How to Know What Users Searched Before Clicking Your Website

Your website analytics won't tell you — search engines stopped passing the search query to websites years ago. The search engines themselves will, though. Google Search Console shows the queries that brought people to each page. Bing Webmaster Tools does the same for Bing. The Google Ads search terms report covers paid clicks, and your own site search shows what people look for once they've arrived.

What none of them does is tie a specific query to a specific visitor. Knowing exactly where that line sits is what turns this data from a curiosity into something you can act on.

Key takeaways

• Your analytics can't show the organic search term for a visit. Google stopped sending individual queries to websites in 2011, and browsers now send only the referring domain by default.
• Google Search Console is the main answer. It shows the queries that led to your site and to each individual page, for up to 16 months.
• The Search Console interface stops at 1,000 rows, and rare "anonymized queries" are withheld everywhere. The free BigQuery bulk export removes the row limit — but only from the day you switch it on.
• GA4 can show Search Console queries, but it can't tie a query to a session or a conversion. The landing page is the bridge between the two.
• Paid search is the exception: the Google Ads search terms report shows the actual terms that triggered your ads, apart from low-volume ones held back for privacy.

Why your analytics doesn't show search keywords

Search keywords used to arrive with the visit. When someone clicked a Google result, the link they came from carried their search query, and analytics tools read it out and showed you the keyword.

That ended in stages. On 18 October 2011 Google announced it was moving signed-in searches to HTTPS, and explained what that meant for sites: visitors would still show as coming from Google, but sites "won't receive information about each individual query". In March 2012 Google said it would use the referrer meta tag to simplify the referring URL, so websites would "start to see origin referrers — Google's home pages". Today Google's own help page states plainly: "When you search on Google, you use SSL."

Browsers reinforce this. MDN documents the default policy most browsers apply, strict-origin-when-cross-origin, which sends only the origin — for example https://www.google.com/ — when a visitor moves from one site to another. Your server learns the visit came from Google. It doesn't learn what they searched.

One caveat keeps this honest. Google's help page also says encryption "doesn't always hide" the search terms you typed, so this isn't an absolute guarantee in every case. But for practical purposes, don't expect your analytics to hand you organic keywords. That's why this data now lives with the search engines rather than in your analytics.

Paid clicks were treated differently from the start. Google's 2011 announcement noted that when someone clicks an ad, the browser "will continue to send the relevant query" so advertisers can measure their campaigns. Paid search is still where you'll see real search terms today, through the Google Ads search terms report covered below.

Where every kind of search term actually lives

Where the search happened	Tool that shows it	What you get	Main limits
Google organic search	Google Search Console	Queries per site and per page, with clicks, impressions, CTR and position	16 months; 1,000 rows in the interface; rare queries anonymized
Google organic, at scale	Search Console bulk export to BigQuery	Every query except anonymized ones, daily	Only from the day you set it up
Google organic, inside GA4	GA4's Search Console reports	Queries and landing pages alongside GA4 metrics	Can't be joined to sessions, users or conversions
Google paid search	Google Ads search terms report	The actual terms that triggered your ads	Low-volume terms omitted for privacy
Bing organic search	Bing Webmaster Tools	Organic keywords, by keyword and by page	Web search only; no data before you added your site
Your own website	GA4 site search	What visitors type into your search box	Only searches made on your site
AI assistants	No one, at query level	The assistant that sent the visit	The prompt itself is never passed on

The rest of this guide takes each row in turn.

Google Search Console: the primary answer

Search Console is Google's own record of how your site appears in its results, and it's where organic queries live. Google's definition is simple: "Queries are the search terms that led to your site." The Performance report shows each query with its clicks, impressions, click-through rate and average position.

The default view covers the past three months. Search Console keeps 16 months of data in total, and Google introduced that retention in 2018.

Two limits shape everything you'll do with it.

Anonymized queries. Some searches are never shown, to protect the people who made them. Google defines anonymized queries as those "that aren't issued by more than a few dozen users over a two-to-three month period". They still count in the chart totals, but they never appear as rows. For a site with a long tail of unusual searches, that can be a large share of the real traffic.

Row limits. The report table "can display a maximum of 1,000 rows", and the interface export is capped at 1,000 rows too. Beyond anonymized queries, Google also notes that Search Console "stores and shows only the most important data rows". So for a large site, the interface shows you the head of your query list, not all of it.

