How To Turn (Not Provided) Into Useful, Actionable Data

We’ve all seen it, lurking in our Analytics reports, nearly always at the top, sucking a huge chunk of data into a black hole of uncertainty and uselessness.

Not provided was predicted as having a single-digit impact on sites. In my research, I’ve found it to have upwards of a 40% impact, especially on smaller traffic sites.

In conversations with clients, and looking at my own sites, I knew the numbers had to be better – but what do you tell a client when chunks of data, in some cases over 30%, are attributed to (not provided) instead of the keywords you’ve given time, attention and tears to over the previous months and years.

Honestly the standard, “It’s happening, but Google isn’t showing us what it is” is just not good enough. Its not good enough for me; why would it be good enough for a client?

I was on a mission… to find some way to make this data useful again. My goal? I needed a way to attribute (not provided) data to a page, or a keyword phrase, without tearing my hair out. My first stop was – it’s definitely a go-to for ready-made reports. I’ve talked before about not reinventing the wheel; well, this is where I started.

Custom Filters

I found a great Custom Filter here that allowed me to see a good chunk of data. With this filter, instead of (not provided) we are allowed to see the page that the (not provided) click landed on. This then gives us an idea of what keyword phrases are driving that traffic. Here’s an example:

From September 7 to October 6, my hobby blog received 5,272 visits from organic search. Of those visits, 1,518 were attributed to (not provided).  I lost nearly 30% of my keyword data. So here is what I saw:

I have implemented the Filter mentioned above on a separate Google Analytics profile.  Any time you’re looking at changing how data is reported with advanced segments and filters, it’s smart to have a profile separate from your main profile, where wonky data reporting wont hinder the true data. If it doesn’t work, you want to be able to undo what you’ve done without permanent adjustments to your tracking.

Luckily, this advanced filter worked, and the not provided data became much clearer. With the Advanced Filter in place, here is what I see for the same report and time period:

In the case of my hobby blog, most of those 1,518 visits were attributable to a single post; this likely will not happen with your site – unless you have one piece of content that drives large amounts of traffic like I do.

You can immediately see the added value in the filtered results. I now know which keyword efforts are driving traffic to my site, when previously the data was buried in (not provided) limbo. I can now use the advanced filtering and exporting to group keywords and landing pages together to see actionable results. I’m working on building a custom report that will do something like this and will share it as soon as I get it figured out.

Setting Up Advanced Filters To Improve Not Provided Results

Setting up a new profile with this advanced filter in place couldn’t be easier. Follow the steps in the link above to set up a new Google Analytics profile for your account.

Here are my tips to make things easier as you set up your new profile:

  • Give your new profile a descriptive name – I used “ – Not Provided” because I created the profile solely to test (not provided) filtering and reporting.
  • Anyone who has admin access to your account will see this new profile, so make sure they know you’re testing and making changes.
  • Don’t skip this step – you want to be able to easily delete the profile if the data becomes corrupted or your filtering and segmentation doesn’t work

Once you’ve set up the new profile, you can immediately create your advanced filter. Click on the “Filters” Tab and click the “New Filter” button.

You’re going to create a New Custom Filter that looks exactly like this:

When finished, click Save and wait for a while for some data to come in. You’ll start seeing “NP –” instead of (not provided) quite quickly.

Custom Report Sharing also featured an advanced segment that just eliminates all (not provided) data from the results. If you’re seeing low single-digit site impact, this might be a good option for you. For me, this piece wasn’t ideal, because such a large amount of data was (not provided), my analysis could have been handicapped without it.

I did some more searching and found a couple of approaches I didn’t try – they didn’t seem to provide any better results than what I had from my examples. If you’d like to try them – check it out.

Note: If you think of data and Analytics as a hobby, and you’re not reading Avinash Kaushik’s blog, you’re missing out. I read his posts 3 or 4 times just to try and glean as many nuggets as I can. 

These tips are designed to help you SEE the data, which then makes it actionable. I say it a lot, and I’ll say it again here, if you’re just looking at data, and not making action plans based on what you see, you’re wasting your time. Data is just words and numbers on a page, analysis and plan derivation from those data points is what makes reviewing your Analytics profitable.

Give this a try and see if you can get a better picture of how your site is performing organically. If you have another way to help you see (not provided) data, I’d love to hear about it in the comments below.

See also our related article on Marketing Land today: Dark Google: One Year Since Search Terms Went “Not Provided”

Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.

Related Topics: Advanced | Beginner | Channel: Analytics | Google: Analytics | How To | How To: Analytics | Search & Analytics


About The Author: is the co-founder of Ignitor Digital, along with long-time colleague Mary Bowling. At Ignitor, Carrie tackles tough technical SEO roadblocks many small business owners don't even know they have. Her experience with analytics and troubleshooting helps her get to the root of issues. When not working, Carrie loves to cook for friends and family, hang out with her pretty awesome kids, and read books that have little-to-no educational value! You can also follow Carrie on twitter, @carriehill.

