Most people looking for “low competition keywords” are really looking for one thing: a query they can rank for without waiting eight months and buying links. That’s a reasonable goal. It’s also achievable, but not by the method most tutorials teach, which is “sort your keyword tool by difficulty score, pick anything under 20, write a post.”
That method fails constantly. I’ve watched sites publish forty posts targeting KD-under-15 keywords and rank for none of them, because difficulty scores measure the wrong thing. What follows is the process I actually use — manual SERP reading first, tools second, and a validation step before a single word gets written.
A note on the timeline in the title: a week is realistic for a genuinely uncontested query on a site that already has some crawl history and a handful of links. It is not realistic for a brand-new domain on a query with three DR-70 sites in the top five. I’ll be specific about which is which as we go.
Table of Contents
What “low competition” actually means
Competition is not a property of a keyword. It’s a property of the results currently ranking for that keyword, relative to what you can produce.
That distinction matters more than anything else in this article. “Best CRM for small business” and “why does my HubSpot workflow keep skipping enrolled contacts” can have identical search volume in your tool. One is defended by seven companies with content teams. The other is defended by a four-year-old forum thread and a support doc that doesn’t answer the question. Same volume. Wildly different competition.
So when you evaluate a keyword, you are not asking “is this keyword hard?” You are asking three concrete questions:
- Can I make something better than what’s on page one right now? Better meaning more directly useful to the person searching, not longer.
- Do the current results actually deserve their positions, or are they ranking on domain strength while answering the wrong question?
- Can my site plausibly be trusted on this topic? Google’s tolerance for an unknown domain varies enormously by subject. A tools comparison post from a new site is fine. A dosage guide is not.
If the answer to all three is yes, the keyword is low competition for you. That last part is not a technicality — a keyword can be low competition for a site with existing topical coverage and high competition for a site starting from nothing on the same term.
Why difficulty scores mislead
Every keyword difficulty score you’ve seen — Ahrefs KD, Semrush KD%, Moz, Ubersuggest, the lot — is derived primarily from the backlink profiles of the currently ranking pages. Ahrefs is open about this; KD is essentially a function of how many referring domains the top-ranking pages have.
That produces three predictable failure modes:
False hard. A query where the top results are pages from large sites that happen to have thousands of sitewide links, but whose specific ranking page has no links and barely addresses the query. The tool sees domain authority and reports KD 45. You look at the page and see a thin category listing. That’s a winnable SERP wearing a scary number.
False easy. A query with KD 3 where every top result is a well-known brand’s dedicated resource page, updated regularly, with a real author. Nobody links to those pages because nobody links to product documentation — but you are not displacing them with a blog post. Low KD, unwinnable.
Intent blindness. The score has no idea whether the SERP is informational, transactional, or navigational. A KD 8 keyword whose entire page one is product pages will never rank your article, regardless of quality.
Difficulty scores are a filter, not a verdict. Use them to reduce a list of 800 keywords to 60 worth looking at. Then look at them.
Reading a SERP manually
This is the skill. Everything else in keyword research is bookkeeping.
Open the query in an incognito window, ideally with location set to your target market — for US-focused work, use &gl=us on the search URL or set the location in Search Console’s preview. Then work down the page.
First, count the ad and feature real estate. If the top of the page is four ads, a shopping carousel, a map pack, and a People Also Ask box before the first organic result, ranking #1 organic may mean very little traffic. This is the most commonly skipped check and it kills more “successful” rankings than any algorithm update.
Then classify every result in the top ten. For each one, note:
- Is it a dedicated page for this exact query, or a broader page that happens to cover it?
- Is it a forum thread, a Q&A site, a Reddit post, or a video?
- Is it a homepage or category page?
- When was it meaningfully updated? (Not the “updated” date on the page — check whether the actual content reflects anything recent.)
Two of those observations matter most. Dedicated pages beating broad pages is normal and means the SERP is settled. Broad pages ranking, or forum threads ranking, means Google could not find a good dedicated answer. That second case is your opening.
If three or more of the top ten are Reddit threads, Quora answers, or a generic listicle from a site that clearly has no expertise in the subject, the query is underserved. That’s the real signal — not the number in your tool.
