Skip to main content

Content Writing

How to Do Keyword Research for Blog Posts: 7 Steps

Start from business topics, expand them with free and paid tools, group phrases by shared results, read the SERP, then choose one primary keyword per post.

Content Writing Square Team Editorial Team 7 min read

Keyword research for a blog post means finding the words people actually search for a topic your buyers care about, checking what Google already shows for those words, and choosing one main query the post can realistically win. The process is: start from business topics, expand them into real search phrases, group phrases that share the same results, read the search results to judge intent and competition, then pick a primary keyword and a handful of related ones for the brief.

Search volume is one input, not the decision. A phrase searched a few dozen times a month by exactly the people you sell to is usually worth more than a big number attached to the wrong audience.

Step 1: Start From the Business, Not the Tool

Open a keyword tool with no starting point and it will suggest thousands of things your business has no reason to write about. Start with three short lists instead:

  • Your services or products, in the words customers use for them.
  • The problems that lead people to you, described as a customer would describe them before they knew you existed.
  • The questions sales and support hear repeatedly. These are often the best seeds of all.

Each item is a seed. Five to fifteen seeds is plenty for a quarter’s worth of posts.

Step 2: Expand Seeds Into Real Search Phrases

Free sources

  • Google itself: type a seed and note the autocomplete suggestions, the “People also ask” questions and the related searches at the bottom of the page. These reflect real searches, phrased the way people phrase them.
  • Google Search Console: the Queries tab shows phrases you already appear for. Queries with impressions but poor positions are often the quickest wins.
  • Google Keyword Planner: free with a Google Ads account. Be aware that accounts with little or no ad spend generally see volumes as broad ranges (such as 1K–10K) rather than exact numbers, a limitation Google introduced in 2016, as Search Engine Journal reported at the time. Useful for direction, not precision.
  • Google Trends: not for volume, but for seasonality and whether interest is rising or falling.

Paid tools

Ahrefs, Semrush and similar tools add volume estimates, difficulty scores, question reports and lists of what competitors rank for. They are faster, not fundamentally different. Treat every number they show as an estimate: their keyword databases cannot include every query, and Ahrefs itself notes in its research that its traffic figures are likely to be underestimates for that reason.

Step 3: Throw Out What Your Buyers Would Never Search

Go through the expanded list and delete anything that fails one question: would someone who might buy from us plausibly type this? Students, job seekers, people looking for free DIY instructions on something you sell as a professional service — they are real searchers, just not yours. This cut usually removes half the list, and it is the most valuable ten minutes of the process.

Step 4: Group Phrases That Belong on One Page

Many phrases are different wordings of one need: “garden design cost”, “how much does a garden designer charge” and “garden design prices” want the same answer. They get one post, not three.

A reliable test is to search two phrases and compare the top ten results. A common working rule is that if three or more of the same URLs appear in both, Google treats them as one topic and one page can rank for both. If the results are completely different, they need separate pages. Our guide to how many keywords to target in one post goes deeper into grouping.

This matters for traffic as well as tidiness. Ahrefs’ study of 3 million searches found that the average page ranking first also ranked in the top ten for nearly 1,000 other related keywords (the median was around 400). A well-chosen page earns traffic from the whole group, not just the phrase you picked.

Step 5: Read the Search Results Before You Decide

For each group, search the main phrase and study the first page. You are answering two questions.

What do searchers want?

Are the results guides, lists, product pages, local listings, videos, tools? If the top ten are all service pages, the query wants a supplier and a blog post will struggle; if they are all how-to guides, a sales page will. The format you see is the format Google has learned people prefer. We cover this properly in how to understand search intent before writing.

Can we realistically beat what is there?

Difficulty scores are a rough guide, mostly based on the backlinks of ranking pages. Reading the results is more useful. Signs of an opening:

  • Forum threads, Reddit posts or Q&A pages ranking on page one — Google is short of a proper answer.
  • Pages that are several years old or visibly out of date.
  • Results that answer a slightly different question from the one searched.
  • Small or specialist sites ranking alongside big ones.
  • Thin pages: definitions with no examples, numbers or detail.

Signs to look elsewhere: every result is a major brand or national publication, every page is comprehensive and recent, and the results exactly match the query.

Step 6: Think in Traffic Potential, Not Single Volumes

Because pages rank for groups of phrases, the volume of your chosen keyword understates what the page could earn. A better estimate is the organic traffic the current top-ranking page receives, which most paid tools show. If you have no paid tool, look at the size of the group from Step 4: a topic with twenty phrasings has more potential than one with two.

