10 Smart Questions to Ask Before Choosing a Keyword

You've got a list of 80 keyword ideas open in a spreadsheet, your cursor is hovering over one of them, and you have absolutely no idea whether it's the right call. So you do what everyone does: you sort by volume, pick the biggest number, and move on. Four months later that page has 300 sessions and zero conversions to show for it.

I've watched this movie more times than I'd like to admit. On my own site, I once spent five weeks writing a long-form piece around a keyword that pulled a respectable-looking 9,400 monthly searches. It ranked. It got traffic. It brought me one newsletter signup. One. Because the searchers wanted a quick definition, and I gave them a 4,000-word essay.

That was the day I stopped treating keyword selection as a numbers game and started treating it as a set of questions—questions I now ask before I write a single word. Here are the ones that actually change the outcome.

What questions should you ask before choosing a keyword?

Six of them, if you want the honest short answer. But they don't carry equal weight, and pretending they do is where most people go wrong.

What is the searcher actually trying to do?

Open a private browser window and type your keyword in. Look at what fills the first screen. Not the titles—the shape of the results.

If you see ten listicles, the searcher wants a list. If you see product pages, they're in buying mode. If you see a wall of "what is X" definitions, they want an answer in the first three lines and then they're gone.

The mistake I made for years: I assumed I could out-write the format. You can't. Google has already tested what works for that query, and it's showing you the answer key. Fighting it is like bringing a salad to a barbecue and expecting everyone to be excited about it.

Ask yourself: does the format I'm planning match the format that already ranks? If not, change your plan, not your ambitions.

Does this keyword bring people who could ever buy?

Here's the uncomfortable truth about high-volume keywords: a lot of that volume comes from people who will never become customers. Students writing papers. Curious onlookers. People who typed a phrase out of boredom.

I keep a rough filter in my head for this. I'll ask: if this person landed on my best page and loved it, would there be a next step for them that makes me money? If the answer is "no, not really," the keyword goes in a different bucket—the one labeled brand awareness, maybe, someday, if I have spare time.

Which is basically never.

How hard is it to actually rank?

Keyword difficulty scores are useful but oversold. They compress dozens of signals into one number, and that number can lie.

What I do instead: I look at who's ranking on page one and ask whether I could plausibly beat them. Are they giant brands with editorial teams? Or are they small sites with thin content and a mediocre structure? The second group is where you hunt.

For context, roughly a third of the pages I've launched over the past few years landed on page one within six months—and almost all of those were against small competitors, not against established publishers. The ones I lost were the ones I picked thinking "I'll just write it better."

Is this keyword already paying rent somewhere on my site?

Cannibalization. Say it out loud a few times, because you will run into it eventually.

Two of your pages, both optimized around the same idea, both competing for the same query. The search engine picks one. It might not be the one you want. And now you've spent twice the effort for less than half the result.

Before I commit to a keyword, I search my own site for the core phrase. If something close already exists, I either improve that page or I deliberately angle the new one toward a distinctly different sub-intent. No half-measures.

How much will this page cost me?

Not in dollars—in hours. Every keyword has a hidden price tag that equals the effort required to do it properly.

A "what is X" definition might take you two hours. A "best X for Y" comparison with original testing might take you three weeks. Both can be worth it, but only if the traffic and conversion math works out.

I keep a simple rule: if I can't see myself producing the content at a quality level that beats the current top three, I don't take the keyword. Not yet. Maybe never.

Is there a clear next step for the reader?

This is the one I almost never see people ask, and it's the one that separates keyword selection from keyword guessing.

After the reader finishes your page, what do you want them to do? Subscribe? Download? Buy? Book a call? If the keyword attracts a searcher who has no interest in that next step—even if they love your writing—you've built a beautiful dead end.

And dead ends don't pay the bills.

How do you prioritize when every keyword looks good?

They won't all look good. But a few will, and that's enough to freeze you.

I use a scoring approach that's deliberately crude. It beats nothing and it beats analysis paralysis.

