How to Do Keyword Analysis: A Step-by-Step Guide

Many sites start a keyword list with "whatever comes to mind" and dump it in a spreadsheet. A few months later, half the pages written still don't show up for their target keyword. The problem is rarely the writing; it's that the keywords didn't match intent, competition or the site's real position. A solid keyword analysis exposes that mismatch before you write a word.

This guide walks through it step by step: where to start, what to look at, how to cluster keywords and how to map them to pages.

What is keyword analysis?

Keyword analysis is the process of finding which phrases people use to search a topic on Google, evaluating them, and deciding which page you should create or improve for each. Google keyword analysis usually answers three questions:

1. What are people searching for, and what do they really want? (intent)
2. How much demand and competition is there? (volume and competition)
3. Where do I stand today, and which page could satisfy this? (current position and mapping)

Google's SEO Starter Guide encourages you to think about how users would search for your page and to build content around those phrases. So the analysis starts with understanding people, not tools.

Step 1: Understand search intent

The same keyword can mean different things to different people. Misread the intent and you'll write the wrong page for the right keyword. There are four basic intents:

The most reliable way to judge intent is to search the keyword and read page one. If the results are mostly how-to guides, a product page will struggle there. If they're mostly product pages, a long blog post may not be what people expect. As Google's visual elements gallery shows, images, videos, featured snippets and similar blocks on the results page also hint at the format users want.

Step 2: Start with the data you already have

Rather than searching for keywords from scratch, first look at which queries your site already appears for. Search Console lists them with impressions, clicks, CTR and average position. It is measurement from real users, not an estimate.

The most valuable group: positions 4–20

If you get impressions for a query but sit between positions 4 and 20, Google considers your page relevant, yet most users click results above you. These keywords usually deliver the fastest wins because you don't need a new page: you improve an existing one.

Here's how:

1. In Search Console, select the last 3 months and sort queries by impressions.
2. Filter for average position between 4 and 20.
3. Flag queries with high impressions but low CTR.
4. Note which page appears for each query.

We explained why average position is a blend of devices and countries in our guide to Search Console average position. If you'd rather not filter by hand, the free keyword opportunity finder lists the same opportunities.

The limit of this data

Search Console only reports queries you received impressions for. Google explains that some rarely searched queries are anonymized and never appear in the list. For keywords you don't appear for at all, you need other sources (Step 3).

Step 3: Collect new keyword ideas

Existing data tells you where you are. New ideas tell you where you could be. Sources:

• The results page itself: Google's "Related searches" and "People also ask" sections give real phrasing patterns.
• Customer language: support emails, sales calls, social comments. People often don't use your industry jargon.
• Competitor pages: the subheadings top-ranking sites use.
• Tools: Google Ads Keyword Planner and SEOLOK's free Google rank checker. Look up a keyword and it shows that keyword's monthly search volume; volumes are country-wide. To interpret them, see how to find keyword search volume.

Don't filter yet; build a wide list first. Filtering is the job of the next steps.

Step 4: Evaluate each keyword on four criteria

Once the list exists, ask four things about each keyword.

Volume

An order of magnitude for demand. It is an estimate and country/language-specific. Don't discard low-volume keywords with clear intent wholesale.

Competition

Who is on page one? If big brands, marketplaces or exhaustive guides dominate, the keyword may be hard for a new or small site. Searching the keyword and inspecting the top ten is often more informative than any "difficulty score"; scores vary between tools and are estimates.

Your current position

If you're already at 11–20, the job is easier. If you don't appear at all, there's a longer road. To follow positions over time, see what is Google rank tracking.

Business value

What is the person searching this worth to you? Buyers, brand awareness, or just curious readers? A high-volume keyword with low business value may not deserve priority.

Step 5: Cluster your keywords

Google often treats close variants as the same intent. "How to do keyword analysis", "how to analyse keywords" and "keyword research steps" may not need separate pages. Treating them as one cluster avoids both duplicated content and your own pages competing with each other.

A practical way to cluster:

1. Group keywords by topic.
2. Search each and compare the top results. If two keywords return largely the same pages, they probably belong together.
3. Pick a "head keyword" (highest volume and clearest intent) and "supporting keywords".
4. Put keywords that return different results in separate clusters.

Example: "keyword rank checker" and "google rank checker" express similar intent, "see where my site ranks for a keyword", and are likely served by the same page.

Step 6: Map each cluster to a page

Every cluster should have exactly one target page, for two reasons:

• One page can fully satisfy the cluster's intent.
• When two of your pages target the same cluster they compete, which is called keyword cannibalization, and both can underperform.

When mapping, choose one of these decisions:

• An existing page is relevant and adequate: strengthen its title, subheadings and content for the cluster.
• A page exists but is weak or thin: expand it. See thin content and zombie pages.
• No page exists: plan a new one.
• The cluster's intent doesn't fit your site: leave it.

Step 7: Build a priority table

You'll end up with dozens of clusters and can't do them all at once. A simple scoring table helps you order the work. The table below is an example; the clusters and scores are hypothetical, so adapt the criteria to your own situation.

Scoring suggestion: each criterion 1–3; volume (high = 3), competition (low = 3), current position (11–20 = 3, not visible = 1), business value (high = 3). Highest total goes first.

The numbers in this table are not real data; they only illustrate the logic. What matters is tying decisions to visible criteria instead of gut feeling.

Step 8: Act, measure, repeat

Keyword analysis isn't a one-off. After you make changes:

1. Log the date of each change.
2. Track the target keywords and watch the position trend over several weeks. Ignore single-day moves; why Google rankings change explains the normal fluctuation.
3. Check in Search Console how impressions, clicks and CTR moved.
4. Update the list: double down on what worked, rethink what didn't.

Google's documentation stresses that rankings aren't static and results are dynamic because the web keeps changing, so run the analysis as a regular loop.

How SEOLOK speeds up the analysis

With a read-only Google connection, SEOLOK's Search Console analytics imports the last 16 months of daily data (query, page, country, device, clicks, impressions, CTR, average position) and adds new days every day. Because the data is stored in SEOLOK's database, days older than Google's 16-month window aren't lost going forward from the day you connect. The monthly PDF report lists opportunities: high-impression searches at positions 4–20, together with the extra clicks you'd expect if they reached the top 3.

You can follow your target keywords with Google rank tracking by country or by any of Turkey's 81 provinces, on mobile or desktop. Plan details are on the pricing page.

Common mistakes

• Looking only at volume: without intent and competition, volume misleads.
• Not checking intent: entering with a product page when page one is all guides.
• One page per keyword: several pages for one intent means competing with yourself.
• Ignoring existing data: skipping the position 4–20 opportunities waiting in Search Console and searching from scratch.
• Not measuring: never tracking rankings and clicks after a change.
• A giant list with no priority: 500 keywords are useless if you don't know where to begin.

A quick checklist

• I verified each keyword's intent by reading the results page.
• I pulled the position 4–20 queries from Search Console.
• I gathered new keyword ideas from customer language and tools.
• I evaluated keywords by volume, competition, current position and business value.
• I clustered similar keywords and assigned each cluster one target page.
• I built a priority table.
• I recorded change dates and the keywords I'll track.

Keyword analysis pays off when it answers not "which keyword should I pick?" but "which page should I improve for which intent?" Start with the data you already have and move step by step; the quickest wins are often already within reach.