Target keywords for SEO: how to pick topics that rank

We wasted three months once chasing a keyword that looked perfect in a tool: decent volume, “low competition,” totally relevant to our product. We shipped the article, polished the screenshots, pushed internal links, did the whole routine. It never broke page 4.

The painful part was not that we failed. It was why we failed.

We had picked the wrong promise.

That is the real work behind target keywords for seo: you are not picking words to sprinkle into a page. You are choosing a specific intent you are willing to satisfy better than the ten pages already winning.

A target keyword is a promise, not a label

Most teams treat keywords like tags. Product page equals “project management.” Blog post equals “best project management tools.” Then they wonder why Google sends them the wrong readers, or no readers.

A target keyword is closer to a contract. If someone types that phrase, they are hiring a page to do a job. If your page cannot do that job in one sitting, you are signing up for bounce rates, weak engagement, and the slow, quiet death of “why is this not ranking?”

Potential friction shows up fast here: people confuse “target keyword” with “topic,” so they choose phrases that describe their product, not the problem the searcher is trying to solve. “Invoice software” feels like a topic. The job might actually be “how to invoice a client as a freelancer,” which is a different audience, different page, different funnel.

We keep the definition simple on purpose. The hard part is the decision system.

The scorecard we use to pick target keywords for SEO (and the kill rule that saves months)

Tools will happily dump 2,000 suggestions into a CSV and let you pretend you made progress. The work is reducing that pile into a handful of targets you can actually win.

We use a three-factor scorecard: relevancy, competition, and monthly search volume. Then we apply a kill rule for intent mismatch.

This is not fancy. It is repeatable. And it forces trade-offs into the open.

Step one: the relevancy test (a harsh yes or no)

Relevancy is not “is this related to what we sell?” Relevancy is: can we create one page with one clear purpose that satisfies the query using our product, our expertise, or our experience.

We ask three questions:

First, can we answer it without twisting ourselves into a pretzel. If the query is “best backyard chicken breeds” and you sell coops, you can contribute, but you are not the authority on breeds unless you actually are. Your page will feel like a sales detour.

Second, can we satisfy it with the page type we are willing to build. Some queries demand a calculator, a template, a comparison grid, a local landing page, or a product category page. If we only want to write a blog post, we score it lower.

Third, can we state the page’s promise in one sentence. If we cannot, we do not have clarity. No clarity means weak structure and vague copy.

We score relevancy as pass or fail. Not 7 out of 10. Not “kind of.”

What trips people up is trying to force a keyword because it has volume. That is backwards. If relevancy is not perfect, you do not have a target keyword. You have a distraction.

Step two: the competition test (manual, messy, and worth it)

Keyword tools often label “competition” based on advertisers, not how hard organic rankings are. That label can still be useful, but it is not the truth.

So we do a manual SERP check. It is low-tech, but it stops us from shipping the wrong format.

Open an incognito window, set your location to match your target, and search the phrase. Then scan the first page and answer:

Are the results actually similar to what we plan to publish.

If the top results are definitions and explainers, and we plan a product page, we are fighting the SERP. If the top results are product pages and “best of” lists, and we plan a 3,000-word educational guide, we are also fighting the SERP.

We also look for strength signals. We are not doing a full backlink audit for every idea, but we do quick proxies: recognizable brands, obvious topical depth, strong internal linking, and pages that look maintained. If page 1 is stacked with household names and government sites, we assume we need more authority than we have.

A note on result counts: some guides suggest using the number of results as a competition proxy, like how “social media company” can show roughly 2 billion results. We treat that number as a vibe check only. It is noisy, and Google inflates it. Still, when a phrase returns a giant SERP size and the results look broad, it is usually a warning sign.

Step three: the volume test (accept ranges, not fantasies)

Once relevancy passes and competition looks survivable, we care about monthly search volume.

Here is the nuance most teams miss: you do not need a massive number for a keyword to be worth it, if intent is tight and the page can convert. A range like 100 to 1K monthly searches can be a great target. The Keyword Planner example we keep in our notes is “best backyard chicken breeds” showing 100 to 1K with low competition. That is a realistic kind of opportunity.

Where this falls apart is chasing high volume phrases that are hard to rank for and full of mismatched intent. Even in that backyard chicken niche, “backyard chicken breeds” can show higher volume but be much harder. That is the trap: bigger number, bigger fight.

