Keyword research is the easy part. Most people can fill a spreadsheet with a hundred terms in an afternoon. The hard part is turning one of those rows into a published page that sends the right people to the right place, and that's where most of the work dies.
This guide walks through the whole pipeline in ten steps, from choosing a keyword to measuring what it did, and shows what each step should produce before you move to the next one.
Why Most Keyword Research Ends in a Spreadsheet Nobody Uses
The usual process runs backwards. Someone exports a list of keywords, sorts by volume, picks the biggest one that looks relevant, and starts writing. Three weeks later there's a post that ranks for nothing, or ranks for something that sends visitors who leave in eight seconds.
The fix is to start from the end. Decide where you want traffic to go, find the question that leads there, then build the post as the bridge between the two. Every step below exists to keep that bridge straight.
Step 1: Start With a Money Page, Not a Keyword
This takes about fifteen minutes and needs nothing but your own website. Before opening any keyword tool, pick the page you want the post to feed. For a software or app company it's usually a signup, trial or pricing page. For a creator it might be a shop, a course page or a newsletter signup. For a service business it's the enquiry form.
Write the goal as a single sentence and keep it somewhere you'll see it during every later step.
| Money page | The single page you want the post to send people to |
| Reader | The kind of person who has the problem your page solves, described specifically |
| Realisation | What they need to understand before the money page makes sense to them |
| Action | The one thing you want them to do after reading |
That sentence rules out keywords that can't lead to the page, and it tells you what the post has to prove before it can ask for a click. Most content that gets traffic but no conversions skipped this step.
Step 2: Build a Seed List From Real Language
Give this about forty-five minutes. You're looking for fifteen to twenty-five raw ideas, and the best ones come from places where people describe their problem in their own words, before anyone has tidied it into a search term. Keyword tools are good at expanding ideas and bad at producing them.
● Search your support inbox, DMs, comments and sales notes for the questions that come up more than three times. The phrasing people use when they ask you directly is often the phrasing they type into search.
● Read the last month of threads in two forums or communities where your audience gathers. Note any question where the replies disagree or hedge. Disagreement means the existing answers are weak.
● Type your core topic into search and record every autocomplete suggestion, then do the same with "how to," "why," and "can I" added at the front.
Write the seeds down as they come, in the words people used, without sorting or judging them. Half the list will go in the next step, and that's fine. What you want here is range, not precision.
Step 3: Qualify Each Keyword With Three Questions
This is an hour's work with a keyword tool for volume and a browser for the results page. Every seed gets the same three questions, scored from 0 to 2, then totalled.
| Question | Score 0 | Score 1 | Score 2 |
| Intent match: does what the searcher wants lead naturally to the money page? | They want something else entirely, such as a free download or a definition | Related, but the path to your page is a stretch | Solving their problem and reaching your page are the same thing |
| Winnable: can a site like yours realistically appear in the top results? | The results are owned by major brands or government sites | Mixed; some strong pages, some weak | The results are thin, outdated or generic |
| Business value: does the searcher have a problem you solve? | Curious browsers, students, people with no budget | Possibly a customer, eventually | Someone with the exact problem, right now |
Sort by total, not by volume. The keyword that scores highest is almost never the one with the most searches. Broad, high-volume terms tend to fail on intent, because they come from broad audiences, and on winnability, because the biggest sites have already claimed them.
A keyword that scores 5 or 6 with a few hundred monthly searches will usually send more useful traffic than one that scores 2 with ten thousand. Pick one keyword, not three. Everything that follows is built for a single target.
Step 4: Map the Intent Before You Outline
Spend thirty minutes on the results page for your chosen keyword. Read the top ten results and any AI summary as if you were the searcher, and write a short note that becomes the top of the brief.
| Who is searching | Their situation at the moment they type the query, as specifically as you can put it |
| What they want | The two or three things they need to know or decide |
| Format that wins | What the top results have in common: length, structure, tables, lists, templates |
| What's missing | The things the searcher would want that none of the results provide |
| Next step after reading | What they'll do once they have the answer |
The last row is the bridge to the money page. Whatever the reader does after they have their answer is the moment your product becomes relevant, and it should appear at that point in the post, not bolted onto the top.
Step 5: Write the Brief
Forty-five minutes here saves hours later. This is the step most people skip, and it's the reason drafts wander. A brief fixes the target, the angle and the boundaries before a sentence is written, so that a writer, including you next week, can execute it without guessing.
| Target keyword | The one keyword from step 3 |
| Secondary keywords | Two or three close variants the post can cover naturally |
| Money page | Which page, how many times it's mentioned, and where |
| Angle | The tone and the stance in two sentences. How should the reader feel by the end? |
| Must cover | The points every top result already makes; you need them too, briefly |
| Unique element | The one thing from "what's missing" that this post will provide and nobody else does |
| Length and author | A range, not a target, and a named author with a reason to be believed |
| Internal links | Two or three existing posts to link to, chosen now so they're not forgotten |
| Do not | No history lesson, no definition opening, no padding |
The "do not" row saves more time than the "must cover" row, because it stops the draft before it drifts. Most bad drafts aren't missing anything; they contain too much.