Equally, do not dismiss phrases a tool reports as zero or near-zero volume if customers ask them. Tools under-report niche and B2B queries, and a post that answers a real buyer question has a job to do in sales conversations even if it never ranks.

The opposite trap is writing on topics with no search demand at all and expecting traffic. In Ahrefs’ analysis of about 14 billion pages, 96.55% received no traffic from Google, and the first reason it gives is that nobody was searching for the topic.

Step 7: Choose the Primary Keyword and Write It Into the Brief

For each post, record:

  • One primary keyword: the clearest, most-searched phrasing of the group. It shapes the title and the opening.
  • Three to eight secondary phrases: other wordings and subtopics from the group. They suggest headings and sections, not sentences to force in.
  • The questions to answer: from “People also ask” and your own sales conversations.
  • The format the results favour, and what the current top pages leave out.

All of this belongs in the writer’s brief — see how to create an effective SEO content brief. Then write naturally. Google’s SEO Starter Guide notes that excessively repeating the same words, even in variations, is tiring for readers; use the primary phrase in the title and opening where it fits, and let related phrasing occur as the topic demands.

A Worked Example

A hypothetical garden design business starts with the seed “small garden design”. After expanding, filtering and checking the results, its notes might look like this (your own results will differ by country and over time):

Phrase What page one might show Decision
small garden design ideas Image-heavy inspiration galleries from big home and lifestyle publishers Skip: wrong format and very strong competition
how much does a garden designer cost A few cost guides, several thin directory pages, a forum thread Write: cost guide with real ranges and what drives them (primary)
garden design prices / garden designer fees Overlaps heavily with the cost results Same post as above (secondary)
do I need planning permission for a garden room Official guidance plus builders’ blogs Write later, only if they design garden rooms
garden designer near me Local map pack and directories Not a blog post: belongs to the service page and Google Business Profile
what does a garden designer do Short explainers, mostly generic Write: process post covering stages, timeline, what the client provides

Six phrases, two posts to write now, one for later, one handed to the service page and one dropped. That ratio is normal.

Mistakes That Waste Keyword Research

  • Choosing by volume alone: the biggest number on the list is rarely the best topic for a business.
  • One post per phrasing: three near-identical posts compete with each other and none ranks well.
  • Skipping the search results: a keyword’s intent is visible in ten seconds of looking, and ignoring it is the commonest reason good posts do not rank.
  • Targeting service-page terms with blog posts: if the results are all suppliers, the term belongs to your service page.
  • Stuffing: repeating the phrase does not help; avoiding keyword stuffing shows what natural use looks like.

What to Do This Week

  1. Write down ten seeds from your services, customer problems and repeated questions.
  2. Expand each with autocomplete, “People also ask” and Search Console.
  3. Cut every phrase your buyers would not type.
  4. Group by shared search results and pick the five strongest groups.
  5. Read page one for each, note intent and openings, and brief the best one.

If you want this done properly across a whole site, our keyword research service delivers grouped, intent-checked topics ready to brief, rather than a raw export.

Questions

Questions about this article

Still unsure about something? Talk to a specialist — no script, no sales pitch.

Yes. Google's autocomplete, 'People also ask' boxes and related searches show real phrasing, Search Console shows queries you already appear for, Keyword Planner gives volume ranges, and Google Trends shows seasonality. Paid tools such as Ahrefs or Semrush add estimates and competitor data that make the process faster, but the core decisions come from reading the search results yourself.

There is no universal threshold. For a niche B2B service, a phrase searched a few dozen times a month by genuine buyers can be very valuable, while a consumer topic might need far more to justify the effort. Judge volume alongside relevance to your buyers, the competition on page one, and the size of the group of related phrases.

One primary keyword plus the related phrases that share the same search results and intent. That might be three or thirty wordings of the same need. If two phrases return largely different results, they usually need separate posts. Forcing unrelated keywords into one article tends to weaken it for all of them.

Keyword difficulty is a tool's estimate of how hard it is to reach page one, usually based mainly on the backlinks of pages already ranking. It is a useful first filter but a rough one. Reading the actual results tells you more: forum threads, outdated pages or thin answers on page one suggest an opening whatever the score says.

Use the primary phrase in the title, the opening paragraph and the URL where it reads naturally, and let related phrasing appear in headings and body text as the topic requires. Google warns that excessive repetition is tiring for readers, and it does not reward stuffing, so write for clarity first.

Ready when you are

Ready to work with the pros?

Get a free proposal in 24 hours — scope, timeline and fixed price. No obligation, no lock-in contract.

Trusted by 13,000+ businesses · 30M+ words published · 50+ industries