How to Find Low-Competition Keywords
You find low-competition keywords by looking for searches where page one is weak: forums, thin posts, and stores no bigger than yours. Start in Search Console and Google autocomplete, shortlist with a difficulty score under 30 if you have a tool, and confirm every pick by reading the page one you would have to beat.
Short answer: Harvest real searches from Search Console and Google autocomplete, keep the ones with buyer words attached, and only write the ones where page one is beatable by a store your size.
What You Need
- A Google account with Search Console verified for your store
- Google itself, in a private window so your history does not shape the suggestions
- A spreadsheet with three columns: phrase, where you found it, and a yes or no
- A keyword difficulty tool if you already pay for one (optional; Google and Search Console do most of the work)
Step 1: Start With the Keywords Search Console Already Shows You
You start with the keywords Search Console already shows you, since Google has already decided your site belongs on those results and only needs a reason to move you up. Most SEO plans start with a list of phrases the site has never ranked for. I start with the ones it already does.
So open the Performance report, which shows clicks, impressions, click-through rate, and average position for every query your site appeared for. Set the date range to the last three months, then sort by impressions and filter position to between 8 and 20.
Read position from the table, not the chart, since the chart averages your whole site and the table shows the number for each query. Those are phrases where Google already shows you on page one’s bottom half or page two, with real demand behind them, and the competition has already been measured: you are in it.
Group the Queries by Subject
Grouping the queries by subject turns a list of 200 phrases into five or six fights you can actually pick. Fixing 20 sunscreen queries as one batch moves a whole topic, while fixing 20 random queries moves nothing. That is how sunscreen got picked for Exclusive Beauty Club: strong volume, weak competition, and the site already had a foothold at position 22. The cluster went from 207 clicks to more than 6,000 in a year.
Also, this report is the fastest place to find quick wins in Search Console, since a page sitting at position 9 with a better title and two internal links usually moves faster than anything new ranks at all.
Step 2: Harvest Candidates From Google Autocomplete
You harvest low-competition keyword candidates from Google autocomplete: every suggestion is a search real people have run, in their own words, for free. Google’s own explanation of autocomplete says it looks at “the real searches that happen on Google” and shows “common and trending ones.”
So a suggestion is proof of demand before you have spent a dollar on a tool. Most people start keyword research by opening a paid tool. I start by opening Google.
From there, work the pattern deliberately. Type your product plus “for” and stop. “Best sunscreen for” handed me face, sensitive skin, kids, babies, body, oily acne-prone skin, and tattoos, and each one became a post.
The same move worked for Fieldsheer: typing “heated ___ for women” handed over vests, gloves, socks, jackets, and slippers, and that list became the whole content plan behind a #1 ranking in two weeks.
Then run the alphabet. Type your product plus each letter, a through z, and write down every suggestion. Open the People Also Ask box on the results page and click it open until the questions repeat. An hour of this produces a bigger list than most stores can write in a quarter.
Mine Your Own Customer Emails
Mining your own customer emails finds the questions no tool has ever listed, since a shopper who emails support is typing the search they could not get answered. Read a month of support tickets and chat logs and write down every question in the customer’s own words. “Will this sunscreen pill under makeup” and “does the heated vest fit over a hoodie” are both keywords, and both have a page one nobody has bothered to win.
These are also the phrases that convert, because the person asking is already holding the product in their head. Your support inbox is the only keyword tool your competitors cannot open.
Step 3: Add the Modifiers Buyers Actually Type
You add the modifiers buyers actually type to move from the crowded generic search to the open specific one. “Fire pit” is owned by Home Depot and Wayfair. “Cast iron fire pit for small patio” is owned by whoever writes the best answer, and the person typing it has a patio and a budget.
Think about what your buyer adds when the generic search fails them: size, skin type, room, climate, budget, use case. Those add-ons are where competition thins out and intent thickens. Ahrefs’ keyword data shows almost 93% of all U.S. keywords get 10 or fewer searches a month, which is the whole case for long-tail over short-tail keywords on a store with no authority yet.
Shopify’s own guidance on long-tail keywords makes the same point from the selling side: it says long-tail searches make up more than 90% of all searches and convert better because the intent is already defined. Specific searches are cheaper to win and worth more when you do.
Just know that specific does not mean obscure. “Sunscreen for tattoos” is specific. “Sunscreen for tattoos in Ohio” is made up. The test is whether a real buyer would type it, and the autocomplete list from Step 2 already answered that.
Step 4: Shortlist With a Difficulty Score
You shortlist with a difficulty score to cut a 300-phrase list to the 50 worth checking by hand, not to pick winners. Semrush’s keyword difficulty scale runs 0 to 100, with 0 to 14 rated “very easy,” 15 to 29 “easy,” and 30 to 49 “possible.”
That score is built mostly from the backlinks and authority of the pages already ranking, which is why it tracks the real fight at all.
The reason backlinks drive the score is that they still drive rankings. Ahrefs’ 2025 study of the top million U.S. keywords found referring domains correlate with rankings at 0.255, the strongest single factor they measured. A keyword where the top ten average two referring domains each is a fight a new store can enter. One where they average two hundred is not, however good the post is.
