How to Do Keyword Mapping: A Step by Step Guide to Assigning Keywords to URLs

A keyword list on its own does very little. Because you still do not know which page owns which search, the list just sits in a spreadsheet instead. Keyword mapping closes that gap. In this guide I walk you through the method I have used on client sites since 2012: inventory, intent clusters, SERP comparison, the mapping sheet and cannibalization fixes.

What is keyword mapping?

Keyword mapping is the planning step where you assign every keyword cluster from your research to exactly one URL on your site. The goal is simple: each page serves one clear search intent, no two pages compete for the same query, and keywords without a matching page become visible gaps.

The result is a table. On the left you have keyword clusters; on the right, the address that answers each cluster. That same table also doubles as your content calendar, your internal linking plan and your list of technical fixes.

For me, the real value of a map is faster decisions. When a new article idea comes up, you check the map first. If the keyword already belongs to a page, you improve that page instead of publishing a new one. If it has no owner yet, the new page starts with a defined name, title and purpose. As a result, the site stays organized as it grows, and every page has one job.

Why does keyword mapping affect rankings?

For most queries, Google shows one page from your site. If you cover the same topic halfway on three different pages, you cannot control which one Google picks. A map, however, puts that choice back in your hands.

The second effect concerns signals. Google's documentation on consolidating duplicate URLs explains that a canonical preference helps search engines consolidate signals, such as links, into a single preferred URL. In other words, when fewer pages do the same job, links and internal links add up on one address.

Then there is a third effect, on the visitor side. Someone searching for a price who lands on a beginner article will leave quickly. Match the right intent with the right page, and both click through rate and conversions tend to improve. That said, let me be honest here. In practice, a map alone does not create rankings. Weak content in a tidy spreadsheet is still weak content. The map only points your effort at the right page.

How is keyword mapping different from keyword research?

Research answers the question "what do people search for?". Mapping answers a different one: "which page on my site will answer this search?". In other words, they are consecutive steps. Skip one and the other stays incomplete.

I covered the research side, meaning how to find keywords close to a purchase, in my article on how to find keywords that drive sales. I will not repeat it here. This guide assumes you already have a list.

The most common mistake I see goes like this. A business owner exports five hundred keywords from a tool and orders one blog post per keyword. As a result, the outcome is dozens of near identical posts, and none of them ranks well. Mapping stops that waste early, because it groups keywords into clusters first and only then connects clusters to pages.

What data should you collect before you start?

A good map rests on three sources: the current state of your site, search data and business priorities. Otherwise, if one of them is missing, the map is either unrealistic or disconnected from revenue.

• A list of all indexable URLs, taken from your XML sitemap and a crawl.
• Query and page data from the Search Console Performance report: clicks, impressions and average position.
• Your research list with keyword, estimated volume and an intent note.
• The words your sales team and customers actually use, from quote forms, call notes and email questions.
• A short list of services or products with high margin or high business priority.

One detail about Search Console is worth knowing. Google's Performance report help page explains that you can group data by query, page, country, device, search appearance and date. The report reaches back up to 16 months. So for seasonal businesses, you can compare against the same period last year.

Step 1: How do you build an inventory of existing pages?

You build the map on your real site, not on an empty sheet. First, list every indexable URL in one table. Each row gets the address, page type, current title tag, H1 and the top three queries the page receives today.

Next, you pull those queries from Search Console. In the Performance report, open the Pages tab, click a URL, then switch to the Queries tab. Now you see which searches that page appears for. This step takes time, but it also teaches a lot. Quite often you discover that a page shows up for queries you never intended.

During the inventory, add a rough label to each page:

• Keep: the page serves the right intent and earns traffic.
• Improve: the intent is right, but the content is thin or outdated.
• Merge: the page does the same job as another page.
• Remove: it brings neither traffic nor leads and has no useful equivalent.

That said, these labels are not final. You will review them once the map is complete. Still, they give you an honest picture of site health from day one.

Step 2: How do you group keywords by search intent?

You cannot assign keywords one by one. First, you sort them into clusters by meaning and intent. A cluster contains different phrases that look for the same answer. For example, "small business website cost", "website design price" and "how much does a company website cost" belong together.

In practice, four rough intent classes are enough to start:

1. Informational: learning searches such as "what is" or "how to".
2. Comparison: pre decision searches such as "best", "vs" or "alternatives".
3. Transactional: decision searches such as "price", "get a quote" or "buy".
4. Navigational: searches for a brand name, "contact" or a specific location.

Still, watch out for one trap. The same root term can split into different intents. A query containing "SEO" might come from a student or from a company ready to hire. So you build clusters around what the searcher wants to do at that moment, not around the spelling of the word. When in doubt, the SERP check in the next step settles it.

Also keep local and language modifiers in mind. Searches with "London", a district name or "near me" often need their own cluster, since the searcher expects service in a specific area. For a single location business, linking these phrases to the main service page is usually enough. If you have several branches, each location needs its own page and its own row. The same logic applies to languages: English and German searches are separate clusters with separate pages.

Step 3: How do you use the SERP to decide which keywords share a page?

The method I trust most is putting two result pages side by side. If Google shows largely the same URLs for two queries, one page can target both. If the results differ clearly, however, you need separate pages.

Here is how I work in practice. I search both queries in a private window and compare the top ten organic results. A high number of shared URLs means one cluster; a low number means two. I do not give a hard threshold, because this is a judgment based on field experience, not a guarantee. Still, when roughly half or more of the results overlap, I rarely find a separate page necessary.

Additionally, the SERP tells you the page type. If the top results are mostly category pages, a blog post will struggle to break in. If guides dominate, a service page will struggle instead. Therefore your mapping sheet needs a page type column next to the URL. This check is slow, so apply it to high value clusters first.

Step 4: How do you assign each cluster to one URL?

You follow one rule during assignment: one cluster, one primary URL. Each cluster has a primary keyword, and the rest attach to the same page as secondary keywords. The primary keyword usually carries the clearest intent and the highest business value. Specifically, it does not have to be the one with the highest volume.

Next, you decide in a fixed order. First, check whether an existing page already answers the cluster well. If it does, assign the cluster there. If not, consider whether expanding a page that partly covers the topic makes sense. Only if neither works do you note a need for a new page.

After that comes my reason for preferring existing pages, and it is simple. An older page usually has links and history behind it. Launching a new page and leaving the old one alone wastes that history. On top of that, you end up with two pages on one topic, which means you create cannibalization yourself. So in keyword mapping, a new page should be your last option.

Which columns should a keyword mapping sheet include?

First, I recommend keeping the map in a shared spreadsheet. Your team can open it, filter it and comment on it. The columns below form the core of my own template. Add columns if you need them, but I would not remove these.

Note that the sample values only show the format; they are not data from a real site. What matters is that every row points to exactly one URL, and no primary keyword appears in two rows.

What is keyword cannibalization and how do you spot it?

Cannibalization happens when two or more of your pages compete for queries with the same intent. The typical symptom: in Search Console, different URLs appear for one query over time, and the ranking never settles.

In practice, the most practical detection tool is the Search Console Performance report. Type your target keyword into the query filter, then switch to the Pages tab. If several pages get meaningful impressions for that query, write them down. Google's deep dive on performance data notes that the interface shows at most 1,000 rows. For large sites, the API or bulk export helps.

A word of caution, though. Two pages showing for one query is not always a problem. A main service page and a supporting guide can appear side by side, and that can signal topical strength. The real problem is two pages giving the same answer to the same intent while the ranking keeps switching. Decide based on intent, not on the count alone.

What are your options for fixing cannibalization?

Once you find cannibalization, you have four basic options. Which one fits depends on how much the pages overlap and what each has built up.

• Merge and 301 redirect: if two pages do the same job, combine the strongest content on one page and redirect the other.
• Canonical tag: if both pages need to exist for users, as with filter variations, mark the preferred URL with rel="canonical".
• Differentiate: if the pages can serve different intents, refocus one of them with a new title, new content and new internal links.
• Remove: if a page has no value and no equivalent, return a 404 or 410.

Google's documentation lists redirects and canonical links as strong signals and sitemaps as a weak signal. Therefore simply dropping a URL from your sitemap will not solve the issue. After setting up a redirect, I suggest checking the chain and status code with a redirect checker. For bulk URL changes, the checklist in how to protect SEO during a website redesign will help too.

What should you do with keywords that have no page?

Once the map is filled in, some clusters have no page next to them. These gaps are the raw material for your content plan. However, not every empty row calls for a new page; you evaluate the cluster first.

So I ask three questions. First, could the person searching this become my customer? Second, can I say something genuinely useful and original on the topic? Third, would this cluster work better as a subsection of an existing page? Only when all three answers point the same way does a new page make sense.

When you create a new page, add its row to the map right away and mark it as new. That way, a second writer does not start another page for the same cluster without knowing it. In teams where several people produce content, this simple habit prevents most future cannibalization. In my own projects, in fact, no content brief starts before its row exists in the map.

How do service, category and blog pages differ on the map?

Above all, page type is the backbone of the map. Transactional clusters go to service, product or category pages. Informational clusters go to blog posts and guides. Comparison intent sits in between and often needs a dedicated comparison page or a thorough guide.

On a typical service site, the picture looks like this. The main service page targets a commercial cluster such as "seo consulting". Blog posts target informational clusters such as "how long does seo take" and link back to the service page. For instance, on my own site the SEO consulting page carries the commercial cluster, while blog posts support it.

On e-commerce and large sites, the category layer comes in. Broad clusters go to parent categories, narrower ones to subcategories, and very specific searches to product pages. I explained how to build that hierarchy in category structure for large websites. Your map and your category tree should mirror each other.

How does keyword mapping connect with internal linking?

The map says which page owns which cluster. Internal links then communicate that decision to Google and to users. You link to the assigned page from other pages that mention the topic, using descriptive anchor text. In effect, the site itself says "this is the main page for this topic".

In practice, the linking pages column in the sheet plans this work. Whenever a new post goes live, I check whether it links at least once to the relevant main page on the map. I also work the other direction and add links from the main page to supporting posts, because readers who want detail should find it easily.

Finally, one anchor text rule matters. I never use the same primary keyword as anchor text for two different pages. If you link to the service page with "seo consulting" today and to a blog post with the same anchor tomorrow, you blur your own map. Anchor text is part of the map too.

Where does the map show up on the page itself?

The mapping decision becomes visible in a few places: the title tag, the H1, the meta description, the first paragraph, subheadings and the URL. The primary keyword appears naturally in most of these. Secondary keywords find their place in subheadings and in the flow of the text.

That said, do not overdo it here. Google's spam policies explicitly list keyword stuffing as a violation. In addition, Google's SEO Starter Guide states that keywords in the domain name or URL path alone have hardly any effect beyond appearing in breadcrumbs. So there is no need to rename old URLs just to squeeze keywords into them.

When you write titles and descriptions, a Google SERP preview shows whether they get truncated. To see whether you repeat a phrase too often, run the text through a keyword density checker. Still, the aim is not to hit a percentage. The check simply catches repetition you did not notice.

Is a keyword map useful for Google Ads too?

Yes, and directly so. The match between the keywords in an ad group and its landing page affects user experience and ad relevance. Likewise, the map you build for SEO tells you which page each ad group should point to.

In my own process, every transactional cluster is a candidate ad group, and the target URL in the sheet becomes its landing page. So someone searching "website design price" reaches the pricing page, not the homepage. When this match breaks, the most common result I see is ads that earn clicks but no leads.

Moreover, both channels feed each other. Search terms that convert well in ads show which clusters deserve organic priority. Where you already rank solidly in organic results, you might shift ad budget elsewhere. I build that balance on the same map in my Google Ads management work.

Does AI search change keyword mapping?

