The winnable set is small on purpose. A new or small site cannot outrank established ones on broad terms, so it looks for specific phrases that bigger sites ignore and page-one results that are visibly weak. Finding those phrases is easy. Proving they are winnable is the work.
Low-competition keywords are specific search terms with low difficulty scores that are easier for new or small websites to rank for. Rank Math's guide to finding them without paid tools builds its whole method around one warning: the difficulty score is where the search starts, not where it ends.
What Makes a Keyword Low-Competition
Four traits recur in the guides.
- Long-tail phrasing. A longer, more specific phrase, usually three words or more, aimed at one precise topic or question. Rank Math's contrast is start a blog, where large, established blogs hold page one, against how to start a blog as a plumber, where a focused page can stand out.
- A low difficulty score. Generally under 30 to 35 on the scales Semrush and Ahrefs use. How keyword difficulty scores are built explains why two tools can disagree about the same phrase.
- Manageable volume. Often somewhere between 10 and 1,000 searches a month, the band big brands skip because it is too small to move their numbers. Rank Math pairs Google Keyword Planner's own low-competition label with a filter of 100 to 1,000 searches for the same reason.
- Weak incumbents. Page one is held by smaller blogs, forums, or question-and-answer threads rather than authority domains.
Use the first three as a filter. The fourth decides whether the filter found anything real.
Why the Score Is Not Proof
A tool that labels a keyword easy has estimated competition from its own data: links, domain strength, and ranking history, weighted by its own formula. It has not read the pages that rank. Rank Math's guide is blunt about the gap: difficulty scores miss the full picture, and the quality of the content already ranking is what decides.
That is why the same score can hide two very different pages. A KD of 20 might mean page one is a set of thin, dated posts nobody has updated. It might also mean a single thorough guide from a respected site holds the top spot with few links, which a score reads as weak and a reader does not.
The Page-One Audit
Open the term in a private browser window so your own history does not shape the results, and read the top five to ten. You are looking for signs that a focused page could take a spot:
- newer or weaker domains holding positions;
- few quality links pointing at the top pages;
- forums, question-and-answer threads, niche blogs, or dated pages ranking;
- few results with the exact phrase in the title or address;
- thin, vague, or off-topic content;
- pages that miss the intent behind the query, answering a different question than the one asked;
- a mixed page one with no dominant winner.
Two examples set the bar. The start-a-blog lineup, with page one full of large, established sites, is the fail case. A phrase like best home workout for beginners with no equipment, surfacing in a fitness forum thread, is the pass shape: specific, question-like, and answerable by one focused page.
When the audit disagrees with the score, the audit wins. Budget ten minutes a keyword; a shortlist of twenty is one careful afternoon, which is cheap next to a month of writing for a page that could never rank.
From Seed to Shortlist
Rank Math's no-paid-tools sequence runs in five steps.
- Start with a seed. The broad topic your buyers already ask about, such as home workouts or budget travel. Ask what customers keep asking, what problem they are solving, and what they type just before they need you.
- Audit the live results for each idea with the checklist above.
- Expand on Google's own surfaces. Autocomplete, the People Also Ask box, and the related searches at the foot of the page each add longer, question-shaped variations.
- Listen where buyers talk. Reddit, Quora, niche boards, and social groups. A question with heavy engagement and thin coverage on Google is an opening.
- Validate with free data. Search Console for queries your site already shows for with many impressions and few clicks; Google Trends for rising interest; Planner's low-competition label with a volume band for longer candidates. Even a Planner low label still needs the page-one check when big brands might be behind it.
Then group what survives by intent and topic, so near-duplicate phrases do not end up competing with each other on separate pages, and prioritize the overlap: low competition, real demand, clear intent. That overlap is the winnable set.
What the Winnable Set Looks Like on a Real Site
The situation that sends most people looking is familiar. A site owner on Reddit's SEO forum describes sixty-plus articles live, daily visits near zero, and a niche that feels crowded, and asks for low-competition topics that still carry volume. Another site owner reports ranking six months in on niche choice alone, without a network of blogs or a paid tool stack, using free tools where needed and fixing only their own pages.
Both stories point the same way. The first site wrote before it checked; the second checked before it wrote. Find the phrase, audit its page one, and only then build the page.
Date Every Pick
A shortlist changes as page one changes. For each keyword you keep, record:
- the phrase, with its difficulty and volume and the tool and month they came from;
- the top-ten lineup on the day you audited it, with your weakness notes;
- the format page one rewards, such as a guide, a list, or a product page;
- the date.
qequals returns the live Google results and suggestions for a query, so an agent can re-run the audit on the same shortlist next month and flag the phrases whose page one got stronger.
A dated weakness note is what turns a finder's list into something you can brief.
