You've got a list of 400 keywords in a spreadsheet. Volume looks fine, the tool says competition is "low," and you still have no idea which five to write first. So you pick the biggest numbers, publish, and six months later the traffic is a trickle that never buys anything.
This guide shows you how to find profitable keywords for my niche (or yours) without that guesswork. You'll get a repeatable process, the free tools that cover most of it, the checks that separate a good-looking keyword from a money-making one, and an honest view of where manual research stops being worth your time.
What makes a keyword profitable in the first place
A profitable keyword is one that brings the right visitor, that you can realistically rank for, and that leads to revenue. Miss any one of the three and the keyword is just a number.
Volume is the least important of them. A term with 90 searches a month that attracts people ready to buy will beat a 9,000-search term full of browsers. Treat volume as a tiebreaker, not a goal.
Here's the filter I'd apply to every candidate:
- Intent fit: does the searcher want what you can give them, whether that's a guide, a comparison or a product page?
- Winnability: do the current top results include sites with a profile like yours, or only giants with thousands of backlinks?
- Money path: can you tie the topic to a product, service, affiliate offer or lead form without forcing it?
- Volume: is there enough demand to matter, even if it's small?
Here's what a failure looks like. Say you run a small espresso-gear site and target "espresso machine," which shows roughly 74,000 searches a month. Page one is Amazon, Breville, a big review publisher and a couple of retailers. You publish a 2,000-word guide, it lands around page 6, and it collects maybe 15 visits a month. Meanwhile "why is my espresso sour," at around 700 searches, had a forum thread in the top three and no real brand in sight.
Before you write a word, count the small sites on page one. If the count is zero, the search volume doesn't matter.
Step 1: Start with narrow seed topics
Good niche research starts narrow. Instead of "coffee," write down "espresso machine descaling" or "cold brew for small cafes." The narrower the seed, the more specific (and less contested) the ideas it produces.
Spend ten minutes listing seeds from four places:
- Your own product or service: every feature, problem solved and customer type.
- Customer questions: support emails, sales calls, DMs, reviews.
- Communities: subreddits, forums and Facebook groups where your audience complains and asks for recommendations.
- Competitor navigation: the categories and blog sections your rivals think matter.
Communities deserve more of your time than they usually get. People type into forums exactly how they'd type into Google, with typos and all. A thread titled "my shots taste sour no matter what I change" on an espresso subreddit turns straight into the query "why is my espresso sour," and the replies tell you which fixes (grind size, water temperature, shot time) the article needs to cover.
Step 2: Expand seeds with free tools
You don't need a paid subscription to build a solid list. Three Google tools cover the basics, and they cost nothing in subscription fees.
Google Keyword Planner
Per Google Ads' own documentation, Keyword Planner can generate ideas from a word or a website and shows estimated volume, advertiser competition and bid ranges. Those bid ranges are quietly useful: a keyword advertisers pay well for usually has commercial value. One catch. It needs a Google Ads account, and some features may ask for billing information, as Digidop's 2026 tool comparison points out. Also, "competition" there means advertiser competition, not how hard it is to rank organically.
Google Trends
Trends lets you compare wording, regions and seasonality. Its scores are relative, not monthly search counts, so use it to decide between "standing desk converter" and "sit stand desk riser," not to forecast traffic. The rising queries section can also surface subtopics before they're crowded.
For seasonal niches, look at a five-year trend in your target market, then publish or refresh before demand climbs. Don't assume when the peak is. Check it. Plenty of "summer" topics start building in February.
Google Search Console
If your site already has pages, Search Console is the most honest keyword tool you own. Look for queries that earn impressions but few clicks. A page sitting at position 11 for a promising term is often a faster win than a brand-new article. Update it, tighten the title, add the missing section.
Autocomplete and People Also Ask
Type your seed into Google and note the suggestions, then scan the People Also Ask box. It's free, instant and shows the questions real people are asking. For a niche like "sourdough baking," you'll see long tails such as "why is my sourdough starter not rising" that no volume-obsessed list would give you.
Step 3: Check search intent before anything else
Search the keyword yourself and read the results like a customer. This takes about 60 seconds per term and saves weeks.
Ask what format dominates. If the first page is all comparison posts, a how-to article will struggle. If it's product pages, a blog post is the wrong tool. If you see forum threads ranking high, that often signals Google can't find great content, which is an opening.
I'll be blunt: skipping this step is the most common reason niche blogs stall. People chase a keyword, write a lovely guide, and discover the whole page one is shopping listings.