Treat the score as a filter, not a verdict. It shortlists fast, but it cannot see that result three is an unanswered forum thread from 2019, and that detail is often the whole opportunity. The score also cannot tell you how many backlinks you need for the specific phrase, since that number is the median of the pages you have to beat, and only reading page one gives you that.
If you do not pay for a tool, skip this step. Steps 1, 2, and 5 find the same keywords, just more slowly.
Step 5: Grade Page One by Hand
You grade page one by hand since it is the only source that shows you the actual pages you have to beat, and no score can stand in for that read. Search the phrase in a private window and answer three questions.
Who Is Ranking?
Who is ranking tells you whether the fight is winnable. Big brands and national publishers with complete answers mean move down your list. Forums, decade-old posts, thin pages, and stores your size mean the door is open. A page one with a Reddit thread in the top five is a page one you can take, often even when you rank without backlinks.
What Format Ranks?
What format ranks tells you what to build. If the top results are buying guides, write a buying guide. If they are product pages, point a collection page at the phrase. If they are forum answers, a clear post that answers the question in the first paragraph beats all of them. This is reading search intent off the page, and it decides the page type before you write a word.
How Good Is the Best Answer?
How good the best answer is tells you how much work the win takes. Read the top result like a customer. Is it complete, current, and specific to the buyer, or does it circle the question for 2,000 words? Beatable, matchable, improvable is the pass mark, and a phrase passes all three or it goes back on the list with the reason written next to it.
I will concede the uncomfortable part: this check is judgment, not math, and the first few reads feel uncertain. It sharpens fast. After twenty results pages you will spot an open door in ten seconds. Weak page one, real demand, buyer intent is the whole formula.
Step 6: Keep the Zero-Volume Winners
You keep the zero-volume winners because a phrase with no reported volume and a weak page one is often the best keyword on the list, and the tool is the thing that is wrong. Tools estimate volume from clickstream samples, and a phrase that gets 40 real searches a month can show as zero.
The Exclusive Beauty Club spokes are the proof. They came from autocomplete, not a volume column, and nobody checked whether “best sunscreen for tattoos” cleared a threshold before it got a post. Together the spokes took the hub to #2 for “best sunscreen 2026.” The full case for zero-volume keywords has its own post, and the short version is that your Search Console, not the tool, is the volume number that counts once the page is live.
So the rule is simple. A phrase that passed Step 5 stays on the list no matter what the volume column says.
Step 7: Assign Each Keyword to One Page
You assign each low-competition keyword to one page so the cluster never fights itself, and you pick the page type page one told you to build in Step 5. Two pages chasing the same phrase split the vote, and Google picks one, usually not the one you wanted.
| Keyword pattern | Page type | Example |
|---|---|---|
| Product + question | Blog post | does mineral sunscreen expire |
| Best product + use case | Buying guide post | best sunscreen for acne prone skin |
| Product + audience or size | Collection page | heated jackets for women |
| Exact product name | Product page | ventani hybrid heated jacket |
On Shopify, the page-type call has one extra wrinkle. A phrase like “heated jackets for women” can be a collection page or a buying guide post, and page one decides: product grids ranking means the collection, articles ranking means the post. Give the phrase to one of them and let the other one link to it, never target the same phrase from both, and never from a tag page, since Shopify tag pages are thin duplicates that fight your real collection for the same words.
Then slot every page into a cluster. The question posts link to the buying guide, the buying guide links to the collection, and the collection links to the products, so each small win feeds the next. That is how a list of low-competition phrases becomes topical authority on the subject, and the site that owns the questions ends up contesting the head term with a running start.
This is the order I run for every client, and the keyword map is the first deliverable in my Shopify SEO services: Search Console first, autocomplete second, page one by hand, then one keyword per page.
What to Do If Nothing Looks Winnable
Every Result Is a Big Brand
When every result is a big brand, keep the keyword phrase and move the angle to what a publisher cannot show: your products, your prices, your customers’ questions. “Niacinamide for skin” is owned by magazines. “Niacinamide serums: which strength to use” is owned by whoever sells them.
If no product angle exists, skip the phrase and spend the month on one that has one.
The List Is All Tiny Phrases
When the list is all tiny phrases, that is the list working. Fifty visits from buyers beat five thousand from browsers, and the head terms will still be there when you are ready. Nothing about starting small locks you out of going big. It means you arrive with a site that has already won something.
Low-Competition Keyword Checklist
- Pull queries at positions 8 to 20 from Search Console and group them by subject
- Type your product plus “for” and plus each letter into Google and harvest every suggestion
- Keep the phrases with buyer modifiers: size, skin type, room, climate, budget, use case
- Shortlist with a difficulty score under 30 if you have a tool, and skip the step if you do not
- Grade page one by hand: who ranks, what format ranks, how good the best answer is
- Keep zero-volume phrases that passed the page one check
- Assign each keyword to one page, in the format page one told you to build
- Link every page into a cluster so each win feeds the next
Want Your First Fifty Winnable Keywords?
Low-competition keywords are found in Search Console and Google itself, not in a tool’s difficulty column. Harvest real searches, keep the ones with buyer words attached, and confirm every pick by reading the page you would have to beat. Weak page one, real demand, buyer intent.
If you want the list without the legwork, book a call. I will pull your Search Console, run the harvest for your niche, and hand you fifty keywords your store can win this quarter, each one checked against page one. If you want a partner on it every month, that is what my Shopify SEO services cover.
