In short: Google Trends is a free Google tool that shows the relative popularity of search terms over time, by location and by category. It does not report how many people searched for something. It divides the searches for your term by all searches in that place and period, then rescales the result to a 0 to 100 index where 100 is the single highest point in the range you selected. Used correctly it answers four questions no keyword tool answers as well: when demand peaks, where it is concentrated, which related terms are accelerating, and whether a term is growing or dying. Used carelessly it produces confident nonsense, because the same term can display as 100 in one chart and 4 in another without demand having changed at all.
What Google Trends is, and what it is not
Google Trends is a public tool at trends.google.com that reports search interest in a term or topic. It has been running since 2006, it is free, it requires no account, and it covers Google Search plus several other Google properties.
The single most important thing to understand is what the tool measures. Google Trends does not publish search counts. Every number you see is a normalised index. Google divides the searches for your term by the total searches in the same geography and time window, so that a growing internet does not make every term look like it is rising, and then it scales the resulting proportions so the largest one becomes 100.
That design makes Google Trends excellent for questions of shape and timing, and unsuitable for questions of size. It will tell you that interest in your category triples in the second week of November. It will not tell you whether that means four hundred searches or four million. For absolute volume you need Keyword Planner or your own Search Console data, which is why the tools belong together rather than in competition. If you are building a full term list, start with our guide to keyword research and use Trends as the timing layer on top of it.
What the 0 to 100 number actually means
Read this section slowly, because almost every bad decision made with Google Trends comes from skipping it.
The value 100 does not mean "very popular". It means "this is the highest point in the data I am currently showing you". Change the date range, the country or the comparison set, and Google recalculates the whole chart from scratch against a new peak. Nothing about real demand changed. Only the reference point did.
A worked example. Suppose you chart the term ski jackets in the United States over the last five years. December of each year will sit near 100 and July will sit near 5. Now change the range to "past 30 days" in July. The same term now peaks at 100 again, because the highest July day is the highest point in the new window. The chart looks identical in shape and means something completely different. A reader who screenshots the second chart and reports "ski jacket demand is at an all time high" has not misread the data, they have misread the axis.
Why two different terms can both show 100
When you compare several terms in one chart, Google normalises them against each other, and 100 belongs to whichever term had the single highest point. This is the only mode in which the numbers are directly comparable between terms. If you look up two terms in two separate searches and put the charts side by side, the two 100s are unrelated and comparing them is meaningless. If you want to compare terms, you must put them in the same chart.
The filtering you do not see
Google documents several things it strips out before you ever see a chart. Searches from very low volume terms are shown as 0 rather than as a small number. Repeated searches by the same person over a short period are counted once. Queries containing apostrophes and similar special characters are filtered. Google also uses sampling rather than the full search log, because it handles billions of searches a day. And notably for 2026, Google states that queries generated inside its own AI products, including AI Mode and AI Overviews, are excluded from Trends data.
That last point matters more each quarter. As a growing share of information seeking moves into AI surfaces, Google Trends increasingly reflects classic typed search rather than total demand for a subject.
The surfaces: Explore, Trending Now, and the API
Most people only ever use one screen. There are three worth knowing.
Explore is the analysis screen. You enter a term or topic and get interest over time, interest by region, and related topics and queries. You can add up to five terms to compare in a single chart, and you can drill geography down to city level. If you leave the search box empty, Explore shows you what is trending broadly rather than a specific term.
Trending Now is the real time screen. It shows what is spiking on Google right now, with an approximate search volume, a trending timeline and the related news articles that explain why. This is the surface for newsrooms, social teams and anyone doing reactive content. It is not the surface for planning a quarter.
The Trends API is the newest addition. Google announced the Google Trends API in alpha on the Search Central blog in July 2025. The important difference from the website is that the API is designed to return consistently scaled data, so that results from separate requests can be joined and compared rather than each being rescaled to its own 0 to 100. Access is application gated, so treat it as a roadmap item rather than something you can wire up this afternoon.
How to use Google Trends, step by step
1. Choose a topic rather than a search term where you can
As you type, Google offers two kinds of entry. A search term matches the literal words. A topic is an entity, and it aggregates across languages, spellings, abbreviations and variations. For a brand, a product category or a public figure, the topic gives a far more stable picture. For a specific phrase you intend to bid on or write a page for, the search term is the correct choice because that is what you are actually targeting.
2. Compare up to five terms in one chart
Comparison is where Trends earns its keep. Put your term next to a competitor term, next to the generic category term, or next to the phrasing you are unsure about. Two phrasings of the same product often differ by an order of magnitude, and that answer costs you thirty seconds here versus a month of live campaign data.
3. Choose the time range on purpose
The default range is usually too short for planning. Use the past five years to judge whether a category is structurally growing or declining. Use twelve months to find the seasonal shape. Use ninety days for campaign level movement, and seven days or less only when you are reacting to something. Bear in mind that Google applies UTC for ranges of thirty days or more and the local time zone for ranges of seven days or less, which is enough to shift a spike by a day near the boundary.
4. Narrow the geography before you conclude anything
Worldwide data is almost always the wrong default for a business. Set the country first, then look at the region and city breakdown underneath the chart. The regional view is frequently the most immediately actionable panel in the whole tool, because it tells you where to concentrate budget rather than merely what to say.
5. Use the category and property filters
Ambiguous terms need a category filter or the data is worthless. "Jaguar" in the automotive category and "jaguar" with no filter are different datasets. The property filter is just as useful in the other direction: switching from Web Search to YouTube Search tells you whether a subject is something people read about or something they watch, and that single switch should decide whether your next asset is an article or a video. News, Image and Shopping searches are available on the same control.
6. Read Related topics and Related queries properly
Below the charts sit two panels, each with a Top and a Rising view. Top shows what is consistently searched alongside your term. Rising is the more valuable of the two, because it shows what is accelerating. Terms whose growth is so large that a percentage would not be meaningful are labelled Breakout, which usually indicates a query that is genuinely new rather than one that merely grew.
Rising queries are the closest thing Google gives you to a free early warning system. They are also the cheapest source of content angles and of negative keywords, because they show you the adjacent intents that your term is being pulled into.
Google Trends vs Keyword Planner vs Search Console
These three free tools are constantly confused with each other. They measure different things and the confusion produces bad reporting.
| Google Trends | Keyword Planner | Search Console | |
|---|---|---|---|
| What it measures | Relative interest, indexed 0 to 100 | Estimated average monthly searches | Your own impressions, clicks and position |
| Absolute volume | No | Yes, bucketed and approximate | Not volume, but real impressions |
| Historical depth | 2004 onwards | Roughly the last 12 to 48 months | 16 months |
| Best for | Timing, seasonality, momentum, geography | Sizing demand and forecasting spend | Finding pages already close to page one |
| Blind spot | No volume, rescales constantly | Groups distinct terms together | Only shows terms you already rank for |
| Account needed | No | Yes, a Google Ads account | Yes, verified site ownership |
The practical rule is that Trends decides when and where, Keyword Planner decides how much, and Search Console decides what to fix first. Any one of them used alone will mislead you.
Seven ways advertisers actually use Google Trends
We manage paid media for retail, ecommerce and B2B advertisers, and these are the uses that survive contact with a real budget.
- Timing the budget curve. Pull twelve months of the category term, find the weeks where interest climbs, and move budget forward so the campaign is already learning when demand arrives rather than starting on the peak.
- Deciding launch dates. A launch into a falling curve looks like a creative failure in the reporting even when the creative was fine. Check the curve before you commit the date.
- Building negative keyword lists. Rising related queries expose the adjacent intents your broad match is drifting into, often weeks before those terms show up in your own search terms report.
- Choosing between two phrasings. Product naming, ad copy and page titles all benefit from a thirty second comparison chart rather than an argument.
- Allocating geographically. The region and city panel is a defensible basis for splitting budget or for deciding where a local campaign is worth running at all.
