Not Provided SEO: What You Can Still Know About Organic Traffic

“Not provided” is what Google returns instead of keyword data for the vast majority of organic search clicks. Since Google moved to encrypted search in 2013 and later removed keyword referral data almost entirely from Google Analytics, SEOs and marketers have been working with a significant blind spot in their organic reporting. The data is gone from the source, but it is not gone from the channel. You can still reconstruct a meaningful picture of keyword performance, and the marketers who understand how to do that have a genuine edge over those who simply accepted the loss and moved on.

I have been in rooms where senior stakeholders have looked at organic traffic reports and asked which keywords are driving conversions. When the honest answer is “we cannot tell you directly from Analytics,” the room gets uncomfortable. That discomfort is understandable, but it is also a symptom of over-relying on a single data source. The real problem is not that keyword data is hidden. The real problem is that most teams never built a reporting structure that could survive without it.

What “Not Provided” Actually Means in Your Data

Before 2013, Google Analytics passed the search query through the referral URL when a user clicked an organic result. You could see it in the keyword dimension of your traffic reports. Then Google switched organic search to HTTPS by default, which stripped the query parameter from the referral string. The keyword appeared as “(not provided)” in Analytics, and over the following months it went from a minority of organic sessions to the overwhelming majority. By 2014, the keyword dimension in Google Analytics for organic traffic had effectively become useless for most sites.

This was not a conspiracy. It was a side effect of a security change, and Google had already been signalling for some time that user privacy in search was a priority. The business logic for Google was also not hard to understand: keeping query data inside its own ecosystem (Search Console, Google Ads) gave it more control over how that data was used. Whether you find that convenient or cynical depends on your perspective. I find it both.

What “not provided” does not mean is that the data has vanished entirely. It means the data is no longer passed through Analytics automatically. Google Search Console still shows you the queries that generated impressions and clicks for your site. The limitation is that Search Console aggregates data at the query level and at the page level, but it does not join those two dimensions in a way that lets you see, with precision, which query drove a conversion. That gap is real, and it matters for anyone trying to close the loop between organic search and revenue.

This connects to a broader point I keep coming back to across the SEO articles in this series: analytics tools are a perspective on reality, not reality itself. If you want the full picture of your SEO performance, the complete SEO strategy hub covers how measurement, positioning, and content strategy fit together as a system rather than a collection of disconnected tactics.

Why the Loss of Keyword Data Is a Bigger Problem Than Most Teams Acknowledge

When I was running iProspect and we were growing the team from around 20 people toward 100, one of the things I noticed was how quickly reporting conventions become institutional. Someone builds a dashboard, it gets signed off, and within six months everyone assumes the dashboard shows what matters. The “(not provided)” problem accelerated that dynamic because it created a visible gap in the data that teams filled with proxies, and then forgot they were using proxies.

The practical consequence is that organic channel reporting often conflates traffic volume with performance. You can see that organic sessions went up or down. You can see that a particular landing page attracted more visits this month than last. What you cannot see clearly is whether the queries driving that traffic are the queries you were targeting, whether they have commercial intent, or whether they are converting at a rate that justifies the investment in the content.

That matters because SEO is not free. Content costs money to produce. Technical SEO requires developer time. Link acquisition takes effort and sometimes budget. If you cannot trace organic traffic back to queries with enough granularity to evaluate whether your investment is working, you are flying with limited instruments. You might still land the plane, but you are taking on more risk than you need to.

There is also a competitive dimension here. Teams that have built better workarounds for the “not provided” problem are making more informed decisions about where to focus their SEO effort. That compounds over time. The gap between teams with good organic measurement and teams without it is not dramatic in any single quarter, but over two or three years it becomes significant.

How to Reconstruct Keyword-Level Insight Without Direct Data

The most reliable approach most SEO practitioners use is connecting Google Search Console data to Google Analytics landing page data. The logic is straightforward: Search Console tells you which queries are generating impressions and clicks to specific URLs. Analytics tells you what those URLs are doing in terms of engagement and conversion. If you join those two datasets at the page level, you get a reasonable approximation of which queries are contributing to business outcomes.

In practice, this means exporting Search Console query data for a given page, then pulling the Analytics performance data for that same page, and mapping them together. It is not perfect. Search Console samples data and caps the number of rows it returns. It also aggregates across all queries for a page, so if a single URL ranks for hundreds of queries, you are looking at a distribution rather than a precise attribution. But it is significantly better than nothing, and most teams are not doing it systematically.

A more automated version of this is available through Google’s Looker Studio, where you can connect both data sources and build a combined view. There are also third-party connectors that pull Search Console and Analytics data into a single reporting environment. The setup takes time, but once it is running it becomes a genuinely useful reporting layer that most organic channel reports are missing.

Beyond that, third-party rank tracking tools like those built into Ahrefs and similar platforms can give you position data for keywords you are actively monitoring. The limitation is that you have to know which keywords to track. You are not discovering unknown traffic drivers, you are monitoring known targets. That is useful but incomplete. The universe of queries that actually send traffic to your site is almost always larger than the list of keywords you chose to track when you set up your campaign.

One approach I have found underused is using on-site search data as a proxy for intent. If you have a search function on your site, the queries people type into it tell you something about what they were looking for when they arrived. It is not keyword data in the traditional sense, but it is behavioural signal that can inform both content strategy and the interpretation of organic traffic patterns. Tools like Hotjar can extend this further by surfacing qualitative feedback from users that helps you understand whether the content they found matched what they were searching for.

What Search Console Actually Shows You and Where Its Limits Are

Google Search Console’s Performance report is the primary tool most SEOs use to work around “not provided,” and it is worth being precise about what it does and does not show.