Seeing the queries that led to one specific page

This is the question most people are actually asking — "what did people search to find this page?" — and Search Console answers it directly:

1. Open the Performance report for your property.
2. Click + Add filter in the filters row and choose Pages.
3. Choose URLs containing and enter the page's URL or a distinctive part of it.
4. Open the Queries tab. Every query shown is one that led to that page.

It works in reverse too. Click any query in the table to filter the whole report to it, then open the Pages tab to see which URLs Google showed for that search. If one query sends clicks to several of your pages, you've found pages competing with each other.

Filtering changes what the totals mean. Google warns that filtering by query or URL "can affect the report totals due to data truncation and the omission of anonymized queries". Compare pages against each other; don't treat a filtered total as exact.

Getting past the 1,000-row limit

When 1,000 rows isn't enough, you have two routes.

The Search Analytics API returns up to 25,000 rows per request, paging through with a start row, up to 50,000 rows per day per search type. Google notes the API "does not guarantee to return all data rows but rather top ones". It's much more than the interface, but still not everything.

The bulk data export is the complete answer. It sends your Search Console data to BigQuery every day, and Google describes it as including "all your performance data, apart from anonymized queries". Crucially, it's "not affected by the daily data row limit". Anonymized rows are still there, flagged with an is_anonymized_query field, with the query itself left empty. Tables are kept indefinitely by default, so you build a history that outlives Search Console's own 16 months.

The catch is timing. The export starts on the day you set it up; for anything earlier, Google points you back to the API or the reports. If you think you'll ever want query-level history in BigQuery, switch it on now, because the days you miss can't be backfilled.

GA4's Search Console reports — and the limit that matters

Link Search Console to GA4 and you get two extra reports. Queries shows search queries with Search Console's metrics. Google Organic Search Traffic shows landing pages with both Search Console and GA4 metrics side by side. The collection is unpublished by default, so find it under Library and publish it before it appears under Reports. Expect a lag: the data appears in GA4 48 hours after Search Console collects it.

Here's the limit that catches almost everyone. GA4's documentation states that "Search Console metrics are only compatible with Search Console dimensions and the following Analytics dimensions: Landing page, Device, Country." In practice, that means you can't ask GA4 "which query led to this conversion?" A query isn't attached to any individual session or user, so it can't be attached to what that user did next.

What you can do is use the landing page as the bridge. Search Console tells you which queries lead to a page, and GA4 tells you what visitors to that page go on to do. Google's own advice for anything finer is to blend the two in BigQuery, using the Search Console bulk export joined on landing page, device and country. That gives you query-to-outcome analysis by page — useful and defensible, but a statistical association, not a per-visitor fact.

Paid search: the Google Ads search terms report

Paid search is the one place where you can still see actual search terms at scale. Google describes the search terms report as a way "to know how your ads performed when triggered by actual searches". Each row pairs a real search term with the keyword in your account that matched it. The Keyword column "tells you which one of your keywords matched someone's search term and triggered your ad."

You'll find it under Campaigns, then Search terms in the Insights and reports menu.

It isn't complete either. "Some search terms that don't have enough query activity are omitted from the search terms report in order to keep with our standards on data privacy." The report lists terms "that a significant number of people have used". Long-tail searches are the ones that go missing.

The report is also the best free keyword research you already own. Search terms that convert well but that you aren't targeting organically are strong candidates for new content. And terms that burn spend without converting belong on your negative keyword list.

Bing Webmaster Tools

Bing's equivalent is the Search Performance report in Bing Webmaster Tools. It lists "the keywords from organic (non-paid) search that are driving traffic to your website". Like Search Console, it works in both directions. Click a keyword to see which of your pages were shown for it, or click a page to see the keywords behind it.

Three details are worth knowing:

• Web search only. Bing notes that keyword and page data applies to Web traffic, "not applicable to Chat". Copilot answers have their own AI Performance report.
• No history before you join. Data collection starts when you first add your site, so you "cannot see backdated data". Add your site now if you haven't.
• Retention is contradictory. Bing's official blog announced in October 2024 that Search Performance offers "16 months of historical data". Its help page still refers to six months. Export regularly, and don't rely on either figure.

If Bing is a small share of your traffic, it's still worth ten minutes. It's a separate search engine with its own users, so its queries won't simply mirror Google's.

What people search after they arrive: GA4 site search

Everything so far covers searches made before the click. Your own search box shows the other half: what visitors still couldn't find once they'd landed. That's some of the most direct intent data you'll ever collect.