It does not change the foundation, but it widens each cluster. In AI supported search experiences, people ask longer questions in a more conversational tone. Also, many of these questions turn out to be sub questions of clusters you already have.

For the map, then, every main page should also answer the side questions of its cluster. For example, a service page should not stop at "what we offer". It should also explain duration, scope and pricing logic in clear sentences. Keeping these side questions on the main page, instead of splitting them into separate pages, prevents cannibalization and makes the page more complete.

I covered how to prepare content for these experiences in how to write content for AI Overviews. In short, the map stays the same. What changes is the depth of the content behind each row.

What are the most common keyword mapping mistakes?

Over the years, I have seen the same mistakes repeat across the sites I audit. Here are the ones I run into most often:

1. Creating one page per keyword and producing posts that read like copies of each other.
2. Building clusters around spelling instead of intent.
3. Creating the map once and never opening it again.
4. Looking only at volume and ignoring low volume clusters with high business value.
5. Targeting every commercial keyword with the homepage.
6. Launching a new page and leaving the old one live without a redirect.
7. Using the same anchor text for different pages.

What these mistakes share is a view of the map as a document you fill in once, rather than a decision tool. In reality, however, the map is a living document. When the content team, the developer and the ads manager look at the same sheet, mistakes surface much earlier. Keep it somewhere everyone can reach.

How often should you update a keyword map?

My suggestion: review the map every three months, and update it immediately whenever a new page goes live. This rhythm is a starting point from field experience, not a guarantee. Adjust it to the pace of your industry.

During the quarterly review, go back to Search Console and check which queries each main URL receives. If a different page appears where the map expects another, make a note. Then add newly emerging queries to the right cluster. Finally, reorder the priority column.

For seasonal businesses, comparing against the same period last year makes more sense, and the 16 month history of the Performance report allows exactly that. If you plan a major structural change, such as merging categories or moving domains, update the map before the change. After all, your redirect plan comes straight out of this sheet.

What does a simple keyword mapping workflow look like for a small site?

A corporate site with twenty or thirty pages does not need complex tools. A spreadsheet, Search Console and some patience will do. The workflow below is the lean version I recommend to small businesses.

1. List every page and write a one sentence purpose for each.
2. Export the last three months of queries from Search Console.
3. Sort the queries into five to ten clusters.
4. Assign each cluster to one page, and note clusters without a page.
5. Merge or refocus overlapping pages.
6. Align titles, H1s and internal links with the map.

You can finish this in a weekend. Seeing results, however, takes a few months, since SEO changes need time to show. Once results start coming in, you expand the map. That way you start small, build the habit, and keep the site organized as it grows.

How do you know the map is working?

Put simply, a map is a tool. You measure its success by changes in search results, not by how neat the sheet looks. I watch three indicators, separately for each main URL.

• Correct page match: is the page Search Console shows for the target query the one you assigned?
• Cluster level impressions and clicks: is the total across all phrases in the cluster growing?
• Conversions: do transactional pages bring forms, calls or sales?

I pay special attention to the first one, because it changes even before rankings move. A few weeks after a merge or refocus, you notice that only one page now appears for the query. That is the first sign Google has accepted your decision.

Cluster level measurement likewise prevents misleading single keyword tracking. One phrase can drop in position while total clicks for the cluster rise. If you only watch one keyword, you draw the wrong conclusion. So I recommend building your reporting around the clusters on the map. Then the table you show your manager and the table you work with speak the same language.

Where should you start with keyword mapping?

First, start with your three most valuable services or product groups. Build the inventory for those areas, form the clusters and assign each cluster to one page. Trying to map the whole site at once usually ends with a half finished project.

Once the map is ready, you have three concrete outputs: a list of pages to improve, a list of new pages to create and a list of pages to merge. These three lists define what your SEO work focuses on in the coming months. When you set the order, put business value first, not volume.

If you would like to run this process together on your site, reach out through my contact page. I work directly with clients, without intermediaries, so we build the map in the same sheet and I explain the reason behind every decision. Whichever path you choose, stopping content production without a map is one of the most efficient moves you can make.