It shows you queries, impressions, clicks, click-through rate, and average position. You can filter by date range, device, country, and search type. You can see which pages are generating the most impressions and which queries are associated with those pages. That is a meaningful dataset. The problem is the ceiling on the data. Search Console only shows queries that generated at least one impression in the selected period. Queries that are close to the threshold may be omitted. The data is also sampled at scale, and the row limit in exports means large sites can miss a significant portion of their long-tail query traffic.

There is also the question of how Search Console handles query attribution. If a user searches for something, clicks your result, goes back to the results page, and then clicks again, Search Console may count that differently from Analytics. The two tools are measuring different things in different ways, and reconciling them is an exercise in approximation rather than precision. That is fine as long as you acknowledge it. The mistake is treating Search Console data as ground truth when it is actually a useful but imperfect signal.

I judged the Effie Awards for several years, and one of the things that struck me about the entries that did not win was how often teams confused measurement with proof. They had data. They had charts. But the data was not actually showing what they claimed it was showing. The “not provided” problem creates the same trap in SEO reporting: you can build a report that looks authoritative while quietly obscuring the fact that you are missing a significant portion of the underlying information.

Building an Organic Reporting Framework That Accounts for the Gap

The practical response to “not provided” is not to find a single perfect data source. It is to build a reporting framework that triangulates across multiple imperfect sources and is honest about the gaps. That sounds obvious, but most organic channel reports I have reviewed in agency and client-side contexts do not do this. They report on what they can measure and imply that what they can measure is what matters.

A more defensible approach uses at least three data layers. The first is Search Console query data, which gives you the closest thing to keyword-level organic performance available from Google directly. The second is Analytics landing page data, which connects organic traffic to on-site behaviour and conversion. The third is rank tracking for your target keyword set, which gives you position trend data that is independent of traffic volume and helps you diagnose whether visibility changes are driving the traffic changes you are seeing.

On top of those three layers, you can add competitive intelligence from tools that estimate organic traffic and keyword rankings for competitor domains. This contextualises your own performance. If your organic traffic dropped in a given month, knowing whether your competitors also dropped (suggesting a market-wide shift or an algorithm update) is more useful than looking at your numbers in isolation. The in-house SEO expertise question is relevant here: building and maintaining this kind of multi-source reporting framework requires genuine analytical capability, which is one reason many organisations underinvest in it.

The output of this framework should not be a single traffic number. It should be a set of signals that together tell a coherent story about organic channel health. That story should be revisable when new data comes in. It should acknowledge uncertainty. And it should be connected clearly to business outcomes rather than just SEO metrics.

The Broader Implication for How SEO Gets Evaluated

The “not provided” issue is a specific technical problem, but it points to something more fundamental about how SEO gets evaluated inside organisations. When I was at Cybercom early in my career, I remember being handed a whiteboard marker mid-brainstorm for a Guinness brief when the founder had to leave for a client meeting. My internal reaction was something close to panic, but the lesson I took from it was that you have to be able to make decisions and defend them with incomplete information. That is true in creative work and it is equally true in channel analytics.

SEO has always had a measurement problem. The channel is long-cycle, attribution is complex, and the factors that influence rankings are numerous and not fully transparent. “Not provided” made that problem more visible, but it did not create it. Teams that treat the absence of keyword data as the core obstacle are often using it as cover for a broader lack of rigour in how they evaluate organic performance.

The more useful framing is that SEO measurement requires honest approximation rather than false precision. You are not going to get a clean, auditable chain from search query to conversion in most cases. What you can get is a set of signals that, taken together, give you enough confidence to make reasonable decisions about where to invest, what content to prioritise, and when the channel is under or over-performing relative to expectations.

Generative AI is adding another layer to this. As AI-driven search features change how results are displayed and how users interact with them, the relationship between impressions, clicks, and conversions is shifting again. The impact of generative AI on SEO and content performance is worth understanding in this context, because it compounds the existing measurement challenges rather than resolving them.

If you are building or revisiting your SEO measurement approach, the complete SEO strategy framework at The Marketing Juice covers how to connect organic channel measurement to the broader strategic picture, including how to set expectations with stakeholders who want cleaner numbers than the channel can honestly provide.

Practical Steps to Take Now

If your current organic reporting is built primarily on Analytics traffic data with no systematic use of Search Console query data, the first step is straightforward: connect both data sources in Looker Studio and build a combined view that shows landing page performance alongside the queries driving traffic to those pages. It takes a few hours to set up and immediately improves the quality of your organic reporting.

The second step is auditing your rank tracking setup. Most rank tracking configurations are built around the keywords that were identified at the start of a campaign and never substantially revised. Run a Search Console export for the last three months and identify queries that are generating meaningful impressions or clicks that are not in your rank tracking tool. Add them. Your tracked keyword set should reflect what is actually happening in the channel, not just what you hoped would happen when you started.

Third, build a regular process for reviewing Search Console data at the page level rather than only at the query level. Pages that are generating high impressions but low click-through rates are often candidates for title tag or meta description improvement. Pages that are generating clicks but high bounce rates may have a content-to-intent mismatch. Neither of those insights requires keyword-level data from Analytics. They are available from the combination of Search Console and on-site behavioural data that you already have.

Finally, be explicit in your reporting about what you can and cannot measure. Stakeholders who understand the limitations of organic data are better positioned to make sensible decisions about the channel than stakeholders who have been given false confidence by reports that imply more precision than the underlying data supports. That conversation can feel uncomfortable, but it is considerably less uncomfortable than explaining why a channel that looked healthy on paper has not delivered the commercial outcomes that were expected of it.

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