Then check whether the results answer the question. Search the actual query, then read the top three results the way the searcher would. If you find yourself thinking “this doesn’t tell me what I asked,” neither does it tell the searcher. Someone will eventually publish the page that does. It may as well be you.
Finally, check the outliers. Is there anything in the top ten from a small or unknown site? One small site ranking on page one is the single most reliable indicator that the SERP is accessible. It’s direct evidence that Google will rank a domain like yours here. If the top ten is uniformly major brands with no exceptions, treat the keyword as long-term, not as a quick win.
A quick scoring pass
I keep a rough tally per keyword rather than a formal score:
Signal |
Read it as |
|---|---|
Forum/Q&A results in top 10 |
Opening — no strong dedicated page exists |
A small/unknown site ranking |
Strong opening — Google will rank a site like yours |
Top results don’t answer the query |
Opening — quality gap you can fill |
All results are dedicated, current, thorough pages |
Closed — pick a different angle |
SERP is product/category pages, you’d write an article |
Wrong format — skip or change format |
Heavy ad and feature stacking above organic |
Low value even if you win |
Three or more openings and no blockers means write it.
The KGR method
The Keyword Golden Ratio is a filter developed by Doug Cunnington. It’s a numeric shortcut for the manual SERP read above, and it works well specifically for very-low-volume, very-specific queries.
The formula:
KGR = (number of Google results with the keyword in the title) ÷ (monthly search volume) Applied only to keywords with monthly search volume under 250.
A KGR below 0.25 indicates the keyword is underserved — meaningfully fewer pages are explicitly targeting it than there is demand for it. Cunnington’s original guidance is that these should rank in the top 50 fairly quickly after indexing, and often top 10 with reasonable content.
How to run it:
- Take a keyword with volume under 250/month. This ceiling matters — the ratio is meaningless above it.
- Search Google using the
allintitle:operator:allintitle:how to clean a burr grinder without water - Note the result count Google reports.
- Divide that count by the monthly search volume.
If volume is 120 and allintitle returns 22 results, KGR is 0.18. That passes.
Reading the result:
- Under 0.25 — underserved, worth targeting
- 0.25 to 1.00 — competitive but sometimes workable if the SERP read is favorable
- Above 1.00 — more pages targeting it than monthly searches; skip
What KGR actually measures. It’s a supply-and-demand proxy. allintitle counts pages that deliberately optimized for the phrase. A low count means few competitors are trying. That’s genuinely useful information that no difficulty score captures.
Where it breaks. Two places, and you should know both:
Google’s allintitle result counts are estimates and have gotten noisier over time. Run the same query twice and you may get different numbers. Treat the figure as an order of magnitude, not a measurement. If it says 18 and you want to know if it’s really 18 or really 30, it doesn’t matter — both pass.
And KGR says nothing about intent or quality. A keyword can pass KGR beautifully and still be a bad target because the SERP is transactional, or because the searcher wants a tool and you’re writing an essay. KGR is one gate. The SERP read is the other. Run both.
Volume floor. Don’t chase keywords under about 20 searches a month as standalone posts. They’re real but the return is too small to justify a full page. Fold them into a larger piece as a subheading instead — you’ll rank for them anyway.
For the full history of the method, the original testing behind the ratio, and the edge cases, see our complete guide to the Keyword Golden Ratio.
Finding candidate keywords
You need raw material before you can filter it. Here’s the order I actually work in.
Free sources first
Google Search Console. If your site has any history at all, start here — it is the only source that tells you what Google already associates with your domain. Open Performance, set the date range to the last three months, and filter for queries where your average position is between 8 and 25. Those are pages Google is already willing to rank; they just aren’t ranking well. A query sitting at position 14 with impressions and near-zero clicks is usually the fastest win available to you, faster than any new keyword, because the trust problem is already solved.
Also filter for queries with impressions but where you have no dedicated page. Google is guessing which of your pages to show. Give it the right one.
Google autocomplete and People Also Ask. Type your seed term and step through the alphabet after it. Then click open every PAA box — they expand as you click, and you can pull thirty question-format keywords out of one search. These are phrased the way people actually search, which matters more for AI search visibility than it used to.
Reddit, forums, and Facebook groups in your niche. Not for volume data — for phrasing. The exact wording of a frustrated question in a subreddit is very often a low-competition long-tail query, because nobody writes content in that register. Search Reddit for your topic plus “how do I” or “why does.”
