To add to Tracy's question above, the table I am using is jivedw_activity_agg_month. I am querying the value for jivedw_activity_agg_month.activity_count where jivedw_activity_agg_month.activity_type = 160 for the particular user_id. Per jivedw_activity_meta table -
160 | Log in | This is fired every time a user logs in. This includes when an account is first created and when a user bridges a piece of content.
Sangee & Tracy - I see some really high values for some of our super users as well. I went down to the activity_fact table and ran what I call a 'Day in the Life' query where I pull all activities for a user and tried to figure out where all the logins were coming from (I recommend using yourself for testing and limiting the Add-ons you use).
Example: These are 3 logins for my user on 4/1 form one of our instances 2 of them are less than 30min apart...
5377;"Log in";"2015-04-01 14:59:31.318";1
5377;"Log in";"2015-04-01 15:27:13.368";1
5377;"Log in";"2015-04-01 22:24:57.929";1
What I *think* I know is every time a new session is started a new login activity is fired. Sessions can be started in lots of ways just via a browser (incognito tab, new window *sometimes*, etc) and gets even weirder if you have some sort of SSO system involved or a ton of Add-ons in use.
I know that probably doesn't help much but might give you a way to try and zero in on the culprit.
Hi Tracy! It's been so long since I was in the depths of the analytics database, I don't feel confident answering the question. As a starting point, I reviewed Definitions of Analytics Events, but Cory wrote that before any of the connectors you mentioned were available - so not super helpful. I've raised this question internally, so someone should get back to you with a more detailed answer soon.
1 person found this helpful
Hi Tracy, I've gotten an answer here. A "login" is the activation of a new session. This will happen under a variety of scenarios including (but not limited to):
- Logging into the app from a browser
- Launching the Jive mobile app
- Connecting MS Office during configuration of the Jive for Office plugin
- Connecting MS Outlook during configuration of the Jive for Outlook plugin
- Connecting Jive Anywhere during configuration of the plugin
The team also looked at your instance and thought the login activity looked a little fishy. They offered that you can submit a case and Support will take a look for you.
Looking at this again, and so the question was pushed a little further on the J4O items:
- If a user is interacting within J4 Office (i.e. working on or viewing a document), would that count as a login?
- If a user is interacting with streams in J4 Outlook, or replying to a thread from within Outlook, would that count as a login?
And then just to be sure on the mobile app question, does it matter which app - Classic or Daily?
David Bastedo - I know Ben is no longer at Jive. Perhaps you can help me get answers here? Thanks!
I think ultimately what we'd like is a way to unique it per day (i.e. one login per day per user) so that we can then figure out how often people are logging in - every day, once a week, once a month, etc. As it is, there doesn't seem to be a good way to come up with these kinds of stats.
Thanks, Benjamin Walker
1 person found this helpful
We use the query below for unique users preforming activities by day, I hope this helps.
select jivedw_day.begin_ts, jivedw_activity_fact.user_id, jivedw_activity_meta.activity_name, count(DISTINCT jivedw_activity_fact.user_id) as "count" from jivedw_activity_fact left join jivedw_day on jivedw_day.day_id = jivedw_activity_fact.day_id left join jivedw_activity_meta on jivedw_activity_meta.activity_type =jivedw_activity_fact.activity_type where jivedw_day.begin_ts >= '04/01/2015' and jivedw_day.begin_ts <= '04/02/2015' group by jivedw_day.begin_ts, jivedw_activity_fact.user_id, jivedw_activity_meta.activity_name
You can remove the DISTINCT to get a true count of activities by user and day if desired.