Connect with the author via: Email | Twitter | Google+ | LinkedIn


Get all the top search stories emailed daily!  


Other ways to share:

Read before commenting! We welcome constructive comments and allow any that meet our common sense criteria. This means being respectful and polite to others. It means providing helpful information that contributes to a story or discussion. It means leaving links only that substantially add further to a discussion. Comments using foul language, being disrespectful to others or otherwise violating what we believe are common sense standards of discussion will be deleted. Comments may also be removed if they are posted from anonymous accounts. You can read more about our comments policy here.
  • audette

    Your approach is similar to what I break down here: although I took it a step farther and used scraped data (via SEMRush) to ‘reclaim’ the lost queries.

    Ultimately though, I think what we’re learning is that (not provided) is not that big of a deal. Keywords are what people use to ‘get’ the stuff the want. The stuff they want resides on our websites. What’s truly valuable are URLs that hold that stuff. The ‘bait’ we use (keywords) is secondary to that.

    Also, there are plenty of good sources to get query data: third-party competitive insight tools, paid search, site search, keyword research tools.

    Until (not provided) reaches 99% I’m not worried.

  • Neil Armstrong

    That’s a great partial solution to a common problem as opposed to just multiplying evrty result by 20% as had been my workaround……

  • Elizabeth Jamieson

    I have been wondering how to make sense of so that data – thanks so much for pointing out a solution. I seem to have more “not provided data” than anything else! What is confusing to me is that – these are not keywords, right? So why were they not inlcuded in referral data – as they look like referrals ….

  • Taylor Gould

    Thanks for posting. My company builds products for Google Apps users, so our [not provided] percentage is upwards of 80%. I’ve tried a number of techniques but this is one of the better ideas and takes fewer steps than most.

  • Carrie Hill

    Hi Adam, Thanks for your comment.

    I also believe for many sites Not Provided isn’t really an issue. For some, like mine, where it’s a good chunk, just knowing the URLS is really helpful, because i basically know what keywords are driving visits to those URLs.

    I’ll check out your breakdown also, thanks so much for sharing!


  • Carrie Hill

    Hi Elizabeth, Actually this is keyword data. Google reports (not provided) when someone performs a search and visits a website while signed into a google account>

    From the page:
    “Keyword: The keywords that visitors searched are usually captured in the case of search engine referrals. This is true for both organic and paid search. If the a visitor is signed in to a Google account, however, Keyword will have the value “(not provided)”.”

  • Yousee

    This is a nice useful tips, something is better than nothing.

  • Jon

    Not just signed in users, it’s anyone on a https:// Google search page, whether they’re signed in or not

  • Pedro Matias

    Thanks, this is the easiest to implement take on the “not provided” I’ve seen so far!

    I also like the assertion that: “if you’re just looking at data, you’re wasting your time”. Lots of people install analytics and then just look at the visitor counts, in many cases not even filtering their own traffic, which is just misleading their own ego.

    Analytics should be at the core of page and site redesign and content planning. I’m right now in the middle of trying to make a small client see how the all picture works.

  • Goran Candrlic

    Go to: Traffic Sources > Search > Organic
    Click on the “not provided”
    Pick secondary dimension: Traffic Sources > Landing page
    You have this data

  • Albert Foekema

    Thats an most nice way to work at least a bit with the ‘lost’ data! Most welcome tip and ill deffo try this out on a campaign! Cheers!

  • Jan

    Just wanted to post the same, don’t get the advantage of creating an additional filter for this purpose.

  • SEO First

    Awesome, fantastic share!

  • Cody Sharp

    The best way to find out detailed info for not provided is simple. Basically you use data sampling. Get the running 3 months average of not provided in organic search (number compared to overall organic search words). Let’s say its 33%. Now just multiply all of your organic traffic keywords by 1.33. After all, it makes sense that the not provided people are searching for the same stuff as the provided visitors.

  • Cody Sharp

    Yeh…but Goran, that wouldn’t fill a 300+ word article :)

  • Cody Sharp

    All Firefox users now have a secure version of Google search used and their results also feed into (not provided).

  • Carrie Hill

    Hi Goran,

    Thanks for your feedback. 2 Things, 1) You can view Landing page as a secondary dimension but it makes kind of a mess out of the report with lots of (not provided) lines in the keyword column.. 2) my version combines not provided URL data with keywords in a cleaner view

    If I was working on the fly I would and have used your version, but what I really wanted this report to do was clarify the keyword report a bit for me, not just show me 100% landing page data. I also want data available without having to sort and filter and choose other dimensions, I need it delivered fast and easily without having to think about it, hence the advanced filter.