Google Trends. Use it for the related queries panel, not the graph. Set to your country, look at “Rising” — those are queries with growing demand and, usually, not much content yet.
Bing Webmaster Tools. Free, and its keyword research tool gives volume estimates without a subscription. The volumes are Bing’s, so scale them mentally, but the relative ordering is useful.
Paid tools where they earn their cost
Ahrefs. The most useful feature for this specific job is not Keywords Explorer — it’s Content Gap. Put in three competitors your size, and it shows what they rank for and you don’t. Filter that list by KD and volume, then manually check the SERPs. Also useful: the “Traffic potential” column, which estimates traffic for the whole topic rather than the single phrase, which is closer to reality.
Semrush. Keyword Magic Tool with the “Questions” filter and volume capped at 250 gets you a KGR candidate list quickly. The Position Tracking tool is worth more than the research tools once you’ve published.
Keywords Everywhere. A browser extension, credit-based rather than subscription, that shows volume and competition inline on Google, YouTube, and several other sites as you browse. For manual SERP work this is the single most convenient paid tool — the data appears where you’re already looking.
Ubersuggest. Cheaper than the above; data quality is noticeably below Ahrefs and Semrush. Reasonable if budget is the constraint, as long as you treat the numbers as directional and lean harder on manual SERP reading.
Google Keyword Planner. Free with an active Ads account, but it groups similar keywords together and gives wide volume ranges unless you’re spending. Useful for confirming that a keyword has some volume; not useful for precision.
Whatever combination you use, the tool’s job is to produce a candidate list. The decision is still made by looking at the SERP.
Validating intent before you commit
This is where most keyword lists fall apart, and it takes about ninety seconds per keyword.
The question is not “what does this keyword mean?” It’s “what does Google currently believe the searcher wants?” Those differ, and Google’s belief is the one that determines what ranks.
Search the keyword and classify page one:
- Informational — guides, explainers, how-tos. Write a guide.
- Commercial investigation — “best X,” comparisons, reviews. Write a comparison, with real opinions and a recommendation. A neutral overview will not rank here.
- Transactional — product and pricing pages dominate. Point a service or product page at it, not a blog post.
- Navigational — someone wants a specific brand. Skip unless it’s your brand.
The mismatch that costs the most time is writing a blog post for a transactional SERP. It never works, no matter how good the post is, because the format is wrong for the intent. If page one is nine product pages and you publish article number ten, you have produced the only result on the page that doesn’t do what the searcher came to do.
Two subtler checks worth doing:
Format consistency. If eight of ten results are listicles, the searcher wants options. If eight of ten are single-solution tutorials, they want one answer. Match the dominant format even if you think yours is better — you can improve the format after you rank.
Freshness sensitivity. Look at publish dates. If everything on page one is from the last few months, this is a query where Google strongly prefers recent content, which means it needs maintenance after publishing. If page one has results from three or four years ago sitting comfortably, you can write it once and leave it. Budget accordingly — a portfolio of freshness-sensitive keywords is a recurring obligation, not a one-time build.
Here’s the failure mode that catches people who do everything above correctly: they build a list of thirty unrelated low-competition keywords, publish thirty disconnected posts, get a trickle of traffic to each, and plateau. The site never gets stronger. Every new post is as hard to rank as the first one.
Low-competition keywords are worth the most when they compound. That happens when they cluster.
Pick a topic, not a keyword. Instead of thirty scattered wins, choose an area you want to be known for and find the twenty low-competition queries inside it. Each individual post is just as easy to rank. But now each one reinforces the others — Google accumulates evidence that your site covers this subject, and the tenth post in a cluster ranks noticeably faster than the first, because the domain has demonstrated relevance in that space.
Structure the cluster. One substantial page covering the topic broadly, then supporting pages each answering one specific question, each linking back to the main page and to its siblings where relevant. This isn’t a formatting exercise — internal links are how you tell Google these pages belong together, and how link equity from any one of them reaches the rest.
Use the easy keywords as the entry point to the hard one. The head term in your topic is probably too competitive today. It won’t be after you own fifteen supporting queries around it. The long tail isn’t a consolation prize; it’s the route.
