How do you measure active members?

(MHCommMgr) #1

I’m having trouble figuring out a good metric for active members. I should probably look around more on Feverbee to see if I can find something.

Can you use negativity in a positive way?
(Sarah Hawk) #2

Can you elaborate? Do you mean you’re wondering how best to define someone as active?

(MHCommMgr) #3

Well, we should probably ask @Darren_Gough what he meant by active members?

Right now I measure: overall traffic across all communities, posts per community, and posts per ambassador. I’ve had trouble working with our product team to figure out what else can be measured using GA.

(Sarah Hawk) #4

What platform are you using?

Active members is fairly commonly defined as someone having made at least one post in the last month.
Some people also count likes/upvotes as activity.

(MHCommMgr) #5

We have our own platform. It has upvoting, not sure how to track it. It’s a new feature so it may be trackable. I’m not sure anyone uses it besides me.

I can track posts made in a month, but not in an aggregate way.

(Richard Millington) #6

Yes this is a tricky one really.

People tend to categorise it in 2 ways.

  1. The total number of people that made a contribution (usually a post) within the past month.

  2. The total number of people that visited the community in the last month.

I tend to believe the latter are lurkers more than active members. So the former interests me the most.

Several ways you can do this, first just pull a list from the database/copying and pasting the list of latest comments/posts etc…of the number of people that posted a comment in the past month and remove the duplicates. This is a pretty easy thing to do in excel.

Second measure it for, two weeks. Individually count the number of unique members who posted (takes a little more time, especially in a very large community), and multiply it by two. You’ll probably need to get a sense of the duplicates in there overall, but it gives you a rough number to work with.

(Darren Gough) #7

@richard_millington has done a great job breaking this down.

(MHCommMgr) #8

If I can measure posts per community in any given period through my admin, this sounds like the same thing. Yes?

And then just putting it into a Doc. So this I am already doing, and we have been doing for some time. And it is a KPI that’s falling.

(Sarah Hawk) #9

Do you mean posts per community member (ie the number of posts per individual) or are you tracking multiple communities?

The former will give you active members. The latter would need to be divided by the number of members.

(MHCommMgr) #10

I am tracking multiple communities (we have over 300 but I am focusing on just 30). I am tracking posts in each.

I am also tracking posts per ambassador.

Though I can see who is a top poster every month, we have hundreds if not thousands of members so it would be tough to track posts per member.

(Sarah Hawk) #11

It doesn’t need to be the specific number of posts per member, just the average.

So total posts/total no of active members

(Alessio Fattorini) #12

I disagree, passive members could be good readers and active members in the future. They consume contents anyway.

(Alessio Fattorini) #13

Discourse Trust Levels are very good in this, counting just posts isn’t enough. People can post a lot and disappear after a few weeks
Great members have to read/write/visit

(MHCommMgr) #14

OK, how are you finding the total active members? Using user data from GA? Or literally counting people who visited and made a post? Our admin actually automatically joins people to a community but I’m not sure that data is captured anywhere. I can investigate tho.

We likely have a lot of lurkers in some communities as our boards have been around so long they contain a lot of useful information.

(MHCommMgr) #15

Is there a way to estimate Discourse Trust Level type data using other resources? We have our own platform. Or can I somehow link Discourse to it? Not sure if it’s an analytics resource like GA or MixPanel.

(Sarah Hawk) #16

I use a couple of reporting packages ( and @DiscourseMetrics) which make it quick, but in the past I have done it more manually by downloading all member stats for the month (from Discourse, but you could do the same with GA), ordering by post count and then totalling those >0.

You could even do it from a very basic stats page like this, if you have one.

Discourse is actually this community platform, and it has its own built in analytics. You can read more about Trust Levels here.