Skip to content

Article 13 min read

Search Volume Traps: The 4 Contaminations We Measured, and How to Find Yours

by Daniel Sócrates published

Say your clients are companies in Brazil, and you open your keyword tool set to that market. You type “cac,” the acronym for customer acquisition cost. The tool shows 49,500 searches a month. 1

Your clients live by measuring what it costs to win each customer. The number glows on the screen. It is the subject you know best, and you want to write the article today.

Before you write, open the Google results page. At the top sits the e-CAC, the online service center of the Receita Federal, Brazil’s federal tax agency. Right below it comes the CAC registration for collectors, sport shooters, and hunters, a firearms license, on the federal government’s website. 1 2

In the Semrush snapshot for Brazil, none of the top ten positions is about customer acquisition cost. 1 The screen shows what Google understands these people want. They want the tax agency’s service center and the firearms registration.

This piece takes the trap apart from the inside, with the searches SEO Genome measured on its own keyword list. Every measurement was taken on Google in Brazil, where we are based, and every term is in Portuguese. I give each one with its translation and the reason it collides. The method at the end is what carries over to your list, in any language.

What the volume number counts

Picture the entrance of a shopping mall with a door counter, a small device that clicks once for every person who walks in. It counts everyone. It has no idea whether that person went to the movies, the pharmacy, or the bank.

Search volume works like that door counter. Semrush defines search volume as “the average number of times a specific search query is entered on a search engine per month.” The average comes from the past twelve months. 3 The count belongs to the string of letters. What the person had in mind stays out of it.

When a keyword’s number includes people looking for a different topic, this is called contamination.

The tool tries to help with a label. Every keyword in its database gets an intent type: informational, navigational, commercial, or transactional. 4 For “cac,” the label is informational. 1 That is accurate, and it does little for your decision. The label tells you the type of search and says nothing about the topic.

The results page is what tells you the topic

The topic of a search shows up in one place only. It is Google’s results page, which the industry calls the SERP.

Google writes that to return relevant results, it first needs “to establish what you’re looking for.” That is “the intent behind your query.” To get there, Google builds language models that try to decipher how the few words you type match up to the most useful content. 5

The guidelines Google gives its human raters spell out the problem. “Many queries have more than one meaning.” Their example is “apple,” which might be the computer brand or the fruit. 6 The guidelines call the meaning most people intend the dominant interpretation. And they warn that not every query has one. 6

Two rules from the same guidelines matter to us, because we work in Brazil. People in different locales may expect different things from the same query. And the meaning of a query can change over time. 6 Google gives an example from its own search. Someone searching “football” in Chicago will likely see American football. In London, the same search turns up soccer and the Premier League. 5

An honest caveat about the guidelines. They guide the people who rate the quality of results, and the text itself says no single rating directly changes how a page appears. 6 They show what Google wants on the screen. The ranking code stays out of them.

In practice, the SERP is Google’s reading of the query, in one country, on one day. The volume comes from a spreadsheet. The SERP comes from whoever decides what goes on the screen.

The four contaminations that hold up

On September 8, our in-house inventory listed eight traps. 7 On September 14, I reopened each SERP twice. Once in the Semrush report for Brazil, and once live, on Google in Brazil. 1 2 I added two searches that other entries and other pieces had already flagged.

In the September 14 measurement, four terms are contaminated by another topic:

Term (in Portuguese)Searches/monthWhat fills the top ten in SemrushThe term for the right topic
cac49,500the tax agency’s e-CAC (5), firearms registration (4), a youth apprenticeship organization (1)custo de aquisição de clientes (customer acquisition cost), 390
atualização do google (Google update)1,600how to update Chrome, the Google app, and the Play Store (10)atualização algoritmo google (Google algorithm update), 30
dados estruturados (structured data)590databases and data science (7), SEO (3)dados estruturados seo (structured data SEO), 40
geo marketing140geomarketing, marketing by location (10)geo seo, 390

cac. On the live screen, one result in eight is about the sales metric. 2 The AI-generated summary on the page mentions only the other two meanings. 2 The spelled-out term gets 390 searches, less than 1% of the acronym’s volume. 1 Those 390 are the ones that matter if you sell to sales managers.