GA4 records it through enhanced measurement. It fires a view_search_results event "each time a user is presented with a search results page, as indicated by the presence of a URL query parameter". By default it watches for five parameters: q, s, search, query and keyword. If your search results URL uses something else, add it in the enhanced measurement settings.

To switch it on, go to Admin, then Data collection and modification, then Data streams, choose your stream, and turn on enhanced measurement. The term people typed lands in the search_term parameter, which populates GA4's Search term dimension.

Put it to work in three ways:

1. Find content gaps. Frequent searches that return weak results tell you what to write next, in your visitors' own words.
2. Fix navigation. If people search for a page that's already in your menu, the menu isn't doing its job.
3. Check landing pages. Searches made straight after landing mean the page didn't answer what brought them there. Pair this with the Search Console queries for that same page, and you can see both what they searched on Google and what they still needed once they arrived.

Connecting search queries to leads and revenue

Since no report ties a query to a person, work at the level the data supports: the landing page. The method takes three steps.

1. Queries per page, from Search Console. For each important landing page, record the queries that bring clicks to it.
2. Outcomes per page, from GA4. For the same pages, take organic sessions, key events, leads and revenue from GA4, filtered to the Organic Search channel.
3. Join on the page. Now you can say "this page, reached mostly through these queries, produced this many leads." For query-level precision, run the join in BigQuery using the Search Console bulk export.

For leads that close later in a CRM, capture the landing page on the lead record itself with a hidden form field. Then the query-to-page link from Search Console carries all the way through to closed revenue.

This is association, not proof of cause. A page that ranks for twenty queries can't tell you which of the twenty brought in a particular lead. But it's accurate at the level it claims, which is more than any tool promising "keyword-level revenue" for organic search can honestly say.

AI search: the queries you can't see at all

AI search is the newest blind spot. Clicks from Google's AI Overviews and AI Mode are counted inside Search Console's Web search type, not separately. Google's generative AI performance report, rolled out to all sites on 31 August 2026, shows impressions for those features — but not the queries behind them. Assistants like ChatGPT and Perplexity pass no query text at all.

That's a big enough topic to need its own guide: how to know which AI search query generated a lead covers what each assistant reveals, GA4's AI Assistant channel, and how to find the prompts that actually drive revenue.

What not to do

• Don't expect keywords in GA4's acquisition reports. They were removed at the source, not hidden by GA4.
• Don't trust a single filtered total. Anonymized and truncated rows make filtered numbers approximate. Compare them; don't quote them as exact.
• Don't wait to switch on the bulk export. Every day before you enable it is a day of query history you can't recover.
• Don't treat the 1,000-row export as your full query list. On a large site it's the head, not the tail.
• Don't buy "organic keyword-level revenue" at face value. No Google tool links an organic query to a conversion. Ask any vendor how they do it; the honest answer involves landing pages or modelling.
• Don't forget Bing's history clock. It only starts when you add your site.

Where an implementation partner helps

Reading Search Console is a skill any marketer can pick up in an afternoon. The harder parts are the plumbing:

• turning on the bulk export correctly;
• building a BigQuery join between Search Console and GA4 that holds up;
• carrying the landing page through your forms into the CRM, so query-level insight reaches closed revenue.

That's where most teams get stuck. If you're on GA4, you can compare GA4 implementation partners or brief one with our GA4 RFP template. For the BigQuery side, see BigQuery partners. If you're not sure what you need, find a partner by need.

Sources

• Google Official Blog — Making search more secure (18 October 2011)
• Google Search Central Blog — Upcoming changes in Google's HTTP Referrer (March 2012)
• Google Search Help — SSL Search
• MDN — Referrer-Policy
• Search Console Help — Performance report
• Search Console Help — Performance report data
• Search Console Help — Performance report table
• Search Console Help — Filtering Performance data
• Google Search Central Blog — Performance data deep dive
• Search Console API — Search Analytics query
• Search Console API — Getting all your data
• Search Console Help — Bulk data export
• Search Console Help — Bulk data export table schema
• Google Search Central Blog — Bulk data export
• Google Search Central — AI features and your website
• Search Console Help — Generative AI performance report
• Google Analytics Help — Search Console reports in Analytics
• Google Analytics Help — Publish the Search Console collection
• Google Search Central — Using Search Console and Google Analytics data
• Google Ads Help — Search terms report
• Bing Webmaster Tools Help — Search Performance
• Bing Webmaster Blog — Extended Search Performance data (October 2024)
• Google Analytics Help — Enhanced measurement