Watch for the shift. You’ll notice it in Search Console — new pages in an established cluster start picking up impressions within days rather than weeks, and you begin ranking for terms you never explicitly targeted. That’s topical authority becoming real, and it’s the point at which the whole approach starts paying compound returns rather than linear ones.
If you want the full framework for planning clusters and mapping supporting content, see our semantic SEO approach and how topical authority is built.
The realistic timeline
Because the promise in the title deserves a straight answer.
A week is achievable when: the query passes KGR, page one contains forum threads or off-topic results, your site is already indexed and crawled regularly, and the keyword sits inside a topic you’ve published on before. In that combination, indexing to top-10 in days is normal.
Two to eight weeks is the common case: a clean SERP read, a new page on a site with modest history, no existing cluster. It’ll sit on page three or four first and climb as it accumulates engagement signals and internal links.
Three to six months or longer: any query where established, genuinely good dedicated pages hold the top spots, or where your domain has no track record in the subject.
Anyone promising the first case as the general rule is selling something. The process above is what makes the first case happen more often.
Putting it together
The working sequence, condensed:
- Pull candidates from Search Console first (position 8–25 queries), then autocomplete, PAA, forums, and your keyword tool.
- Cut anything with volume over 250 out of your KGR pass; run
allintitle:on the rest and keep KGR under 0.25. - Read every surviving SERP manually. Look for forum results, small sites ranking, and answers that miss the question.
- Classify intent and format from page one. Match it.
- Check whether the query needs ongoing freshness.
- Group what’s left into topic clusters rather than a flat list.
- Publish the supporting pages first, interlink them, and let the head term become reachable.
That’s the whole method. It’s slower per keyword than sorting by KD, and it produces a list roughly a tenth the size — which is the point, because the list is real.
FAQ
What is a low competition keyword? A low competition keyword is one where the pages currently ranking can be outperformed by content you’re able to produce. Competition depends on the existing search results and on your site’s standing in that topic, not on the keyword itself. The same keyword can be low competition for one site and unwinnable for another.
Is keyword difficulty score reliable? Only as a rough filter. Keyword difficulty scores in Ahrefs, Semrush, Moz, and similar tools are calculated mainly from the backlink profiles of ranking pages. They don’t account for search intent, content quality, or format mismatch. Use them to shorten a long list, then judge each keyword by reading its search results directly.
What is the Keyword Golden Ratio? The Keyword Golden Ratio (KGR) is the number of Google results containing the keyword in the page title, divided by monthly search volume, applied only to keywords under 250 searches per month. A result below 0.25 indicates the keyword is underserved and can usually be ranked for quickly. You find the title count using the allintitle: search operator.
How do I check KGR? Search Google for allintitle: followed by your exact keyword, note the number of results Google reports, and divide it by the keyword’s monthly search volume. If volume is 150 and allintitle shows 30 results, KGR is 0.20, which passes. Only apply the formula to keywords under 250 monthly searches.
Can you really rank a new website in a week? Sometimes, on genuinely uncontested long-tail queries. A week is realistic when the keyword passes KGR, the current results include forum posts or pages that don’t answer the question, and your site is already being crawled. For most keywords on a new domain, two to eight weeks is more typical, and competitive terms take months.
How do I find low competition keywords for free? Google Search Console is the strongest free source — filter for queries where you already rank between positions 8 and 25. Then use Google autocomplete, People Also Ask boxes, Google Trends related queries, Bing Webmaster Tools, and question phrasing pulled from Reddit and niche forums. Validate every candidate by reading its search results manually.
How much search volume is too low? Below roughly 20 searches per month, a keyword usually isn’t worth a dedicated page. Cover those queries as a section inside a larger article instead — you’ll typically rank for them anyway, and the larger page earns its position more easily.
Why do my low competition keywords still not rank? The most common causes are intent mismatch (publishing an article for a SERP full of product pages), format mismatch (a single-answer tutorial where searchers want a list), no topical support around the page, or a difficulty score that concealed strong, well-maintained competitors. Re-read the search results for the keyword and check which of those applies.
Internal links used:
- our complete guide to the Keyword Golden Ratio
- our semantic SEO approach
- how topical authority is built
Additional suggested internal links to place contextually: