Not Provided Keywords: What You Can Still See

Not provided keywords are organic search queries that Google encrypts and withholds from analytics platforms, replacing them with the label “(not provided)” in your traffic data. You cannot recover the exact query data directly from Google Analytics, but you can reconstruct a reliable picture of your organic keyword performance by combining Google Search Console, on-site behaviour data, and a structured approach to URL-level analysis.

The data is not gone. It is just sitting in a different place, and most teams are not looking in the right one.

Why Google Removed Keyword Data in the First Place

In 2011, Google began encrypting search queries for users who were logged into a Google account. By 2013, it had extended that encryption to all searches, regardless of login status. The official reason was user privacy. The practical effect was that the keyword referral data that had sat inside Google Analytics for years, telling you exactly what someone had searched before landing on your site, disappeared almost entirely.

What replaced it was the “(not provided)” label, which by 2014 accounted for the vast majority of organic keyword data in most accounts. If you were running an agency at that point, as I was, you spent a fair amount of time explaining to clients why a report that used to show them their top 50 organic keywords now showed them almost nothing. It was not a comfortable conversation, but it forced a more honest discussion about what keyword data had actually been telling us, and what it had not.

The loss was real. But it was also, in retrospect, a useful forcing function. Teams that had been using keyword-level attribution to justify organic investment had to build more sophisticated measurement frameworks instead. That is not a bad outcome.

Google Search Console Is the Most Direct Answer

If you want to see what keywords are driving organic traffic to your site, Google Search Console is where you start. It is free, it is directly connected to Google’s index, and it shows you query-level data that Google Analytics does not.

In Search Console, go to the Performance report and select “Search results.” You will see total clicks, impressions, average click-through rate, and average position across your organic search traffic. By default, the Queries tab shows you the actual search terms people used to find your site, ranked by clicks or impressions depending on your sort preference.

This is the data that Google withholds from Analytics but makes available in Search Console. It is not a workaround. It is the authoritative source, pulled directly from the same system that decides how your pages rank.

A few things worth knowing about how to use it properly. First, the date range matters. Search Console retains 16 months of data, so if you are trying to understand seasonal patterns or year-on-year shifts in organic query mix, you need to be deliberate about the window you are analysing. Second, the data is sampled at scale, meaning very low-volume queries may not appear. Third, Search Console reports on impressions and clicks at the query level, but it does not show you what happened after the click. That is where the connection to Analytics becomes important.

For teams building out a broader organic growth strategy, this kind of query visibility feeds directly into how you prioritise content and target new audience segments. If you are thinking about that more structurally, the Go-To-Market and Growth Strategy hub covers the strategic layer that sits above individual channel tactics.

How to Connect Search Console Data to Landing Page Behaviour

The limitation of Search Console in isolation is that it tells you about pre-click behaviour but not post-click behaviour. You can see that a query generated 400 impressions and 30 clicks last month, but you cannot see whether those 30 people converted, bounced immediately, or read three pages and signed up for something.

The fix is to link Search Console to Google Analytics 4. Once connected, GA4 surfaces Search Console data within the Acquisition reports, specifically under “Google organic search traffic.” This gives you a blended view where you can see which queries drove traffic and then filter by the landing page those sessions hit, along with engagement metrics like session duration, pages per session, and conversion events.

The practical approach I have used across a number of client accounts is to work backwards from your highest-converting landing pages. Identify the pages that generate the most organic conversions in GA4, then cross-reference those pages in Search Console using the Pages filter alongside the Queries filter. This tells you which specific queries are driving your most commercially valuable organic traffic, which is a far more useful insight than raw click volume.

It is also worth using Search Console’s “Page” dimension alongside the “Query” dimension simultaneously. When you filter by a specific URL and then look at associated queries, you get a clear picture of the keyword universe driving traffic to that page. Do this across your top 20 organic landing pages and you have a working keyword map that is grounded in actual performance data, not estimated search volumes from a third-party tool.

URL-Level Analysis as a Structural Workaround

One of the more underused approaches to recovering not provided keyword intelligence is systematic URL-level analysis. The logic is straightforward: if you know which pages are receiving organic traffic, and you know what those pages are about, you can make highly reliable inferences about the queries driving that traffic, confirmed and refined by Search Console query data at the page level.

Start by pulling your organic landing pages from GA4, sorted by organic sessions. For each page, you know the topic, the target keyword, and the content structure. In Search Console, filter by that page URL and review the associated queries. In most cases, you will find that a page is ranking for a cluster of related terms, not just the primary keyword you optimised for. That cluster is your actual keyword footprint for that URL.

This matters strategically because it often reveals ranking opportunities you did not plan for. I have seen pages rank strongly for queries that were never part of the original brief, simply because the content addressed related questions naturally. When you map that across a full site, you start to see gaps and opportunities that a keyword research exercise alone would not surface.

The other dimension worth tracking at the URL level is position movement over time. A page that has dropped from position 4 to position 9 for its primary query will show a significant impression decline in Search Console. That is a signal worth acting on, and it is only visible if you are monitoring at the page and query level consistently, not just looking at aggregate organic traffic in GA4.

Third-Party Tools and What They Actually Tell You

SEMrush, Ahrefs, Moz, and similar platforms offer keyword data for your domain based on their own crawling and modelling of search results. This data is estimated, not exact. It is based on what these tools observe in SERPs, combined with their own click-through rate models and search volume data from various sources. It is useful, but it is not the same as first-party data from Google.

Where third-party tools add genuine value in the context of not provided is in competitive keyword intelligence. You can see which keywords your competitors are ranking for, which queries are driving estimated traffic to their pages, and where there are gaps in your own keyword coverage. That is a different use case from recovering your own not provided data, but it is a legitimate and valuable one.

I spent a period judging the Effie Awards, which requires you to evaluate marketing effectiveness claims rigorously. One thing that experience reinforced is the difference between data that is directionally useful and data that is precise. Third-party keyword tools fall into the first category. They are good enough to inform strategic decisions about content priorities, competitive positioning, and keyword targeting. They are not reliable enough to use as the basis for precise attribution claims or to substitute for Search Console data when you are making investment decisions.

Tools like SEMrush’s market penetration analysis can be useful when you are trying to understand your organic share of voice relative to competitors, which is a strategic question that goes beyond individual keyword rankings. That kind of analysis sits comfortably alongside Search Console data rather than replacing it.

The Measurement Mindset That Makes This Work

Earlier in my career, I placed a lot of weight on lower-funnel data. Keyword rankings, direct conversion attribution, last-click performance. It felt rigorous because it was specific. What I came to understand over time is that specificity is not the same as accuracy, and a lot of what performance data appeared to be telling us was a reflection of demand that already existed, not demand we had created.

Not provided keyword data is a version of the same problem. The data is not actually gone. It is just not sitting in the place you expected it, in the format you were used to. Teams that treat not provided as a measurement dead end are usually the same teams that were over-relying on keyword-level attribution in the first place, using it as a proxy for understanding whether their content was working rather than building a proper framework for organic performance measurement.

A proper framework combines Search Console for query-level visibility, GA4 for post-click behaviour, URL-level mapping to connect the two, and third-party tools for competitive context. None of these individually gives you the complete picture. Together, they give you something more useful than the keyword referral data that existed before 2013 ever did, because they force you to think about organic search as a system rather than a list of keywords.

This connects to a broader point about how growth-oriented teams think about measurement. Forrester’s intelligent growth model makes the case that sustainable growth requires a more sophisticated approach to understanding what is actually driving outcomes, rather than optimising the metrics that are easiest to measure. That applies directly to how you approach organic search measurement when the data you want is not available in the format you expect.

Practical Steps to Implement a Not Provided Workaround

If you want to move from understanding the problem to actually solving it, here is the sequence that works in practice.