atualização do google (Google update). This one joined the list after the inventory. It came up in the check for another of our pieces, the one on organic traffic drops. 8 Anyone writing about algorithm updates is aiming at 1,600 searches. The top ten are about updating the browser, the Google app, and the app store. 1 Live, the AI summary walks you through updating Chrome. 2 The phrasing that means the algorithm gets 30 searches a month. 1

dados estruturados (structured data). This one fools even people who check the SERP in a hurry. Google’s documentation on structured data holds first place, in Semrush and live. 1 2 Below it, seven of the ten explain tables, databases, and data science, with IBM, AWS, Oracle, and Databricks. 1 Whoever writes for SEO is competing for three spots. The snapshot matches the one from September 8, position by position. 1 7

geo marketing. In Semrush, all ten are about geomarketing, the strategy that puts sales on a map. 1 Live, two of the nine results already talk about GEO for AI, generative engine optimization. Three of the ads on the page sell a course or a tool for that GEO. 2 The contamination still stands, and it has started to give way.

The one that left the list: aeo

This is the part that teaches the most, because it shows a SERP changing hands.

On August 31, we measured “aeo” in Brazil. Four of the ten positions were about the Authorized Economic Operator, a customs program. 9 Our piece on the boundary between LLMO, GEO, and AEO recorded that finding.

Reopening it on September 14, I have to correct our count. The eighth position went unclassified in August. It is a customs certification school. In the Semrush snapshot, the split is five to five. 1 9

Live, the screen looks different. Seven of the eight web results explain answer engine optimization. One customs brokerage result was left. 2 Related searches still suggest “aeo comex” (AEO foreign trade) and “aeo certificação” (AEO certification). 2

Two findings come out of this. The first confirms Google’s guidelines, since the meaning of a query changes over time. 6 The second is about method. The Semrush snapshot I pulled on September 14 is identical, position by position, to the one from August 31. The live screen shows pages dated July 2026 that are missing from the snapshot. 1 2 And the report I used does not say when the capture was taken.

The verdict for now: on the live screen, “aeo” is no longer contaminated. In the tool’s snapshot, it is still split. Our target remains the spelled-out name, until the next measurement.

Now a trap from a different family. Think of someone who types the name of their bank into Google. What they want is the login screen for their account. This is called a navigational search.

The rater guidelines describe the search of someone trying to reach a specific website or page. The example is “reddit login,” where the person wants to see Reddit’s login page. 6 The same guidelines ask that, for this kind of search, the desired result sit at the top. 6

Two searches on our list work this way:

  • google meu negócio (Google My Business, the Portuguese name of what is now Google Business Profile), 135,000 searches a month. Five of the top ten belong to Google itself. The tool is in first place, support shows up twice, and the login screen twice. The snapshot is identical to the one from September 8. 1 7 In a browser signed in to an account that manages business profiles, the live search put a shortcut to the account’s profile at the top. 2
  • google search console, 60,500 searches a month. Five of the ten are Google’s own: the tool three times, the documentation, and the help center. 1 Live, “google search console login” shows up among the related searches. 2

Here the topic is right. The volume comes from people who want to get into the tool, and few spots are left for anyone who teaches. With contamination, you get the topic wrong. With a navigational search, you get the topic right and the moment wrong. The person wants to use the tool, and you are offering them something to read.

In these two cases, the tool’s label got it right. It marked both terms as informational and navigational at the same time. 1

Territory drift: the same words, a different conversation

The term “autoridade de marca” (brand authority) gets 90 searches a month. The top ten in Semrush and the eight on the live screen are about branding. 1 2 The phrase means what we mean by it. The conversation, though, is about positioning, trust, and events. Search engines come up, when they do, in passing.

The term stays on the list, and it calls for a bridge. The person searching is a business owner thinking about their brand. The text has to cross from their vocabulary to the machine’s. We made that crossing in the piece on brand authority.

List contamination: a clean SERP and a dirty spreadsheet

The last family lives in the spreadsheet. The results page is clean, and the list of variations the tool returns is mixed.