The kill rule: if intent does not match page 1, we discard

This is the rule that would have saved us that three-month failure.

If we search the phrase and page 1 is clearly serving a different intent category than what we can satisfy, we kill it. Even if volume is high. Even if a tool says low competition.

No heroics. Just move on.

Seed to cluster: generating long-tail ideas without drowning in duplicates

We start with seed keywords because they keep us anchored to reality. A seed is not a target. It is the start of a family.

We get seeds from sales calls, support tickets, and the phrases customers use when they are frustrated. Moz has good guidance here: talk to customers. We do it because marketing language lies. Customer language pays.

Then we expand in Google itself.

Autocomplete gives us real phrasing. People Also Ask gives us the questions that Google thinks are adjacent. Related Searches shows the variants people try when they refine.

We keep a working rule for long-tail: 3 to 4+ words. Not because word count is magic, but because specificity usually means clearer intent and less competition.

Potential friction: you will collect a pile of near-duplicates and “questions” that your page cannot satisfy. The fix is to cluster by intent, not by wording. “how to choose running shoes for flat feet” and “best running shoes for flat feet” are close in topic, but the intent differs: one is educational guidance, the other is shopping.

Our trick is to name clusters like jobs, not keywords. “Pick a shoe” is a job. “Treat plantar pain” is a different job. Then we pick targets inside the cluster.

Anyway, we once had a spreadsheet where we argued for 45 minutes about whether “for beginners” counts as a modifier. We were tired and hungry. Back to the point.

Google Keyword Planner: the workflow that matches real constraints

Keyword Planner is still useful because it is first-party data and it forces you to see volume bands, not made-up precision.

You do need a Google Ads account to access it. You do not need to run ads or have active spend. That detail matters when people assume there is a paywall.

Once inside, the flow we use is simple.

Go to Tools, then Keyword Planner. Choose Discover new keywords. Start with seed keywords, or start with a website URL if you want Google to read a competitor and suggest themes. We use both depending on the niche. For new markets, starting with a URL can surface language we would never think to type.

Set location. Always. Geo-targeting changes volumes and the shape of questions, and it prevents you from building a US content plan using global demand. We often set United States first, then rerun for other regions if the business is international.

Export the ideas. Work offline. Keyword tools are built to keep you clicking. Spreadsheets are built to help you decide.

The annoying part is interpreting the numbers. Keyword Planner often gives broad ranges, and its “competition” column reflects advertiser density, not organic difficulty. We use those fields as filters, not as the final answer.

If a term sits in a band like 100 to 1K and relevancy is perfect, we keep it in play. If a term is huge volume but the SERP is dominated by big sites and a different format, we do not talk ourselves into it.

Manual SERP validation: the fastest way to stop publishing the wrong thing

Most keyword guides say “check search intent” and then move on. That is not enough. The actual problem is not naming intent. It is deciding what to build when the SERP is mixed, and spotting SERP features that steal clicks.

We do this in two passes: intent labeling, then feature impact.

The 10-result intent labeling method

Take the top 10 organic results and tag each one:

Informational: guides, how-tos, explanations.

Commercial investigation: comparisons, “best,” “top,” reviews.

Transactional: product pages, categories, sign-up pages.

Local: map packs, location pages, “near me.”

Navigational: brand pages, login pages, people trying to reach a specific site.

Then count.

If 7 of 10 share one intent, we match it. We do not fight it.

If it is split, we pick a primary intent and design the page to win that, then we capture the secondary intent with an FAQ section driven by People Also Ask.

This is where real teams get tripped up: they see a keyword that feels relevant, then publish the wrong format because they are thinking about what they want to write, not what Google is already rewarding.

Example: if page 1 is “best X” lists and roundups, and you publish “what is X,” you might write a great article and still lose. Google is trying to help a shopper. Your definition post is not the best tool for that job.

SERP feature impact: the click-thieving checklist

Even if you rank, you might not get traffic.

Before we commit, we look for SERP features and adjust expectations:

If People Also Ask boxes are huge, we plan to answer those questions directly, clearly, and early. If you can win a PAA result, you can steal attention even without a number one ranking.

If there is a featured snippet, we look at what it rewards. Is it a list, a short definition, a comparison paragraph, a set of steps. Then we structure a section to compete for that shape.

