What Data Would You Want From Hundreds Of Online Communities?

(Richard Millington) #1

We’re working on an ambitious project to extract useful insights from a huge data-set of hundreds of online communities.

A quick question for you all, what would you want to know?

If you could get any benchmarks you like from hundreds of online communities, what would you love to know? Average time to first response? Active members? etc…

Please be as ambitious as you like. What data would most help you in your work that doesn’t exist already?

(Mark Baldwin) #2

Is there any evidence to suggest that certain groups of people are more valuable to communities than others? For example, does age, gender or occupation have a noticeable impact across all types of communities?

Is language a massive barrier to engagement when you have a worldwide community?

(Nick Emmett) #3

What sort of evidence exists, and what benchmarks are there to aim for, in terms of community members spending habits, in terms of cross/upsell, churn reduction etc. you read a lot about this in various places and to some extent it’s a reasonably obvious statement that those that are engaged with your brand/company will spend with you, but it would be good to see how this looks across the industry rather than specifics from various organisations.

(Richard Millington) #4

Love the idea Nick.

I think though it’s best to assume that while we can get all their
community data, we probably can’t get their financial data :slight_smile:

(Nick Emmett) #5

If we don’t ask, we don’t get :slight_smile:

(Sarah Hawk) #6

I often get asked for growth benchmarks at specific stages of the lifecycle. E.g. how many members should I have at the 1 year mark? What should my growth rate be at the 2 year mark?

(Mark Williams) #7
  1. I’d be curious on the overall trend/health of the on-domain “forum” ecosystem since the rise of “social media”. If you have enough data on multiple communities over a long enough timeline, it would be interesting to see a general trend of “participation”. It would be even more interesting to splice by industry cohorts. I’m intentionally leaving the specific definition of participation, cohort and specific time points to the analyst.

  2. Is there a general type of “helpful” user across multiple types of communities. What does that user look like and can we identify specific measurable features that would make it stand out? This ties in to my long desired ability to identify the ascenders in my users while they are still early in the process of becoming “super-users/VIPs/etc”, especially in large user sets.

(Shreyas) #8

Apart from all the great questions that are already here, it would be interesting is to know the time frame for first action and hopefully see some similarity or pattern in that.

Example: In Communities of Interest, the general time frame between a person signing up to making the first post is x hours/days, whereas in B2C communities it’s y hours/days. This might be because of z.

Another data point might be related to platforms. While I understand that we must choose the platfrom depending upon our community, it would be good to have an overall idea, especially for new community managers to see some examples of what platforms are available and where do they fit in.

Example: X communities use Discourse to interact with their online community. 80% of Discourse communities are y.

(Anton) #9
  1. Characteristics common to all topics that last longer and engage most people.
  2. Characteristics common to the first paragraph of all topics that get explosion in terms getting first responses very soon after topic creation.
  3. What traits of your community are publicly stated as good/competitive when compared to the communitites which the authors recently left (and moved to yours). For example, in our community people often talk publicly that our forum gives them the most positive feelings compared to all our competitors.

(Bas van Leeuwen) #10

This is a very interesting idea. What type of characteristics would you think have a postitive impact?

I’d be wary of concluding that Buzzfeedy titles and hooks are the key to an active community (it’s always easy to over-optimize). But it could be interesting to see the difference between sentences with/without questions. Or the difference between long and short opening posts :slight_smile:

(Anton) #11

I suspect this:

  • thought-provoking from the very first sentence; asking some question that would not otherwise come to everyone’s mind immediatey;
  • even though the topic needs to be pretty specific, it should still give enough room for discussion to many people about many various cases

For example, compare these versions of the very first sentence:

There is a milk sterilization method that is not thermal yet kills all bacteria.


UV-filtering of milk is the way to sterilize it without using high temperatures

I’d use the first one, as it does not answer the question immediately, yet is strong enough to let you read further. The more times you can let someone read further, the more engagement will happen - that’s my guess though.

My thinking here is that if one does not like or knows nothing about UV, he might stop reading if he is not in the mood of learning something new. So, better not to tell about UV immediately, but rather make some appetite for some new knowledge first.

(Bas van Leeuwen) #12

That is really quite interesting :slight_smile: And something that I’d love to learn as well!

(though I suspect it would be a massive challenge to properly quantify this)

(Erlend Sogge Heggen) #13
  • Out of every user who has posted at least once, what proportion of users has 2 posts or more?
  • When does a user’s 2nd post usually happen? A day/week/month after the first post?
  • What proportion of newly registered came back at least one more time the next day/week/month?
  • What proportion of registered users are online every day/week/month?
  • What’s the average lurker-to-registered user ratio?
  • What’s the average time spent online on the community platform compared to other website assets (i.e. the “time spent online” stat of your community.example.com site compared to example.com, app.example.com, shop.example.com etc.)
  • What appears to be the minimum amount of active users necessary for a community to function? (very ambiguous, I know!)
  • What’s the retention rate of non-staff first-responders? How long do they usually stay at the top of the “leaderboard”?


  • What proportion of community users are on mobile?
  • How many people write posts on mobile?
  • How many people use any kind of formatting on mobile? (again, a tricky one, but it could shed some light on a disagreement that could use some data ;))