Semrush’s Broad Match returns keywords containing your seed keyword “in any variation and word order.” 10 It matches letters. On our list, that produced four surprises:

  • The SERP for “seo” is entirely about SEO, with Google’s guide in first place. 1 2 The broad match list for the same term is another story. Of its 40 highest-volume terms, 26 revolve around Korean names, like “park seo-joon,” with 18,100 searches. 1
  • The same list includes “seo rosa,” with 1,000 searches a month. Live, it is a bar and restaurant in Campinas, a city in the state of São Paulo. 1 2 Among the questions containing “seo,” one gets 170 searches: “como saber seo nome ta no spc.” 1 That “seo” is a typo for “seu,” which means “your.” The whole search asks how to find out whether your name is on the SPC, a Brazilian credit bureau.
  • Anyone researching “golpe seo” (SEO scam) to write about dishonest agencies gets judo variations, like “golpe seoi nage.” 1 In Portuguese, “golpe” means a scam, and it also means a throw in martial arts. Live, a video of that same judo throw appears on the screen. 2
  • On September 8, we noted “meu site não aparece no google” (my site isn’t showing up on Google) as contaminated by a browser problem. 7 On September 14, the top ten are about visibility in search, led by Google’s help pages. 1 2 The browser problem lives in variations with no volume. One of them is “as imagens dos sites não aparecem no google chrome” (site images don’t show up in Google Chrome). 1 Triage for that symptom is in the piece on when your site isn’t showing up on Google.

The risk here is the total. If you add up the whole list to size the market, you are adding Korean names, a restaurant in Campinas, and a judo throw.

The method, for your list

The method needs the tool you already have and one more column in the spreadsheet, next to volume.

  1. Pick out the terms you plan to turn into pages. The list can come from anywhere, including a keyword gap against your competitors.
  2. Search each term on Google, in your client’s country. Write down the date and where you searched from. Google uses location, search history, and search settings to choose what it shows. 5
  3. Read the top ten and mark each result with one word: topic, tool, other, or neighbor. Topic is about your subject. Tool is a login, dashboard, or support page. Other is a namesake. Neighbor is the same meaning in another discipline.
  4. In the new column, write how many you marked as topic. That number decides the target before volume does.
  5. If there are only a few, look for the spelled-out term, or add a word that pins down the meaning. Measure its volume. That is how “cac” became “custo de aquisição de clientes” for us.
  6. Before you add up the volumes on a list, delete the variations about another topic.
  7. Reopen the SERP before you publish. The one for “aeo” changed between our August measurement and the one on September 14. 2 9

One detail about step 7. Rechecking volume within the same month changes nothing, because Semrush updates its volumes once a month. 3 The numbers on our list on September 14 are identical to those from September 8. 1 7 The SERP changes without warning.

The question you ask from now on

When a keyword shows up with high volume in your spreadsheet, your first question changes. The old question was how many searches it gets. The new one is what the top ten are about.

Volume is still useful, because it measures the size of the door. Only the SERP tells you where the door leads.

Open your spreadsheet now. Take the five highest-volume terms you plan to go after. Create the new column, search each one, and write the count next to the number. The technical fronts that come after this check are in the technical SEO section.

Frequently asked questions

Is the tool’s volume wrong when the SERP is about another topic?

The number may be right. Semrush defines volume as the monthly average of times the query is entered, calculated over twelve months. 3 The mistake is reading the whole number as if all of it belonged to your topic.

Does the tool’s intent label solve the problem?

It helps, without solving it. Semrush sorts every keyword into four intent types. 4 For “google meu negócio” and “google search console,” it marked navigational intent. For “cac,” it marked informational, and the entire SERP is about another topic. 1

How often should I reopen the SERP?

