With so much difficulty in measuring the value of social programs and platforms, Tumblr is a blessing from the Social Platform heavens.

tumblr-measureable-platform

Tumblr pages are customizable and therefore allows for a number of measurement methodologies that most 3rd party social platforms do not.  Ultimately, this means we are able to have a holistic view of the impact that Tumblr as a platform is having on our overall business objectives.

The first question I always get as it relates to Tumblr measurement is, “Where should I put the data?”  Allow me to outline the what I believe to be the best data architecture for not just Tumblr data, but all social data in general.  To do so, I will use Oxygen Network as an example, because of their reputation of early-adoption of emerging Social Channels (This does not imply that they are leveraging the measurement practices outlined in this post):

Oxygen has a primary .COM site, and a number of different show sites.  Typical data collection best practices would have them structure their deployment with a separate Report Suite (data collection silo) for each show, so they can report on each show independently, and then also have that data roll up into a global Report Suite as shown:

oxygen--properties

Now, you will also find corresponding social profiles for each show and for Oxygen in general.  The best practice is to report on the Tumblr data in the same way you would collect data for each digital property:

Screen Shot 2013-03-13 at 2.50.51 PM

Now that we have a home for this data, let’s talk through each of the key measurement and analysis opportunities that Tumblr provides:

  • Brand Page Measurement – Real-time data collection for branded Tumblr profile.
  • Owned Post Clickthroughs – Measuring the impact of posts that include links back to conversion pages
  • Tumblr Monitoring – Listening to all of the Tumblr posts that are related to your brand / products
  • Influential ‘Tumblrs’ – Identifying individual Tumblr accounts that are driving onsite clickthroughs and conversions
  • Tumblr Segmentation – Creating Tumblr segments and analyzing them in comparison to visitors that do not engage on Tumblr

Brand Page Measurement:

Adding measurement to the customized branded pages is as easy as editing the page template itself.  As a logged-in admin of your Tumblr pages you will notice a ‘Customize’ link in the top right:

Customize-Template

From here you will see an ‘EDIT HTML’ button which will allow you to customize the template as it is currently setup:

Edit-Tumblr-HTML

Within the HTML files, you can add the Adobe Analytics page code as downloaded from the code manager in the Admin Console:

add-s_code-to-tumblr

You should then follow your standard page name, site section, and other custom measurement conventions and apply them to the Tumblr page templates.  Doing so will enable the following types of reporting and Analysis as it will now show up as an extension of your digital properties:

Analyze Tumblr as another Site Section:

tumblr-site-sections

Better understand how users flow from Tumblr through to other Site Sections, etc.

Screen Shot 2013-03-13 at 2.54.36 PM

Look at key Tumblr posts/pages that are accounting for most engagement and interactions:

tumblr-pages-report

Owned Post Clickthroughs:

A best practice for Tumblr, as well as all content we publish to social platforms, is to have campaign IDs appended to the end of the links back to owned properties.  Target does an excellent job in providing desirable content to their Tumblr community, but also ensures to append campaign codes to the end of the link URLs to make sure that they are able to look at revenue driving posts from Tumblr.  Here is an example of what I mean:

Target-tumblr-campaign-codes

Adobe Social automatically appends campaign codes for Facebook, Twitter, and Google+.  Look for Tumblr, LinkedIn, and YouTube publishing and automatic campaign codes for these platforms are in near term releases of Adobe Social.

The data will then allow a social marketer / analyst understand how each post is performing with regards to onsite conversion, and also compare each post to each other to benchmark which messages / types of content are working well:

Tumblr-campaign-codes

Tumblr Monitoring:

Adobe Social provides broad listening capabilities across a number of major social platforms and tens of thousands of blogs and forums across the social web.  We recently announced on  January 31 support for monitoring Tumblr as well:

Screen Shot 2013-03-13 at 2.58.51 PM

This helps to ensure that we are aware of all of the conversations and key trending topics on Tumblr and can alert us to different opportunities and/or threats that may arise on Tumblr organically.  You can then filter all the conversations by Tumblr specifically and look at how the conversations differ on this platform compared to others to better inform content/community managers on what types of content will resonate well:

Adobe-Social-Tumblr-Monitoring

Influential ‘Tumblrs':

A key differentiator between Adobe Social and other competitive point solutions is our ability to tie social activity back to tangible business metrics.  One way that we do this is by taking the commonly known ‘social influencer analysis’ to the next level by showing not just followers and Klout scores of social authors, but identifying how much are these authors driving actual onsite conversions.  Similar to our methodology for identifying Twitter authors we are able to find key ‘tumblrs’ that are driving onsite success.  For example, as Adobe monitors Tumblr for the term “Creative Cloud” they would pick up the following post with a link back to Adobe.com:

Adobe-tumblr-post

The author name is ‘lovendbeloved’  and once his followers click through on his link, we are able to start giving credit back to this individual for the traffic and ultimately revenue that he/she is driving:

tumblr-post

In the sample report below you can see Visits and Revenue, but we can also add mentions and sentiment scores for this author as well.  A good calculated metric to include would be revenue per mention to see which authors can get the most bang for their individual posts:

tumblr-influencer-analysis

Tumblr Segmentation:

Up until now, we have spent most of our time collecting data on different Tumblr interactions, but the power of this data really shines through when we leverage the segmentation capabilities of Adobe Analytics (SiteCatalyst v15, and Discover).  To illustrate what I mean, I will use the most basic of all segments:  “Visitors that came from from Tumblr at any point during the last 30 days”, what does their consumption behavior look like relative to those visitors that do not engage on our Tumblr community?

Tumblr-segment-builder

When performing this segmentation analysis for an actual Media & Entertainment company here is what the data shows:

Tumblr-segmentation-analysis

It is showing that Tumblr visitors consume roughly 40% more content per visit and 60% more per visitor than the average site visitor.  They also spend around 1 minute more per visit and are 40% less likely to have a single-page visit than the average site visitor.  This was encouraging for this business to see such a promising social platform.  As you can imagine, even though they were seeing relatively low volumes of traffic coming from Tumblr, it took on a much more strategic tone because they were beginning to understand the VALUE of this segment.

Now, don’t run off to your CMO with the above analysis and ask for a large investment for you to dump into Tumblr.  My recommendation is to 1) Collect the data from Tumblr as outlined above, and 2) Perform your own segmentation analysis and find out how this segment is impacting your own business metrics.

Here are a few examples of other potential segments that can be leveraged for this as well:

Tumblr-segment-filters

In summary, Tumblr is a platform that allows you to get as close to “Value of Social” as any other platform out there.  Whether or not Tumblr is right for you and whether it should be a strategic part of your social marketing program is all up to the data you collect and analysis you perform.  So, get out there, and start implementing the methodologies outlined above and feel free to sound off below on any additional ideas that I may have overlooked.

 

 

 

0 comments