    Your way certainly works, you’re welcome to it, but it didnt do what I needed in several instances, namely report on what data was “lost” in the “Not Provided” hole.

  • Adam Whittles

    Hi Carrie! This is an excellent post and I’ve already started testing it on my sites. I have a question though.

    Unfortunately I am still seeing a handful of (not provided) keywords even after implementing this method. Any ideas what could be causing this?

  • Carrie Hill

    Hi Adam, I’ve not run into this issue – so forgive my elementary line of questioning.

    1) is the timeframe you’re using 100% after you implemented the filter?
    2) did you double and triple check the filter steps, comparing what you have above with what you did? even missing something as simple as a period or coma can throw the filter off
    3) if you’d like me to take a look – shoot me an email at carrie (at) keyrelevance (dot) com.

    Thanks for the feedback :)

  • Adam Whittles

    Thanks Carrie, I double checked the filter and stupidly I had put (not provided) in field A as opposed to (.not provided.). Doh!! I’ve restarted with a new profile and filter so I’ll let you know if I run into the same problem (I doubt I will now this has been fixed!) Thanks again Carrie.

  • Carrie Hill

    @twitter-464439706:disqus – you’re welcome! Glad you got it fixed!

  • Warren Lee

    Given the rise in personalized search, and recent adoption of secure protocol by browsers like firefox, using browser market share data I’d estimate that 40% is most common. Even with this % of data going into the void i’m sorry to say this filtering method sounds like a waste of time to me.

    Why not do it this way instead… Apply the % contribution of each of the known keywords to the “not provided” number and apply the result back to each keyword. My understanding is that this is the method of choice for most in-house SEO managers at leading brands. However at the end of the day, the method that one chooses in order to make their data based insights actionable must match the point that they are trying to make. That said I agree with your point about the value of making the data actionable but might encourage more elaboration regarding some of the actionable insights you might provide after using this filter. In short, how does using this filter make the data any more actionable, other than it being a little easier to see? For example, some really actionable examples of using the % contribution method to drive insights can include making a better case for new content or other optimization efforts, testing co-optimization efforts between paid and natural search where you need data at the keyword level, or driving keyword level insights into performance attribution.

  • rama

    Sorry Carrie, this doesn’t serve any purpose.Waste of time

  • Todd Walbridge

    Good stuff I feel compelled to think twice about data that is actionable or not. I am one of those that studies data on a weekly basis. My take is different in that I have my websites at a comfort level and am looking for problems. To that end if it is not going down I consider it non actionable. If I was looking to grow I would have to confer 100%. Thanks for a good read

  • Gordon Campbell

    Carrie, this is genius! As Goran said you can get the information using the method that he mentioned but your version is so much more organised and is a big time saver. Excellent post!

  • Manish Chauhan

    Not sure how do you see this a time savor when you can simply do the same as Goran said.

  • Manish Chauhan

    well said..seems author has attempted to look a bit creative, however I really don’t see any logic to use these steps when landing page details are available by default.

  • Marc Perez

    No tool can give you the not provided info.

  • dan barker

    hi, Manish, I actually came up with this method in the first instance. SEL said they were looking into why it had been posted here with no mention of the original post, but I never heard back.

    (original was here: )

    Anyway, you’re right – you can get this as a secondary dimension. The limitations of doing it as a secondary dimension was one of the things that prompted me to look at other options. It’s not right for every account, but it can help in many cases & I’ve found it useful in a lot. Here are some advantages:

    1. It gives you all data straight within the keyword field, whether you’re using it in the standard report, in a dashboard, via the API, or via any other means.
    2. If you use ‘keyword’ as a secondary dimension in another report, this data is present there too.
    3. You can quickly & easily filter it in/out of reports simply using ‘np -’ in the filter field, and can filter out particular URLs (eg. filter out the homepage on the assumption it’s brand, etc).
    4. When using multichannel funnels, this appears in ‘keyword path’ too.
    It’s not always appropriate, and mileage varies depending on your URLs, the scale of your site, etc.

    Hope that helps,


  • Jon


Get Our News, Everywhere!

Daily Email:

Follow Search Engine Land on Twitter @sengineland Like Search Engine Land on Facebook Follow Search Engine Land on Google+ Get the Search Engine Land Feed Connect with Search Engine Land on LinkedIn Check out our Tumblr! See us on Pinterest


Click to watch SMX conference video

Join us at one of our SMX or MarTech events:

United States


Australia & China

Learn more about: SMX | MarTech

Free Daily Search News Recap!

SearchCap is a once-per-day newsletter update - sign up below and get the news delivered to you!



Search Engine Land Periodic Table of SEO Success Factors

Get Your Copy
Read The Full SEO Guide