.
Why paid tools mislead you on this specific task
I pay for one keyword tool. I use it for volume estimates and competitor tracking, not for deciding what to write.
The reason is simple: difficulty scores were designed to be comparable across millions of keywords, which means they flatten out exactly the details that matter for a small site. A page ranking on the strength of one viral Reddit thread and a page ranking on a genuinely authoritative domain might score identically. They are not the same problem.
So treat the score as a filter, not a decision. Anything above your ceiling, discard. Anything below, put it in a shortlist and go look at the actual search results with your own eyes.
The manual method: five minutes per keyword, zero dollars
This is the part the tool-first crowd skips, and it's where the real advantage lives. Google hands you competitive intelligence for free. You just have to ask in the right syntax.
Step one: search your phrase in quotes
Put the exact phrase in quotation marks. Google returns only pages containing that precise string. The result count at the top is your first signal.
If a phrase in quotes returns under 10,000 results and the top page is a forum thread, congratulations — you've found something worth a closer look. If it returns 2 million results and the first page is dominated by established publishers, move on. Don't negotiate with yourself here. I've done it, and the negotiation always ends badly.
Step two: use allintitle: to find real competition
Type allintitle: followed by your phrase. This returns only pages that use the phrase in their title tag — a far better proxy for "someone specifically targeted this" than the general result count.
A keyword with 400,000 total results but only 30 allintitle matches is dramatically softer than it looks. That gap between total results and allintitle results is one of the most useful signals I've found, and it costs nothing.
Step three: read the search results like a human, not a spreadsheet
Now the part that requires actual judgment. Open the top ten results and ask three questions about each:
- Does this page actually answer the query, or is it a broader article that mentions the topic in passing?
- When was it last updated — and does the content show its age? (Broken screenshots, dead links, references to products that no longer exist.)
- Is this a page built to rank, or a page that happens to rank? Product pages, category pages and forum threads are all opportunities.
I once found a keyword where positions two through five were all Amazon product listings and position one was a manufacturer's PDF spec sheet. No actual article existed. I wrote one, 1,100 words, and it was ranking on page one within six weeks. That's not typical — but it's the kind of opening you only spot by looking.
Step four: check search intent and the AI overview
Two things worth checking before you commit.
First, intent. If the top results are all transactional — product pages, pricing pages — and you're planning an informational article, you're pushing against the grain. Google has already decided what that query wants.
Second, and this is the 2026 addition: does the query trigger an AI-generated overview at the top of the results? For simple factual questions — "what is X," "how tall is Y" — the overview often answers the question completely, and clicks to organic results drop sharply. I stopped targeting definitional queries almost entirely for this reason. The traffic is still theoretically there. It just doesn't reach you.
The free tools that actually earn their place
You can do this work without spending anything. Here's what I keep open, and what each one is genuinely good for.
| Tool | What it's actually good for | Where it falls short |
|---|---|---|
| Google Keyword Planner | Broad volume ranges; discovering related terms you hadn't considered | Volume is bucketed, not precise; no real difficulty signal |
| Google Trends | Spotting whether interest is growing, flat, or dying off | Relative data only — never gives you absolute search numbers |
| Google autocomplete & "People also search for" | Real phrasing your audience uses, in their own words | No volume data at all |
| Search operators (quotes, allintitle:, site:) | The most honest competition read you'll get | Manual, slow, doesn't scale past a few dozen keywords |
| Paid suites (Ahrefs, Semrush) | Bulk filtering, backlink analysis, tracking | Difficulty scores flatten nuance; subscription cost |
Notice what's missing: a single "find me low competition keywords" button. It doesn't exist, because the judgment is the work.
Using Google Trends without fooling yourself
Trends gives you relative interest, normalized to 100. A term sitting at 8 that has held steady for three years is a far safer bet than a term that spiked to 100 last month and is now collapsing. Seasonal keywords can be excellent — you just need to publish two to three months ahead of the peak, not during it. I learned that the hard way with a gift-related term I published in December. By the time it ranked, the wave was gone.
Finding low competition keywords for YouTube
Video is a different game, and the metrics that matter barely overlap with written content.
On YouTube, the competition signal isn't backlinks — it's the gap between what a video promises and what it delivers. Type your phrase into the search bar and look at the top results. If the first five videos are all over ten minutes long and none of them answer the question in the first thirty seconds, you've found an opening. YouTube rewards watch time and satisfaction, and a tight four-minute video that answers the query immediately will outrank a padded twelve-minute one surprisingly often.
Two more checks worth running:
- Look at the view counts relative to subscriber counts. A video with 40,000 views from a channel with 300,000 subscribers underperformed. One with 40,000 views from a channel with 2,000 subscribers is a signal that the topic is genuinely in demand.
- Check the upload dates. If the top results are three or four years old and the topic has moved on, there's room.
The same phrase-in-quotes trick applies inside YouTube search, though it's less reliable than on Google. What works better is simply watching the first minute of the top three videos and asking whether you'd do it better. If the answer is an easy yes, that's your keyword.
Where to find candidates nobody else is looking at
Every keyword tool draws from the same pool, which means every competitor is fishing in the same water. The advantage comes from sources tools don't index well.
Reddit is the single most underused source I know. Not the big subreddits — the small, specific ones. Search a niche subreddit for phrases like "how do I," "anyone know," or "struggling with." Sort by top posts of the year. Those threads are full of questions phrased exactly the way real people type them, and they frequently have zero well-optimized content competing for them. A thread with 400 upvotes and 90 comments, and not a single article answering it properly, is a keyword with a built-in audience.
Your own inbox works too. The questions you answer repeatedly by email are questions other people are searching. I built an entire content cluster from eleven recurring support questions. Nothing clever about it. Just paying attention.
The mistakes I kept making, so you don't have to
Chasing volume over winnability. My first niche site targeted ten keywords between 5,000 and 20,000 monthly searches. Six months later, three of those pages had never ranked at all. The keywords I eventually succeeded with were all under 800 searches. Slower individually, but they compounded — and each one made the next easier to rank.
Trusting the difficulty score alone. Already covered, but it bears repeating: a low score with strong intent-matched competitors is a trap. Look at the pages.
Ignoring the exact-match problem. If the top three results are pages whose titles mirror the query word for word, and those pages are decent, the query is more contested than any tool will tell you. Exact-match targeting is a strong signal that someone built a page specifically for that term.
Writing before checking intent. I once spent a full weekend on a 3,000-word guide for a query where every ranking result was a comparison table. Google wanted a table. I gave it an essay. It ranked on page four and stayed there.
Putting it together: a workflow you can run this afternoon
Start with twenty candidate phrases. Run each through the quotes and allintitle: checks. Discard anything above your difficulty ceiling or with a dense exact-match SERP. For the survivors, open the top ten results and score each on whether it genuinely serves the query.
You'll end up with three to six keywords worth writing about. That's plenty. A focused shortlist beats a spreadsheet of two hundred maybes every time.
And when you find one where the competition is a two-year-old forum post and a product page — write that article first. Not the one that feels most impressive. The one you can actually win.
The uncomfortable truth is that most people searching for low competition keywords are really searching for a shortcut around judgment. There isn't one. The tools will hand you a shortlist. Deciding which names on it you can beat is the skill, and it only sharpens by looking at search results until you can read them the way you read a room.
