Some of the KPI's and metrics we measure include the following:
- All the metrics provided in the Community Manager Reports plug-in
- Drop-Out rate: The % of total community members/ followers with no activity in the past 90 days.
- Creator-rate: The % of members/followers with at least one "create" activity in the past 90 days.
- Avg. # of activities per user per day
- Proportion of View Activity vs. Create Activity
- Cross-Pollination levels: Proportion of discussions with replies from users residing in 1 location, 2 to 3 locations, 4-6 locations and 7 to 10 locations (we use this as an intranet where we are interested in the cross-department/market user collaboration levels)
These are really good metrics. For the last one about cross-pollenation, is that a report you built in the SAP BI OD analytics module? I'd love to hear more about how you were able to set that up, if you're willing to share.
I did not use the analytics module but if that is able to execute raw SQL queries against the Analytics Data Warehouse then you should be able to replicate.
In our case each user has a set of profile fields that are pulled in from our LDAP directory when the user registers so all users have a location and market profile field that is filled out for more than 90% of users because they are added to LDAP when employed. If you do not have something similar to populate the fields your challenge will be to get users to complete their profile fields so that you can extract meaningful/accurate results.
I will post the queries as a blog entry in a bit but basically it works like this:
1. 2 temp tables are created one for all user locations and the second for user markets
2. The select query returns all the content (blogs, discussions, documents, etc...) created over a user specified date range
3. The query has a join on the location and market tables to add the location/market columns to content data for all user create activity that occurred for each piece of content (ie. replies)
3. The select query also returns an aggregate sum of the activity count, # unique locations, # unique markets, # unique users
Then, in my front end report which I built using a Microsoft BI tool called SSRS but the same calculations could be done in Excel if you brought in the dataset, I wrote some logic to create scorecards (amount of content with >5 locations >10 etc..) amongst other scorecard metrics that can be calculated from the returned dataset.