First, verify that Search Console is properly configured for your property and that your GA4 account is linked to it. This sounds basic, but a surprising number of accounts I have audited over the years had either a broken Search Console integration or a property-level mismatch that meant the data was not flowing correctly. Check it before you assume it is working.

Second, pull your top 30 organic landing pages from GA4 by organic sessions over the last 90 days. Export this list. These are your priority pages for keyword analysis.

Third, in Search Console, use the Performance report with both Pages and Queries dimensions active. Filter by each of your priority pages in turn and export the associated queries, clicks, impressions, and average position. Do this for each page and build a consolidated keyword map that shows which queries are driving traffic to which pages.

Fourth, cross-reference this keyword map against your conversion data in GA4. Which pages are converting? Which queries are associated with those pages? This is your high-value keyword set, and it is the one that should inform your content priorities, optimisation decisions, and any paid search strategy you are running in parallel.

Fifth, set up a monthly reporting cadence that tracks this data over time. Position movement, click-through rate changes, impression shifts, and conversion rate by landing page. Not provided is only a problem if you treat keyword data as a one-time snapshot. When you track it as a system over time, the patterns become clear.

There is a useful parallel here with how go-to-market strategy has become more complex in recent years. The data environment has changed, the tools have multiplied, and the temptation is to either over-engineer the measurement stack or give up on precision entirely. The answer is neither. It is building a framework that is honest about what each data source tells you and what it does not.

What Not Provided Actually Tells You About Your SEO Programme

There is a version of this conversation where not provided becomes a convenient excuse. I have heard it in agency settings more times than I would like: “We cannot report on organic keyword performance because of not provided.” That is not a measurement problem. That is a team that has not built the right reporting infrastructure.

When I was growing an agency from a team of 20 to over 100 people, one of the things that separated the teams who retained clients from the ones who lost them was the quality of their measurement narrative. Not the precision of the data, but the quality of the narrative. Could they explain what was working, why, and what the plan was? Keyword data, provided or not, was one input into that narrative. The teams who treated it as the only input were the ones who struggled when the data changed.

Not provided forces you to build a more complete picture of organic performance. It pushes you toward understanding user intent through content behaviour rather than just query strings. It encourages you to think about keyword clusters and topical authority rather than individual keyword rankings. These are better ways to think about SEO than the keyword-by-keyword approach that dominated before 2013.

The teams that are most frustrated by not provided are usually the ones whose SEO strategy was built around tracking a fixed list of target keywords in a rank tracker, reporting position movements to clients, and calling that measurement. That approach was always a proxy for business outcomes rather than a measure of them. Not provided just made the limitation more visible.

If you are thinking about how organic search fits into a broader growth framework, including how it connects to paid, content, and market penetration strategy, there is more on that in the Go-To-Market and Growth Strategy hub, which covers the strategic decisions that sit above individual channel tactics.

Using On-Site Behaviour to Validate Keyword Inferences

One dimension that gets underused in the not provided conversation is on-site behaviour data. Tools like Hotjar, which offers session recording and heatmap functionality, let you observe how organic visitors actually interact with your pages. This does not tell you the query someone used, but it tells you whether the page is satisfying the intent behind that query.

If a page is receiving significant organic traffic from Search Console data but showing a high scroll depth with low conversion, that is a content-intent mismatch. The page is ranking for queries it is not fully serving. Fixing that is a more valuable intervention than simply knowing the exact keyword string that drove the visit.

Similarly, internal site search data, available in GA4 if you have it configured, can reveal what organic visitors are searching for once they arrive on your site. If someone lands on a category page from organic search and immediately searches for a specific product, that search query is a signal about the intent behind their organic visit. It is not a substitute for keyword data, but it is a useful triangulation point.

The broader principle here is that not provided is a constraint on one specific type of data, not on your ability to understand organic search performance. Every data source has limitations. The question is whether you are building a measurement approach that is honest about those limitations and uses multiple sources intelligently, or whether you are using one limitation as a reason to avoid measurement altogether.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions