Google Trends is Google's free tool that shows how much a term is searched for relative to all searches, over a period and in a geographic area you choose.
The thing to know before you open it: it does not return search volumes. It returns an index from 0 to 100 that changes every time you change the date range or the country. If you need a figure to put in a proposal or a business plan, that figure does not come from here.
How to read the 0 to 100 number
Google documents the calculation in two steps in its FAQ about Google Trends data. First, each data point is divided by the total searches for the area and time range it represents, otherwise the most populous regions would always come out on top. Second, the resulting numbers are scaled from 0 to 100 based on the topic's share of all searches.
Three practical consequences follow, and I often see them ignored.
- 100 is the highest point within the range you selected, not an all-time record. Move the start date and the 100 moves.
- Two regions with the same value do not have the same number of searches. Google says so: "Different regions that show the same search interest for a term don't always have the same total search volumes".
- Comparing two terms makes sense; reading one term on its own much less so. The chart for an isolated keyword tells you the shape of demand over time, not its size.
What Google removes from the data before showing it
This is the part few guides mention, and it changes how you interpret a flat chart. Trends starts from a sample of real searches, anonymised, categorised and aggregated. Before anything reaches the chart, though, Google excludes:
- searches made by very few people, which appear as 0 even when real volume is above zero;
- repeated searches by the same person over a short period;
- queries containing apostrophes and other special characters;
- searches generated by Google's own products, including internal searches from AI Mode and AI Overviews.
The third point matters more than it looks in English. "What's the best air conditioner", "men's trainers", "children's party ideas": these are perfectly ordinary ways to phrase a search, and they fall out of the sample. If a query with an apostrophe gives you a flat line, try the version without it before concluding that nobody searches for it.
Then there is something that surprises people who look at isolated spikes: to protect privacy, Google adds statistical noise to the data, small random fluctuations that do not correspond to real behaviour. The noise is always there but shows up mostly on queries with low or no interest. That lone spike in the middle of a niche keyword's chart very probably never happened.
One last technical detail, useful when you compare Trends with other sources: in the Explore charts the time zone is UTC when the granularity is daily, weekly or monthly, whereas for ranges of 7 days or less Google uses your browser's time zone. Two readings of the same curve an hour apart may not match for this reason.
Explore and Trending now are two different tools
The getting started guide on Google Search Central clearly separates the two. Explore lets you query terms and topics of your choosing, with interest over time, regional breakdown, related topics and related queries. Trending now shows what is rising right now, with an approximate search volume, a timeline and related news articles.
Inside Explore there is a distinction that makes a real difference in keyword research: a search term and a topic are not the same thing. A topic aggregates languages, misspellings, variants and related acronyms, so it holds up better when your audience misspells a word or uses both British and American spellings. For long-tail keywords the exact term remains easier to read, because the topic swallows them.
Besides web search you can filter by Google Images, Google News, Google Shopping and YouTube. Web search data goes back to 2004, the other channels to 2008. Google Trends has been online since 11 May 2006, according to Wikipedia.
A UK example: when people really search for "air conditioner"
We queried Google Trends on 26 September 2026 for the term "air conditioner", United Kingdom, past five years, web search (via DataForSEO). Interest hits 100 in the week of 21 to 27 June 2026 and sits at 1 or 2 from October to April. In the UK this is not a gentle seasonal curve: it is a series of spikes that follow heatwaves, and the 2026 peak is so large that earlier summers look small on the same chart. Within the same range, the 2022 peak reaches only 25, 2023 reaches 18, 2024 reaches 10 and 2025 reaches 19.
The detail that matters to anyone planning a content calendar is not the peak itself, it is that the peak moves. In the same series, the yearly high falls in the week of 17 to 23 July in 2022, mid-June in 2023 and 2025 (with a second spike in early July 2025), and late June in 2024 and 2026. Anyone who schedules the campaign on last year's date is likely to arrive late: the working rule is to publish before the curve starts rising, which in this series means early to mid-May, not when it is already at the top.
Trends, Keyword Planner or Search Console: which number to look at
Google's three free sources answer different questions, and most misreadings come from mixing them up.
| Tool | What it measures | Unit | Question it answers |
|---|---|---|---|
| Google Trends | Relative search interest, from a sample of searches | Index from 0 to 100 | When do people search, and is it rising or falling? |
| Keyword Planner (Google Ads) | Search volumes designed for advertisers | Average monthly searches | How big is demand, and how much does it cost to buy? |
| Search Console | Queries that bring impressions and clicks to your site | Clicks, impressions, average position | What do people already find you for, and with what result? |
Google itself puts the difference between the first two in writing: the Google Ads search terms report is designed for advertisers' average monthly volumes, Trends for more granular data closer to real time. In day-to-day work the loop is this: start from the queries you already have in Search Console, check in Trends whether demand is growing or shrinking, then look at the SERP for the search intent before deciding the page format. When a drop in traffic cannot be explained by anything technical, Trends is the first place to look to see whether it was the sector that fell rather than the site.
Since 2025 there is also an API
On 24 July 2025 Daniel Waisberg and Hadas Jacobi announced the Google Trends API in alpha on the Google Search Central blog, with access by application. This is not a developer footnote: it removes two long-standing limits of the web interface.
| trends.google.com | Trends API (alpha) | |
|---|---|---|
| Data scale | Rescaled 0-100 on every request | Consistent scale across requests |
| Terms you can compare | 5 | Dozens |
| Time window | From 2004 | Rolling 1,800 days (about 5 years), up to 2 days ago |
| Aggregation | Automatic, based on the range | Daily, weekly, monthly, yearly |
| Geographic detail | Country, region, metro area | Region and sub-region (ISO 3166-2 standard) |
The consistent scale is the change that counts. On the website, to update a time series you have to download the whole period again each time, because every request redefines the 100. With the API you can pull just the latest month and append it to the data you already had, which makes demand monitoring inside a dashboard possible without the manual work that puts almost everyone off today. The starting point does not change: the API also returns search interest, not absolute numbers.
The comparison that makes Trends reliable
A habit I rarely see, and one that makes Trends far more reliable: never look at a keyword on its own; put a stable reference term next to it. Google suggests this explicitly as a way to gauge the scale of a trend, proposing a comparison with common words such as food names, celebrities or "weather". If your product sits at 4 and "weather" sits at 100, you know what scale you are working at before you even open a paid tool. It is the check that separates a well-read curve from a slide with an arrow pointing up, and the same reasoning we apply when we set up an SEO strategy for a client.
Google Trends FAQs
Yes. It has been free since its launch in 2006 and does not require a Google account to use. Signing in with an account only lets you save comparisons and get notifications about a term.
It means that point is the moment of peak search interest within the range and area you selected. It is not a volume: 50 means half the interest of the peak, not fifty searches.
Because Google only shows data for terms with enough searches: below that threshold the value becomes 0 even if real searches are not zero. If the query contains an apostrophe, try the version without it: Trends filters out queries with apostrophes and other special characters.
No. For an estimate in searches per month you need Google Ads Keyword Planner or a paid tool. Trends tells you how demand moves, not how big it is.
No. In its data FAQ, Google lists among the excluded searches those generated by its own products and services, including internal searches from AI Mode and AI Overviews.
Web search goes back to January 2004. The other channels, namely images, news, Shopping and YouTube, start in 2008. The Trends API in alpha works on a rolling window of 1,800 days instead.