A simple scoring framework you can steal

Rate each candidate keyword on four dimensions, from 1 to 5:

1. Intent match—does the format I plan match what ranks?
2. Business relevance—could this searcher ever buy from me?
3. Difficulty odds—can I realistically beat the current top results?
4. Effort fit—do I have the time and resources to do this properly now?

Total score out of 20. Anything below 14 goes back in the pile. It's not scientific. It doesn't need to be. The point is to force yourself to stop evaluating on volume alone.

I'll admit I resisted scoring for a long time because it felt like busywork. Then I noticed I was picking keywords based on how interested I was in writing about them, which is a terrible strategy. The score doesn't remove judgment. It just stops my mood from being the deciding vote.

When to ignore the volume number

Volume is a signal, not a verdict. A few cases where I'll happily choose the smaller number:

• The keyword has clear commercial intent and the bigger one doesn't.
• The keyword matches a topic I've already written about, so I can link it into an existing cluster.
• The keyword targets a question my audience asks me in emails. That's free research, and it beats any tool.
• The bigger keyword is dominated by pages I can't touch—news sites, government portals, enormous marketplaces.
• The smaller keyword has a longer tail I can expand into over time.

Notice that in each case, the deciding factor is something other than search volume. That's not a coincidence.

Impact versus effort, in plain terms

Here's a table I keep in my notes. It's saved me more wasted weeks than any single tool.

Scenario	Typical volume	Effort	Worth it?
Short definition keyword	High	Low	Yes, if your page leads somewhere
Buyer-intent comparison	Medium	High	Yes, usually the best ROI
Broad "how to" tutorial	High	Very high	Only if you can genuinely win
Niche problem-solving query	Low	Medium	Yes, especially as part of a cluster
Trend-based keyword	Volatile	Medium	Rarely, unless you move fast

The buyer-intent row is the one I'd defend to the end. It's almost never the flashiest keyword. It's almost always the one that pays.

What changes about keyword selection in 2026?

The landscape has shifted, but not as dramatically as the panic headlines suggest.

Answers are borrowed more often now

Conversational search interfaces surface an answer before anyone clicks. That means informational keywords designed purely for "tell me what this is" are the weakest they've ever been. The click was already shaky; now it's often gone entirely.

The counter-move is simple: pick keywords where the reader needs to do something, not just know something. Comparisons. Templates. Walkthroughs with real decisions embedded. These are harder to satisfy in a single paragraph of synthesized text.

Zero-click is not a death sentence

Yes, a large share of searches now end without a click. But the searchers who do click tend to be more motivated than they used to be. If someone ignores the on-page answer and visits you anyway, they're not window shopping.

I've noticed this in my own analytics: my sessions are down year over year on a couple of informational pages, but the conversion rate on those pages is up. Fewer people, better fit. I'll take that trade every time.

AI overviews reshape your question list

When a query's answer can be paraphrased in three sentences, treat it as brand awareness, not traffic acquisition. When it can't—because the answer depends on nuance, sequence, or context—treat it as a real opportunity.

This one filter alone has changed which keywords I write about more than any tool update in the past two years.

Questions I ask myself out loud before saying yes

By now you have the framework. Here's the quick version I actually run through, in order, whenever a keyword tempts me:

• Can I describe the reader in one sentence?
• Does the format I'm planning match what already ranks?
• Is there a realistic path from this page to revenue, or to an audience that will eventually become one?
• Is anyone on my site already chasing this query?
• What's the honest effort, and do I have the bandwidth right now?
• Am I picking this because it's good, or because it's interesting to me?

That last one is the trap. I fell into it constantly when I was starting out, and I still catch myself doing it. The keywords that excited me most were usually the ones nobody was searching for.

Pick the keyword that answers your reader's question, matches your business, and fits your capacity this week. Everything else is a distraction dressed up as opportunity.

And if you're staring at that spreadsheet right now, undecided between three options—close your eyes, ask the six questions, and go with the one that scores highest on business relevance rather than the biggest number. Your future self, four months from now, will quietly thank you.