How to find free keyword search volume

Free search volume numbers are lies of varying sizes. When you type a query into a free tool, you are not getting a precise count of human intent. You are getting a bucketed estimate, a scaled proxy, or a scraped API response from a tool that is guessing. Making product decisions on bad data costs months of engineering time. If you want to know how many people are looking for a solution, you need to understand exactly where free numbers come from and the error bars on each source.

The architecture of free search volume

Every free keyword tool on the internet gets its data from one of two places: the Google Ads API or a clickstream panel. Google Ads data is designed for advertisers, not product builders. It groups similar keywords together. If you search for "build an app" and "make an app", Google combines the volume. This artificial inflation makes markets look bigger than they are.

Clickstream data comes from browser extensions that track user behavior. Companies buy this data, anonymize it, and extrapolate it to estimate search volume. If a data provider tracks two million users in the United States, and 400 of them search for a specific keyword, the provider multiplies that number by a demographic constant to estimate the total volume. The error bar here is heavily biased by the demographics of people who install browser extensions. Tech workers and gamers are vastly overrepresented. If you are researching demand signals for developer tools, clickstream data will overestimate the volume. If you are researching demand signals for local plumbing software, clickstream data will severely underestimate it.

Relying on a single free tool guarantees a blind spot. You must triangulate demand using raw sources that do not hide their methodology behind a proprietary metric.

Google Search Console and the honeypot method

If you own a domain, Google Search Console is the only free tool that provides exact impression data. It does not bucket numbers into massive ranges. If your page ranks in the top 100 results and a user searches for that term, Google logs an impression. The error bar for your specific ranking position is zero.

To use this for new keyword validation, you must build a honeypot page. This is a minimum viable landing page targeting the exact keyword you want to measure. You do not need a product. You just need a headline, a description of the solution, and an email capture form. Publish the page, submit it to Google via Search Console, and wait three weeks.

When the data populates, look at the impressions for your target keyword. You will likely rank somewhere between position 40 and 80. If a keyword gets 300 impressions in a week at position 50, you can apply basic click-through rate math to estimate the total market. Pages on the fifth page of Google only capture a fraction of total searchers. A baseline rule is that position 50 sees roughly 5 percent of total search volume. If you have 300 impressions at a 5 percent visibility rate, the total weekly search volume is roughly 6,000. Multiply that by four to get a monthly search volume of 24,000. This is real data based on actual users seeing your link. It is infinitely more valuable than a generic bucket from a free tool.

Google Ads Keyword Planner forecast trick

Keyword Planner is the source data for almost every free SEO tool on the market. When you use it without an active ad spend, Google gives you useless ranges like 10K to 100K. A keyword with 11,000 searches and a keyword with 99,000 searches look identical. You cannot build a business model on a 900 percent variance.

You can tighten this range without spending a dollar by using the forecast feature. Create a Google Ads account and skip the campaign creation. Open the Keyword Planner and select the option to get search volume and forecasts. Enter your target keyword and add it to a plan. Navigate to the Forecast tab.

By default, Google estimates how many clicks you will get based on a low bid. Change your maximum cost per click to an absurdly high number like $500. This tells Google to project your metrics as if you were winning every single auction. Look at the Impressions metric for the next 30 days.

This forecast number is much closer to the actual search volume than the generic bucket. If your bucket was 10K to 100K, the forecast might show 34,500 impressions. The error margin shrinks from 900 percent down to roughly 20 percent. The forecast accounts for seasonality and device trends, making it the most accurate free number you can get without ranking a page yourself.

Autocomplete depth as a proxy for width

Google Autocomplete does not give you a number. It gives you a map of demand signals. When you type a query, Google suggests the most frequent completions based on real user activity. You can measure search volume by testing the depth of the alphabet.

Type your core keyword followed by the letter A. Count how many highly specific suggestions appear. Repeat this for every letter of the alphabet. This is a manual process, but it reveals the exact shape of the market. Give yourself one point for every relevant suggestion. If a keyword has unique autocomplete suggestions for 20 different letters, it has massive search volume. A score over 50 means you have found a deep market. A score under 10 means you are in a niche too small to support a standalone product.

For example, typing "invoice generator a" might yield "invoice generator app", "invoice generator australia", and "invoice generator api". If you type "how to build a custom CRM in" and Google runs out of suggestions by the letter C, the volume is microscopic.

The error bar here is difficult to quantify in raw numbers, but the directional accuracy is flawless. Autocomplete cannot be gamed by bots in the same way search volume metrics can. It shows you exactly what features users are trying to find. If you are building an iOS app, Unthinkable surfaces these exact demand signals directly from web data. Otherwise, you will need to manually map out the autocomplete branches to see where the market is pulling.

YouTube view velocity

YouTube is the second largest search engine in the world. Unlike Google, it exposes exact view counts on every piece of content. You can use the view velocity of recent videos to estimate search volume for a topic.

Search for your target keyword on YouTube. Filter the results by videos uploaded in the last 30 days. Look at the top five results. Divide the total views by the number of days since the video was published. This gives you the daily view velocity.

If a new video about "how to scrape real estate data" gets 14,000 views in seven days, the daily velocity is 2,000 views. This proves immediate, active demand. The error bar here depends on the creator's subscriber count. A channel with one million subscribers will skew the data through notification feeds. To get a clean number, look for videos from small channels that are ranking purely based on search intent. If a channel with 100 subscribers gets 5,000 views in a week, that traffic is coming from search. This method is incredibly reliable for consumer software, tutorials, and developer tools. It acts as a perfect proxy for Google search volume because user intent is identical across both platforms.

Ranking the sources by decision weight

Not all free data should carry the same weight when deciding what to build. You must rank these sources by their reliability and base your engineering decisions on the heaviest signals.

Rank one is Google Search Console. The honeypot method takes two hours to build and gives you exact impression data. This is the only metric that should carry a heavy financial or engineering decision. The data is unassailable because it comes from your own server logs and Google's direct reporting.

Rank two is the Keyword Planner forecast trick. It is the best way to get a specific number before writing a line of code. Use this to filter out ideas that have absolutely zero demand. If the forecast shows fewer than 100 impressions a month, kill the idea.

Rank three is YouTube view velocity. This proves that people are actively seeking solutions right now. It is a live pulse check on the market. If the view velocity is high, the search volume is high.

Rank four is autocomplete depth. This does not give you a total market size, but it gives you the feature roadmap. It tells you exactly what modifiers people care about.

Rank five is any free third-party SEO tool that gives you a specific number without asking for a credit card. These tools are either caching old Keyword Planner data or multiplying clickstream data by arbitrary constants. Their error bars are unknown and their methodology is hidden. Do not base product decisions on a black box. Build your validation around the raw sources and you will stop building products for empty rooms.