Save over 40% when you secure your tickets today to TNW Conference 💥 Prices will increase on November 22 →

This article was published on December 1, 2016

Marketing the TNW Way #17: 50+ Custom Dimensions in Google Analytics


Marketing the TNW Way #17: 50+ Custom Dimensions in Google Analytics

In this blog series, we shed some light on our marketing approach at The Next Web through Web analytics, Search Engine Optimization (SEO), Conversion Rate Optimization (CRO), social media, and more.

During MeasureCamp London a few months ago I shared parts of our set up and promised to blog more about the custom data that we save. In a previous blog post I talked more about the 24 calculated metrics that we’re using. But it’s about time I do that as well for the 59+ custom dimensions what we have. So in this blog post I’ll try to share the pros and cons for the data that we save and how it helps us provide more value to our Google Analytics data.

We love geeking out over Google Analytics. With all the custom information we can monitor, we felt it was about time to give our readers a peek a how we’re using it at The Next Web. Today, we’ll look at custom dimensions.

Why so many custom dimensions?
Data is Dumb! Period. Having a ton of hits in your web analytics tool doesn’t really cut it, that’s why you need to bundle data with additional context to make sense of it and come up with better ideas. We all want to know how many pageviews a certain post has, but if we could gain additional information at the same time? This is what custom dimensions allow us to accomplish.
 
What got us to 59+ custom dimensions?
Last year we started to add additional custom dimensions. When we switched over to Google Analytics 360, we went from the 20 we had to our current 59. This allows us to ensure that all the information we have about our post is covered in our Schema.org setup but also in our Google Analytics setup.
 
What to track & how it’s providing context?
In our case, we’re a publisher, so we always want to know more about the content that we provide — whether the topic is a company, person, an event or just the time it was published. See what I did there? I just listed four custom dimensions. Of these, the last one — date published — can be set up in multiple ways as you want to know the date/time, the hour, and the day of the week so you can later compare this data. At this rate it’s easy to see how the number of custom dimensions you use can spiral out of control.

In addition, you might consider saving your data both on User and Session Level. I would like to know if we identified the User, in general, but I’d also like to know which session the user happened upon. Two custom dimensions on session + user level are able to help me answer that question.
 
Let’s do it: 40+ examples of custom dimensions
Wait, wut!? Not sharing the 59 examples that you’ve talked about before? I’m sorry, some of the 9 examples are pieces of information that only would work for The Next Web, are outdated, and/or custom to our situation:

  1. Time – Session Level
    Google Analytics does save the timezone of a user and presents it in a nice report. But what if you get a lot of global visitors? The reports don’t make sense anymore. To make sense of it, we’re sending the hour the user is currently in to get a better overview of when you’re visiting the site.
  2. Tags – Hit Level
    All the tags that are related to an article.
  3. Post Type – Hit Level
    We have different post types: articles + video posts. So we’d like to know the difference while analyzing.
  4. Author – Hit Level
    Makes sense, I guess: the name of the Author.
  5. Post Date – Hit Level
    The post date of the article: YYYY-MM-DD.
  6. Adblocked – Session Level
    Is the user currently using an Adblocker in his session. We’re analyzing this via a JavaScript trick.
  7. Company – Hit Level
    What kind of company is this article about: Google, Apple, Samsung? We’re saving the companies mentioned in this article based on a list of ±150 companies that we maintain internally.
  8. Profile – Hit Level
    What kind of people do we write about: Eric Schmidt, Steve Jobs? Much like companies, we’d like to know who you’re interested in.
  9. Visitor Type – Session Level
    Are you a new visitor or returning visitor? We’re looking at a cookie to see what your current status is.
  10. A/B Testing – Session Level
    Are you currently in an A/B tests, if so: which one and what variant are you in?
  11. Event – Hit Level
    Is this post related to an event like: SXSW, Mobile World Congress or one of our own conferences?
  12. Comments – Hit Level
    The number of comments on this post.
  13. Sponsored Post – Hit Level
    Is the post sponsored, Yes or No? In another custom dimension we save who the sponsor would be.
  14. Category – Hit Level
    What WordPress category was this posted in?
  15. User ID – Hit Level
    The User ID that we maintain for this user. We’re also saving this within the userId feature of Google Analytics, but having this in a custom dimension makes it easier to use within the GA interface.
  16. Homepage – Hit Level
    Our homepage block can have multiple identities, so are you seeing 1 article or multiple? We save the blocks structure.
  17. Canvas – Hit Level
    Are you exposed to a canvas Ad? If so, which one?
  18. Sponsor – Hit Level
    Is the article sponsored? If so, we want to know who’s sponsoring it. This makes it easier for our sales team to filter down on all articles of a specific sponsor.
  19. Sections – Hit Level
    Next to categories, we’re using Sections on The Next Web for editorial purposes (the ones you see in the top navigation).
  20. Syndicated – Hit Level
    Is this article syndicated? Did we publish the article based off another site (always in agreement with the original owner)?
  21. Country – Session Level
    What country are you from? We’re saving this data based on the GeoIP lookup table so we can use it for personalization.
  22. GTM Container Version ID – Session Level
    The version of our Google Tag Manager container, great for debugging if at some point we would need to debug what went wrong with publishing a new GTM container version.
  23. Roadblocked ads – Hit Level
    Advertisers can sponsor an article with ads. We like to know how they’re doing.
  24. Signed up for newsletter – User Level
    Are you signed up for the newsletter on the site? If so, we use this to hide the newsletter signup box for you. This gives you a different experience then the majority of users, we want to be able to segment on that.
  25. First visit – User Level
    When was the first ever visit you made to The Next Web?
  26. Last visit – User Level
    When you are a returning user, when did you last visit us?
  27. Publish date and time – Hit Level
    We saved the published date before, but this time we’re including the time that it was published.
  28. Post ID – Hit Level
    The Post ID from WordPress, great for connecting some data outside of Google Analytics.
  29. Client ID – Session Level
    The Client ID that Google Analytics is setting, so we can more easily identify users in BigQuery but also in Google Analytics if needed.
  30. Session ID – Session Level
    The same for a sessions ID, we want to know what kind of users are visiting us. So we want to follow a user in a specific session by this ID.
  31. Hit timestamp – Hit Level
    The timestamp of the hit is not available in the reporting interface, it is in BigQuery, so we’d like to have that data in the reporting interface too.
  32. URL string – Hit Level
    What kind of URL string did you have on this page? We need to know what specific parameters might have been included in your page path (we filter them out in our main view).
  33. Query string – Hit Level
    What kind of query strings do you have on this URL?
  34. Full referral path (Hit) – Hit Level
    Google Analytics is not great in recognizing traffic coming from mobile applications, but if you save the full referral path you can snag some information still. It’s very useful to segment these kind of users.
  35. Region – Session Level
    What continent is the user located in, related to the Country that we saved before?
  36. Shares – Facebook – Hit Level
    The number of shares for the article on Facebook, also used to calculate the share counts on The Next Web.
  37. Shares – Twitter – Hit Level
    The number of shares for the article on Twitter, also used to calculate the share counts on The Next Web.
  38. Shares – LinkedIn – Hit Level
    The number of shares for the article on LinkedIn, also used to calculate the share counts on The Next Web.
  39. Shares – GooglePlus – Hit Level
    The number of shares for the article on Google+ (yes, it still exists), also used to calculate the share counts on The Next Web.
  40. Shares – Reddit – Hit Level
    The number of shares for the article on Reddit (we love traffic spikes from Reddit), also used to calculate the share counts on The Next Web.
  41. Count Images – Hit Level
    How many images are embedded in the post. This will tell us more about the structure of the post and what could influence you to stay longer on the page.
  42. Count Videos – Hit Level
    Same for the videos, if we embed two videos of a certain length, we might be able to keep you on the page even longer. Since it’s contextual data on an article, we’re saving it.
  43. Published Date – Week – Hit Level
    The week the article was published. Provides us with an easier data point to do week over week analysis on the performance of our editorial team.
  44. User Known – Session Level
    Do we know this user, as in, do we have a User ID? We’re saving this both on User and Session level. When you’re doing multiple visits to the site we’d like to know in which session we started identifying you.
  45. User Known – User Level
    Do we know this user at all?
  46. Returning User – User Level
    Is this User a returning one or is this his first visit?
  47. Returning User – Session Level
    Is this User returning in this session, or are we identifying this as the users first session to The Next Web?
  48. Length Buckets – Hit Level
    What is the length of the post? We’re bucketing this in different buckets so we don’t end up with too many to monitor. By analyzing this we know if you like shorter or longer posts (and we can obviously combine this with the # of images & videos).
  49. Full referral path (Session) – Session Level
    From what kind of referral path did this user come before. The same as the referral path, but this time saved on a Session level.

Note: It’s important to know that we don’t save any personal identifiable information (PII) in Google Analytics. Also this data is never sold to advertisers so they can identify users.

What’s next?
Like always, I’ll shine my light on what we consider to be the next step for this. So far, over the course of the last year, we’ve added around 10-15 new custom dimensions. Every month we’ll come up with something missing that would be beneficial to our set up. That’s why we’ll probably be doing the same for the next year.

I must admit, there is currently nothing really exciting around custom dimensions on our roadmap. While we’re working on some projects around The Next Web it will mostly be maintenance to make sure we can support some other data elements for, for example TNW Answers.

If you missed previous posts in this series, don’t forget to check them out: #1: Heat maps , #2: Deep dive on A/B testing and #3: Learnings from our A/B tests, #4: From Marketing Manager to Recruiter, #5: Running ScreamingFrog in the Cloud, #6 What tools do we use?, #7: We track everything!, #8: Google Tag Manager , #9: A/B Testing with Google Tag Manager, #10: Google Search Console, #11: 500 Million Search Results and #12: How are you engaging with this page?, #13: Supporting Schema.org, #14: Calculated Metrics in Google Analytics, #15 How TNW uses Google Optimize 360, #16 Using Google Tag Manager in AMP

This is a #TNWLife article, a look into the lives of those that work at The Next Web.

Get the TNW newsletter

Get the most important tech news in your inbox each week.

Also tagged with