- Tracking brand demand against competitors. Branded search interest is the cleanest available proxy for whether upper funnel work is doing anything, and it is one of the few competitive metrics that is genuinely public.
- Choosing the format. The YouTube Search property answers whether the audience for a subject wants to watch rather than read, which decides where the production budget goes.
If you are turning these into live campaigns, our guides to campaign management and to Performance Max cover the execution side.
Six ways Google Trends will mislead you
Every one of these has produced a wrong decision in a real meeting.
- Reading 100 as a volume. It is a position on a rescaled axis. Two charts with a 100 in them are not comparable unless they are the same chart.
- Comparing terms in separate searches. The only valid comparison is inside one chart with the terms added together.
- Trusting low volume curves. Thin data is noisy and Google reports very low volume terms as 0. A jagged chart on a niche term is usually sampling noise, not a trend.
- Forgetting the category filter. Ambiguous terms silently blend unrelated audiences.
- Mistaking a news spike for demand. A single event can move a curve for a week without changing commercial intent at all. Cross check against the related news in Trending Now.
- Assuming Trends covers all search behaviour. It does not include queries made inside Google's AI products, and it never included what people ask other assistants.
A twenty minute workflow you can run this week
Open Explore and set your country. Enter your main category term as a topic, set the range to five years, and note whether the line is structurally rising, flat or falling. Switch to twelve months and write down the two or three weeks where the curve climbs fastest. Add up to four competing or adjacent terms to the same chart and see which phrasing genuinely leads. Scroll to the regional panel and list the top five regions. Open Related queries, switch to Rising, and copy every Breakout term into two lists: one of content ideas and one of candidate negatives. Finally switch the property filter to YouTube Search and check whether the shape changes.
Twenty minutes of that produces a seasonality calendar, a geographic split, a content shortlist and a negative list. That is a better return than most paid tools deliver in a week.
Sources
The product behaviour described here is drawn from Google's own documentation: the Search Central guide Get started with Google Trends, the Google Trends Help article on how Trends data is adjusted and normalised, and the Search Central blog announcement introducing the Google Trends API in alpha, published in July 2025. Where this article goes beyond Google's documentation it is describing how our own team applies the tool to client campaigns, and it is marked as such in the text.
Frequently asked questions
How do I use Google Trends as a beginner?
Go to trends.google.com, type a term, and change two settings before you read anything: set the country to your market and set the time range to the past twelve months. Then look at three things in order, the shape of the line, the regional breakdown underneath it, and the Rising related queries. That is ninety percent of the value of the tool.
How does Google Trends work?
It takes a sample of Google searches, divides the searches for your term by all searches in the same place and period so that overall search growth does not distort the picture, and then rescales the result to an index where the highest point in your selected range equals 100.
How do I access Google Trends?
It is a free public website at trends.google.com. No Google account and no advertising spend is required. There is also an application gated API, announced in alpha in July 2025, for programmatic access.
What can you do on Google Trends?
Compare up to five terms in one chart, see interest over time from 2004 onwards, break interest down by region and city, filter by category and by Google property including YouTube, read Top and Rising related queries and topics, and see what is spiking right now in Trending Now.
Is Google Trends free?
Yes. The website is free and requires no account. There is no paid tier.
Does Google Trends show real search volume?
No. It only ever shows relative interest on a 0 to 100 index. Trending Now shows an approximate search volume for currently trending items, but the main Explore charts do not. For volume estimates use Google Keyword Planner.
How is Google Trends different from Google Keyword Planner?
Keyword Planner estimates how many searches a term gets each month and needs a Google Ads account. Google Trends shows no volume at all, but shows timing, momentum, seasonality and geography far better, goes back to 2004, and needs no account. They answer different questions and are usually used together.
Can I use Google Trends for YouTube?
Yes. In Explore, change the property filter from Web Search to YouTube Search. The curve for the same term is often noticeably different, which is a fast way to decide whether a subject deserves an article or a video.
