Effectively reporting on community metrics is a continued challenge for many Jive customers. We’ve taken new approach of using big data analytics tools that we normally use for IT data, have seen some amazing results. Skip to the “Playing with data” section and absolutely watch the video to see the results. If you are interested in my experiences leading up to this, start from the beginning.
Reporting is hard
I’ve been working with Jive as a user, consultant and developer for 6 years now, and one of the most persistent topics in my own work and in discussions with other Jive customers is (you guessed it): Reporting and Analytics.
When you are running a platform like Jive knowing what is happening in your community is relevant on many different levels:
- Community Managers need to know which groups are active and which can be retired to keep the environment relevant and less cluttered
- Your marketing team wants to know which product areas are most frequented and discussed
- The customer support team wants to know if their FAQ articles are found and reach the right audience
- Sales enablement would likely kill if they could find out how a successful sales rep’s usage of Jive differs from an unsuccessful one
The list goes on, and I’m sure anyone who is responsible for a Jive powered community has had similar requests.
Now, all reporting solutions for Jive that I’ve experienced so far (Community Reports, Analytics, Business Object powered analytics, Community Manager Reports, JBA, Resonata and a myriad of home grown solutions customers have built) have one thing in common:
They make assumptions on what data you need/want.
Don’t get me wrong, I am a big fan best practices in the form of out of the box reports. However, it always seems that in my day to day work with our own data and customers, something is missing. So it seems I’m constantly exporting data to Excel or running database queries to get to the needed.
Example: The places activity CMR is great, but if I need to know who is using a group that I want to retire, I can assume it’s the group owner and ask them (more often than not, they don’t know either), or I can start running SQL queries against the Jive database. Not ideal, especially if this is a recurring task.
Likewise: The new Jive Business Analytics looks awesome, but what if my customers have a different opinion on success criteria?
A different approach
My path to Jive reporting and analytics nirvana (at least for me) started when we started working with the new data export API available in Jive 7 and Cloud. We ported the functionality to a plugin for a customer running an air gapped on premise installation of Jive.
To work with and compare the exported activity data I used one of the tools we use in our IT operations, called Splunk.
If you are not familiar with Splunk, it is a data analytics tool, typically used in IT and security operations to manage large amounts of machine data. It is very intelligent in regards to making sense of unstructured data and works amazingly well even with large datasets (think billions of records and terabytes of data not uncommon in IT).
So after creating a data connector between Splunk and the Jive data export service, it became clear very quickly that this was the answer to the vast majority of my reporting and analytics requirements.
Playing with data
So let’s get started. Here we have an overview dashboard that presents some community stats for me.
A global map of community activity, a high level health indicator (based on the level of activity compared to the average level of activity), and an overview report showing which activity happens where.
There are two things about Splunk that have changed how I go about working with community data dramatically:
• You always have access to the underlying raw data and you can always drill down to it
• You can always change the source data for a report on the fly, no need for report building, SQL and Excel sheets
So in the example above I can see in the Place Activity report that most of my activity is happening in social groups. Let’s say I want to find the social groups with the least activity, so I can start retiring them. I can do this in two clicks:
• Click on the socialgroup link (which takes me to the underlying data)
• View the groups with the lowest activity
Since I do have access the raw data, I can actually take things a step further and click on the group with the lowest activity:
So now I can directly see that Anna from the HR department was the only one using this group recently, and I can now reach out to her. Or I can use the Jive Anywhere cartridge we’ve built for Splunk to share this report with my colleagues. All of this within 3 clicks.
Of course this is just one example, but this complete flexibility has been amazing.
For more formal reporting, we’ve also created data models on top of the raw Jive data.
Let’s say I want to analyze content creation and consumption to figure out which departments create the most content and which departments consume it.
Create a new pivot using the content activity data model
I start out with a new pivot dataset holding all events
Add additional data points (Department and Activity Type)
Activity Types added
Switching to bar chart for easier visual analysis
All reports remain full interactive for drill down, and can be added to a dashboard scheduled and emailed if necessary
And to blow your mind completely: This data is actually coming from two separate internal Jive communities for one organization, seamlessly combined. So if you have more than one community for historical or other reasons, like many companies do, you can now elegantly combine your data.
I hope this was of interest to you. The examples I’ve included are clearly only scratching the surface of what’s possible, but as I mentioned above, my approach on how to report and analyze Jive community metrics has improved dramatically. We’ve also started enriching the Jive data
We are currently working to package this, so it can be shared. If you are interested in leveraging this integration for yourself, please feel free to reach out to me.
I've started using Splunk to analyze and report on Jive data. It is awesome.