Step 4: Judge competition the way Google does
Ignore the tool's difficulty score as your only verdict. Treat it as a hint and confirm in the live results.
When you open the top ten, look for these signals:
- Site type: are results from huge brands, or from small blogs and niche sites like yours?
- Content quality: is it thin, outdated or generic? A 2019 post ranking first is a gift.
- Specificity: does anyone cover your exact angle, or just the broad topic?
- Freshness: are the winners recent, or has nobody touched this in years?
If two or three small sites are ranking, the door's open. If it's all household names, move on or go more specific. We break down this process in more depth in how to find winnable keywords on Google, and if you want terms with real demand but thin competition, this guide to low-competition, high-volume keywords covers the trade-offs.
Step 5: Test the money path
A keyword is commercially promising when it connects naturally to something you can sell or recommend. Here's a quick way to sort candidates by how close they are to a purchase.
| Keyword type | Example | Buying intent | Typical use |
|---|---|---|---|
| Problem-aware | why is my espresso sour | Low to medium | Trust-building guide with a soft product mention |
| Comparison | best grinder for espresso under $200 | High | Roundup or affiliate page |
| Solution-specific | descaling solution for [machine model] | Very high | Product or category page |
| Broad informational | what is espresso | Low | Skip unless you need topical coverage |
Prioritise the middle two rows first. They're where niche sites make money. Broad informational terms feel safe because volume is big, but they rarely convert and they attract the toughest competition. For more on picking terms that actually turn into customers, see how to find low-competition keywords that convert.
If you can't name what a reader would do after finishing the article, rethink the keyword. A genuinely useful post with no next step is a hobby, not a strategy.
Step 6: Score and shortlist your niche keywords
Now you've got candidates, so make a decision. Put them in a sheet with columns for estimated volume, intent, winnability (1 to 5), and money path (1 to 5). Add the last two and sort. That sheet is the whole answer to how to find profitable keywords for my niche once the raw research is done.
A few practical notes:
- Treat volume as an estimate. Keyword Planner may give broad ranges, and each SEO tool uses its own model, so two tools will disagree. Validate in the live results.
- Don't discard tiny keywords. Ten posts at 100 targeted searches often beat one post chasing 5,000.
- Cluster related terms. One strong page can rank for dozens of variations, so you don't need a separate post for each.
Two or three clusters at the top of the sheet is usually plenty for the next month.
Do you need paid tools?
Not at the start. Google's free trio is enough to build a first batch of 20 to 30 targets. Paid tools earn their place once you need competitor gap analysis, better difficulty estimates or rank tracking at scale.
As for cost, reported November 2026 prices run from Ahrefs Starter at $29/month to Semrush Pro at $139.95/month, according to Appterly's beginner tool roundup. That's a real range, so don't buy the expensive one before you've exhausted the free options.
My recommendation: free tools for your first month, then one paid tool only if you can name the specific question it'll answer that the free ones can't.
A weekly routine for how to find profitable keywords for my niche
Research only works if it becomes a habit. Here's a routine that takes about two hours.
- Monday: pull new queries from Search Console and flag pages with high impressions and low clicks.
- Tuesday: add five fresh seeds from customer questions or communities, then expand them in Keyword Planner and autocomplete.
- Wednesday: check intent and competition in the live results for your top ten candidates.
- Thursday: score, cluster and pick two to publish.
- Friday: check where your existing pages rank and queue refreshes.
Where manual research breaks down
The process above works. It's also slow. Doing intent checks and competition reviews for 30 keywords is an afternoon, and then you still have to write the posts, in a consistent voice, and publish them. Most small teams stall at that point, not at research.
This is the part Spook takes over. It looks for Google and ChatGPT queries where a site like yours has a real shot, writes the article in your brand's voice, and publishes it to your website. In practice, the Wednesday and Thursday steps above (live-result checks, scoring, picking) stop being your job, and what's left is reading what went live and deciding what to refresh. Manual research is still the right call if you're publishing a handful of posts a year. Past that, the afternoon per batch adds up fast.
Whatever route you take, start with Search Console. Find three queries with impressions but few clicks and check the live results for each. The pattern is usually the same: a page stuck around position 11 to 15 with a weak title, which you can fix in twenty minutes before writing anything new.
Spook is an AI-powered SEO tool that identifies winnable Google and ChatGPT queries, writes optimized content in your brand's voice, and publishes it to your site. It's built for website owners and marketers who want the keyword research and the content done without spending every week on spreadsheets.