Step 6: Outline Against the Brief
Thirty minutes turns the brief into a reading order. Each heading is the next question a real searcher would ask, phrased the way they'd ask it.
● Headings are questions, or close to it. A single-noun heading is a label. A heading that asks what the reader is wondering tells both people and search systems what the section answers.
● The order follows the reader's thinking, not yours. The first heading is the thing they need settled before anything else makes sense. The last is what they do next.
● The unique element goes in the middle, not the start. The commodity questions come first because searchers need them answered to trust you, and they stay short because the answers are known.
Aim for five to eight headings. Under each one, write the first sentence of the section as a direct answer. If you can't write that sentence from the outline alone, the heading is wrong or the research is incomplete, and it's far cheaper to find out now.
Step 7: Draft in Three Passes
Budget three to four hours across two sittings. Trying to write well and completely in one go is how drafts stall. Three passes, each with one job, is faster and tighter.
● Answer pass: write the first sentence of every section as a direct answer, then fill each with the plainest explanation you can manage. No evidence, no polish. An hour gets the whole shape down.
● Evidence pass: add the specifics from the brief. The figures, the examples, the sources, the table, the template. This is where the post stops being an opinion and becomes something a reader can act on.
● Cut pass: read every paragraph and ask what it's doing. Anything that restates, hedges, or warms up gets deleted. A draft usually loses 15 to 20 percent here and gets better with every cut.
The typical casualty of the cut pass is the paragraph that opens a section by explaining that the answer depends on your situation. It nearly always does, and saying so tells the reader nothing. Delete it and start with the answer.
Step 8: On-Page SEO Without the Anxiety
Give this forty-five minutes. On-page work isn't a list of forty boxes. It's making sure the page tells search systems the same thing the brief told you, and most of that is already done if the outline was built right. This step is mostly the title.
Write five options and pick by asking which one a person with the exact problem would click. Most titles fall into one of these patterns:
| Pattern | Strength | Weakness |
| The keyword as a "how to" plus what's included in brackets | Matches the query and promises the missing thing | Can run long; keep it under about 60 characters |
| The keyword followed by "the complete guide" | Familiar | Promises length, not an answer |
| A question the searcher is asking | Reads naturally | Weak if it leads with the least interesting question |
| A punchy benefit statement with the keyword buried | Sounds good out loud | Ranks poorly because the query isn't in it |
| The keyword plus three specifics separated by commas | Precise; says exactly what's inside | Often too long; trim to two specifics |
Then the rest, quickly: a URL of three or four words, a meta description that states the promise in one sentence, the internal links from the brief placed where a reader would want them, alt text on every image, and the basic signals that help a blog post rank.
Step 9: Publish and Do the Three Things That Move It
An hour on publish day, then fifteen minutes a week. After publishing, three actions matter and everything else is noise for the first month.
● Go back to the related posts named in the brief and add a link to the new page from each, with anchor text that describes it. Then find one more older post that touches the topic and link from there too.
● Post it in one place where the question was actually asked: the community thread from step 2, a relevant group, or your own newsletter. One place, done well, beats ten done carelessly.
● Reply to the next few messages you get on the topic with a link to it. Real readers who arrive with the exact problem and stay to the end are the best early signal a page can get.
What you don't do is check the ranking daily. Positions move for weeks. Leave it alone.
Step 10: Measure Against the Money Page
Twenty minutes, once at thirty days and again at ninety, with Search Console and your analytics open. The question isn't "did it rank." It's "did it send the right people to the page I chose." Four numbers answer that.
| Number | Where to find it | What it tells you |
| Impressions for the target keyword | Search Console, filtered to the page | Whether the page is being considered for the query at all |
| Clicks and click-through rate | Search Console | Whether the title and description are earning the visit |
| Clicks from the post to the money page | Analytics, internal link tracking | Whether the bridge is working |
| Conversions with the post as the landing page | Analytics, conversions by landing page | Whether the whole pipeline paid off |
At thirty days you're looking for impressions and early clicks. At ninety days the last two rows decide what happens next: update the post if it's working, sharpen the money page link if traffic arrives but doesn't cross, or revisit the keyword if the wrong people are showing up.
The Pipeline on One Page
The whole process, compressed. The total is around ten hours for one post, with most of it before and after the draft rather than in it.
| Step | Time | What you end up with |
| Money page | 15 min | One-sentence goal |
| Seed list | 45 min | 15 to 25 raw questions in real language |
| Qualify | 1 hr | Scored table, one keyword chosen |
| Intent map | 30 min | Five-line note |
| Brief | 45 min | One-page brief |
| Outline | 30 min | Headings as questions |
| Draft | 3 to 4 hrs | Clean draft after three passes |
| On-page | 45 min | Title, URL, meta, links, images |
| Publish | 1 hr | Linked, shared, in front of readers |
| Measure | 20 min, twice | Four numbers, one decision |
The Verdict
The keyword with the most searches is rarely the one worth writing for. The one worth writing for sits between a real problem and a page you want people to reach, and finding it takes more care than any keyword tool can give you.
One post built this way is worth more than five built by picking the biggest number and hoping. Not because the writing is better, though it usually is, but because every step points at the same page. A post that knows where it's sending people is the only kind that earns its keep.