If video packs dominate, we ask whether the query is better served with a video-first asset. Sometimes the right play is to publish a written guide and pair it with a short video, then embed it.

If a local pack is present, and you are not a local business, you might be fighting for scraps.

None of this is theoretical. We have watched a page rank number 2 and still underperform because the featured snippet and PAA ate the clicks.

Turning a keyword into a page plan (without cannibalizing yourself)

After you pick a target, you still have to avoid a common self-inflicted wound: publishing multiple pages that satisfy the same intent, then splitting ranking signals.

We use a practical compromise.

One primary keyword per page. Then we include close variants through natural stemming, meaning we write like humans and cover the same intent using the phrases people actually use. RankYa calls this keyword stemming and frames rankings as working on a “one word basis.” We do not take that literally, but the underlying idea is useful: you do not need to force exact matches if your page thoroughly covers the job.

We create a simple mapping rule: if two phrases would produce the same outline and the same call to action, they belong on one page. If they require different formats, different examples, or a different audience, they get separate pages.

Potential friction: teams create “SEO landing page” and “SEO landing pages” and “landing page SEO” as separate URLs because the keywords look different in a spreadsheet. Google sees one intent. You just diluted your own authority.

On-page placement that helps, and the tripwires that hurt

Once you have the right target, on-page work is not mysterious. It is also easy to overdo.

Put the primary phrase or a close variation in the URL if it fits naturally. Use it in the title. Mention it early in the page, where it belongs. Use descriptive headings that reflect the sections a reader expects. Add it to the meta description if it reads like English.

Images matter mostly for clarity and accessibility. Rename files and write alt text when it naturally describes the image. If you start stuffing the exact phrase into every alt attribute, you will make the page worse.

Over-optimization is real. Repeating keywords unnaturally can drag quality down, and quality is the whole game now. Write helpful, reliable, people-first content, then make sure your targets are visible to both readers and search engines.

The 200-character keyword list rule: what to do with outdated advice

You will still find guidance that says you should keep your target keyword list under 200 characters and even use meta keyword tags. That comes from older playbooks, and it is not how modern SEO works.

Google does not use the meta keywords tag in the way people imagine, and rigid character limits for a “target list” miss the point. What is still useful in that old advice is the discipline: pick fewer targets, make them specific, and stop chasing vague one-word terms.

If you want a modern version of the constraint, use this instead: can you explain your page’s target intent in one sentence, and can you list 3 to 6 close variants that share the same intent. If the list grows to 30, you are probably mixing jobs.

A fill-in template you can run every time

When we are tired, we rely on process. Here is the scorecard template we paste into a doc and fill in. Keep it blunt.

• Relevancy (pass or fail): Can we satisfy this query with one page, one purpose, using our product or expertise without a bait-and-switch.
• Competition (low, medium, high): What do page-1 results look like, are they strong brands, are they deep and maintained, is the format the same as what we can publish.
• Volume (range): What is the monthly search band in the target location, and is it enough given expected conversion.
• Kill rule (yes or no): Does page 1 clearly serve a different intent category than what we can serve.
• Tie-breaker: If two keywords are close, we pick the one with clearer intent and a SERP where we can match the dominant format.

That is the whole thing. No secret sauce.

If you do this consistently, you stop shipping content that was doomed at keyword selection. You start picking fights you can win.

And when you still lose sometimes, which happens, you will at least know why. That alone is worth the extra hour on the SERP.

FAQ

Which keywords are best to target in SEO?

The best keywords are the ones where you can match the dominant page-1 intent with a single clear page. Start with perfect relevancy, then choose the lowest competition SERP you can realistically compete in, even if volume is only in the 100 to 1K range.

What is the 80/20 rule for SEO keyword targeting?

Spend 20 percent of your effort on keyword selection and SERP validation, because it prevents 80 percent of content that will never rank. If the intent and format match page 1, most on-page optimization becomes straightforward.

Is SEO dead or evolving in 2026?

SEO is still alive in 2026, but it is more intent and distribution driven than keyword stuffing driven. You still validate in Google, but you also plan for SERP features like snippets, People Also Ask, video packs, and other click-stealing elements.

How do we avoid targeting multiple keywords that compete with each other?

Use one primary keyword per page and merge phrases that would produce the same outline and the same call to action. If two keywords require different page formats or serve different audiences, split them into separate pages.