Our rule is to reopen it before declaring the target and again before publishing. Semrush updates volumes once a month. 3 The results page changes on no schedule at all, as “aeo” showed. 2 9

Notes and sources

  1. SEO Genome measurements in Semrush, Brazil database, September 14, 2026. Reports on volume, intent, and difficulty per term, on the top ten organic results, on Broad Match, and on questions. Volumes: cac 49,500; google meu negócio 135,000; google search console 60,500; atualização do google 1,600; aeo 1,000; dados estruturados 590; geo marketing 140; autoridade de marca 90; custo de aquisição de clientes 390; geo seo 390; dados estruturados seo 40; atualização algoritmo google 30; seo rosa 1,000. In the Broad Match list for “seo,” “park seo-joon” appears with 18,100. The organic results report does not state the date on which Semrush captured each SERP.
  2. Live check on Google in Brazil, September 14, 2026, with the language set to Portuguese, the country set to Brazil, and the parameter that requests non-personalized results, in a browser signed in to a Google account, from a single location. It is the screen of one place and one moment. Terms checked: cac, atualização do google, dados estruturados, geo marketing, aeo, google meu negócio, google search console, autoridade de marca, seo, seo rosa, golpe seo, and meu site não aparece no google.
  3. Semrush, “What is Search Volume in Semrush?”, https://www.semrush.com/kb/683-what-is-search-volume-in-semrush. Accessed September 14, 2026. Passages: “Search volume is the average number of times a specific search query is entered on a search engine per month”; the monthly average is “calculated by dividing the total searches for the keyword over the past 12 months by 12”; and “Semrush search volumes are updated on a monthly basis”. The page describes the calculation as third-party data layered over historical browsing data, processed by a machine learning model.
  4. Semrush, “What is Keyword Overview?”, https://www.semrush.com/kb/257-keyword-overview. Accessed September 14, 2026. Passage: “The main types of search intents are: informational, navigational, commercial, and transactional.” The same page says intent was identified automatically for each keyword in the database.
  5. Google Search, “How Does Google Determine Ranking Results,” https://www.google.com/search/howsearchworks/how-search-works/ranking-results/. Accessed September 14, 2026. Passages: “To return relevant results, we first need to establish what you’re looking for”, followed by “the intent behind your query”; “To do this, we build language models to try to decipher how the relatively few words you enter into the search box match up to the most useful content available”; “we use information such as your location, past Search history, and Search settings to determine what is most relevant for you in the moment”; and the example of “football” in Chicago and in London.
  6. Google, “Search Quality Rater Guidelines” (General Guidelines), version of September 11, 2025, https://static.googleusercontent.com/media/guidelines.raterhub.com/en//searchqualityevaluatorguidelines.pdf. Accessed September 14, 2026. Section 0.1: “No single rating can directly impact how a particular webpage, website, or result appears in Google Search”. Section 0.0: “Search results for a specific website or webpage should have that desired result at the top.” Section 12.3: “Users in different locales may have different expectations for the same query.” Section 12.5: “Many queries have more than one meaning”, with the “apple” example and the definition of dominant interpretation, “what most users mean when they type the query. Not all queries have a dominant interpretation.” Section 12.6: “Query Meanings Can Change Over Time”. Section 12.7.3: the intent of a Website query “is to locate a specific website or webpage”, with the example “reddit login”.
  7. SEO Genome, full target inventory (Semrush, Brazil database, September 8 and 9, 2026), with group C of traps in eight entries and the contamination note on “meu site não aparece no google”; SERPs for “google meu negócio” and “autoridade de marca” opened on September 8, 2026; and the SERP for “dados estruturados” opened on the same date. Internal records.
  8. SEO Genome, contamination check for the piece on organic traffic drops, Semrush, Brazil database, September 14, 2026. Internal record.
  9. SEO Genome, SERP competition analysis, Semrush, Brazil database, August 31, 2026. The “aeo” SERP recorded there has the same nine classified positions as the September 14 snapshot, and left the eighth one unclassified. Internal record.
  10. Semrush, “How to Use the Keyword Magic Tool,” https://www.semrush.com/kb/617-keyword-magic-tool-manual. Accessed September 14, 2026. Passage: “Broad Match: Returns keywords containing your seed keyword in any variation and word order, giving you the most results.”

Want to see SEO Genome more often on Google? Add it as a preferred source.

Daniel Sócrates

Founder and CEO of SEO Genome

Author of AI Search Optimization. Writes about what the company applies in client work.

danielsocrates.com.br

Let us look at what the machine understands about your company

A diagnostic conversation, with no slide deck. You bring your website address. We bring what Google and the AI assistants already know about it today.

Talk to us on WhatsApp