Tracking Engagement

I like to take a Quarter, Month-over-Month, and Daily (last 30 days) view of a few key metrics. I then lay them out left to right: Quarter, Monthly, Daily in rows per metric

• Active Users = Unique users with a visit to our community in that timeframe (our private community requires a login, so all visits are tied to specific users)
• Visits = Cumulative total of all visits
• Contributing Users = Users with an activity on the community (excluding visit); they took an action (like/post/vote/reply/marked answer) and contributed either content or a data point (likes, votes, answers) on that content. 

Then we have views of a few things by Account (company) in the last 90 days

• Active Users per Account 
• Total Contributions per Account 

Then we have a few tables of top users in the last 90 days

• Top visitors (total visits for that user)
• Top contributors (total contributions for that user)

I also like to use a metric I call “Recurring User,” although that had to be calculated in Tableau based on the data logged via our SFDC integration. Recurring User is defined for us as a user who visits the community on three separate days within that month. 

Additionally, I calculate Visits Per User = Total Visits / Active Users in that time frame. 

I feel it gives me an engagement funnel of three levels:  

1. Active = visiting and logging in
2. Recurring = coming back regularly
3. Contributing = adding some kind of value/activity 

And Visits Per User is a good pair with those three levels of the funnel. 

I’ll eventually add a 4th level related to those who post/reply, but Insided’s “Users with at least 1 post” will give you that number if you need/want it. 

*Important note: The Salesforce integration only writes activities for community records it can match to a contact. In order for all of those metrics (and our dashboards) to be accurate, our IT team had to find a way to create contacts in our Salesforce instance for community member records (emails) that weren’t in our system. 

Tracking Growth

I also have a a Growth dashboard to show registrations over time and per Account. However, I use that one far less. Total users is a “noisy” metric as users will become inactive due to job change, etc. Instead I use my Quarterly Active Users as a proxy for how “big” my community actually is. 

That said, it is to look at trends on how your community is growing. But Insided’s user dashboard will give you some decent baseline stuff there. 

Dashboard Best Practices

On most of those metrics outlined above, I use simple column charts with the x-axis as either Quarter, Month, or Company. While there are fancier visualization tools, I’ve seen best practices in the data/analytics community that columns are often the most effective to compare values. Our eyes can clearly see the difference between the two and it’s often easier to read. 

For daily metrics, I like to use line charts as a way to show the trend. That said, those are a bit messy in Salesforce dashboards, especially over a long range of time. Your better tool there is using Google Analytics. 

As I mentioned at the top, laying them out left to right is a good way to compare the same metric over different time intervals. Most western cultures read left-to-right, so it’s clearer and easier than stacking the same metric in columns on the page. And I like to start with the highest level (quarter) on the left and then narrow it down with Month in the center and Daily on the Right. Finally, I set expecatiations with our leadership that we’ll goal set to quarter numbers, benchmark and report monthly, and monitor daily trends to ensure month and quarter numbers are on the right track. 

@Kenneth R reply above puts on “paper” what I was explaining to you earlier today during our check-in.

This is awesome @DannyPancratz - I love how you’ve been thinking about recurring and contributing users as part of your engagement funnel.  I’ll be curious to see how you eventually design your 4th (and beyond?) level of the funnel.  

Having a deep dive walkthrough webinar or class for this would be amazing and something I’m sure a TON of people would take a lot of value out of! 

@DannyPancratz We are working through our Salesforce integration so that we can track deeper segmented activity. I know that there are ‘historical data’ limitations with the Salesforce integration. What has been your experience with setting up dashboards within SFDC? Any things I should keep in mind. 

@allison_able 

Talk to support and/or your CSM. Historical data can be a problem, but they were very supportive of us as we got our integration and reporting up-and-running.

Specifically, they were able to do a full sync of historical data once or twice as we ironed out the kinks. That gave us all our data (comes with the first sync) and not just net-new. 

One of those re-syncs came after this first consideration: 

1. Automate creating a contact in SFDC when someone signs up (only for if they don’t already exist) 
   
   This ensures the Insided record has a match and all their activities will sync going forward from day 1 of when they join.
   
   The one challenge is this requires me to spend 5-30 minutes a way cleaning up SDFC data (matching them to accounts)
    
2. Your dashboards will be as good as your SFDC data. If you want to segment by Account Type, Account, or other attributes, you can! But you have to ensure the community record is matched to a contact record with all those features. 
    
3. We do most of our reporting based on rolling up user activities. Insided will also push the summary and latest data (the types of things you get in .csv exports, last login etc), but the real value is in the many additional data points (activities like visits, likes, votes, marking answers, etc) that you get. 
   
   Note: Visits is an awesome activity. Login cookies can last days, visits actually show you how frequently they visit in a day/month/quarter/year. I don’t use logins for anything; I use visits. 
   
   Thus, we’ll roll up activities in reports with filters that can show them over time. 
   
   For example, our MAUs are all activities (really only need visits) rolled up and counting unique users within that month. Our recurring users are those with activities on three different days within the month. Our contributing users filter OUT visits and show those who take actual actions like posting, liking, voting, marking answers. 
    
4. You can use the same reports to build different charts for the same metrics in different time intervals. Since most cultures read left-to-right, I put comparable metrics (active/recurring/contributing left-to-right to show the engagement).
    

   

   Next row below this is Quarterly sign ups → monthly signups
   
   then the same thing as top row but the quarter view (monthly row is at the top, as I’ll check that more/report on it more often)
   
   Next is a series of bar charts showing Active, Recurring, Contributing, and Signups by company (these are nice and scrollable down)
   
   Next is Active Visits by quarter, month, and day
   
   Finally, my last row is % of Total Question replies by Account Type: by quarter, by month
    

5. Subtitles are helpful to explain your definitions like in my image above
    

6. One dashboard can do the trick if you can add in filters for things like Account Type / Account, but that will depend on your licensing. 
    

7. A helpful filter is filtering the full dashboard to show only that for the previous month. It’s especially helpful for EOM/EOQ reporting. Whether live or in screen shots; you want your last data point in the charts to be the one you’re talking about. If you’re doing that with the current month showing, all your charts will be pointing down. This filter gives our a clean view. 

I agree Contributing Users is a good metric. Is there a way to see this data in Insided/Gainsight Or do you have to import all the data into SFDC?

@Wilder-Clari right now you can only get all of those things via SFDC. BUT they’re working on making that same data available via a Data Lake feature in the future (this quarter or next, I think is the estimated timeline)