Tagging your audience for segmentation


(Sarah Hawk) #1

One thing that Discourse is a bit weak on is functionality to easily segment an audience. Profiles aren’t a big focus (they can be customised but not without involving a dev). I’d like a way to tag interests or passions on the platform so that I can easily call people into topics of interest.

Does anyone do this? Do you get good results?


Tagging and metrics in Discourse
(Alessio Fattorini) #2

I’m been working on it for a while but I didn’t get a definitive solution. Tried with a spreadsheet, groups on discourse and still waiting for discourse staff notes. Nothing, tagging users to segment the audience is better, description notes aren’t so effective. My aim is the same, easily call people into topics of interest.


(Colleen Young) #3

@hawk we just started doing this on the Mayo Clinic Social Media Network. The member profile to date is weak to non-existent and we have an abysmally low completion rate. First step is to redesign the profile which will include pre-determined interest categories. Then we’ll encourage members to complete their profiles to connect with others within same interest area, geographic area, etc. Ultimately, I’d like to visualize a mapping by different segments, but that’ll be for another day.


(Peter Staal) #4

@HAWK I have experience tagging and segmenting members. I use Intercom.io. It is incredible tos tag and segment users based on the number of contributions, pages they visit, videos they watch etc. Based on those segementations it can send aut automated drip campaigns. I highly recommend using it!


(Sarah Hawk) #5

I have staff notes activated here, but I can’t currently find a way to easily pull the data. My other idea is to add custom fields to the profile so that people can add their own tags, but again, I’m not sure how to pull that data.

@colleenyoung Sounds like exactly the strategy that I’m hoping for here. Let me know how it goes.

@Peter_Staal Good to know. Sounds like a no-brainer. Thanks for the input.


(Kristen Gastaldo) #6

We do quite a bit of segmentation. The platform lets us segment by product interest - since we have 18 or so products listed on the same community platform (and they all have their own forum/event/resource/blog pages).

From there we can segment further. I break my community down by job role (marketing/events/database admins/membership etc). I send them topics and resources that are relevant to their field. I always see an uptick in engagement when these go out. If I message them as a group, I let them know it’s because someone else in their field needs advice. They are generally keen to help each other.

It does allow us to segment based on actions as well, so we could message our most active posters, or those who have clicked on a certain resource, or just most active in the space. It’s a missed opportunity that I haven’t done anything with those options yet!


(Sarah Hawk) #7

When do you collect that data from them? At sign-up?


(Kristen Gastaldo) #8

I used to stalk and kept an excel sheet. We had less than 1000 members at that point and only about 30 or so joined every month. You had to buy the product to join the community, so it was manageable stalking.

Now it’s an option in onboarding. It’s not required, but most people add it in.


(Sarah Hawk) #9

That’s my current strategy (the stalking). I’d like to make it more automated because I don’t have the time any more.

On the subject of stalking, I made a horrible mistake once, by actually using the word stalking when someone asked how I knew something about them. It didn’t go down well at all! It was entertaining for one of us though. :confused:


(Sarah Hawk) #10

Update on this. I’ve hacked together a solution using the Staff Notes plugin for Discourse. Ping @ale_fattorini

Here are the steps for tagging and searching for “Discourse”

Step 1: Add Staff Note

Step 2: Query data

Use the data explorer to query the Staff Notes data

Here is that query:

WITH psr AS (
 SELECT regexp_matches(plugin_store_rows.key, '[0-9]+') AS psr_key_value
   , plugin_store_rows.plugin_name AS psr_plugin_name
   , plugin_store_rows.value AS psr_value
 FROM plugin_store_rows
)
SELECT users.id 
 , users.username
 , users.email
FROM users 
JOIN psr
ON
 users.id = CAST (array_to_string(psr.psr_key_value, '') AS int)
WHERE 
  psr.psr_plugin_name LIKE 'staff_notes'
AND  
  psr.psr_value LIKE '%Discourse%'

Step 3: Download results


Re-Engagement Message
(Alessio Fattorini) #11

Wow it seems workable :slight_smile: thank you for your explanation, it’s very clear to me. Data axplorer is within my plugins list while Staff notes not yet, I’m trying to sort it out.


(Alessio Fattorini) #12

Plugins activated, it works like a charm!
Export to CSV is extremely useful for an automatic message by mailchimp as well :slight_smile:


(Alessio Fattorini) #13

Just a question, do you have figure out how to list tags used yet? :slight_smile:
Are you saving them separately?


(Sarah Hawk) #14

Do you mean do I keep a seperate list so that I know what the options are?
No – I just make the tags intuitive.
You can pull the entire list from that query though, by using a wildcard, if you need it.


Re-Engagement Message
(Richard Millington) #15

The work that @edfryed and @Mary_Green were doing here was really impressive at inbound.

I don’t quite know how, but I’d love to figure this out and set aside some budget to develop something if Discourse doesn’t make it happen.

Looking at Mary Green’s re-engagement message I featured the other day, that was a really powerful message. So being able to set something like that up would be an incredible way of boosting activity and getting more useful information out there.

Very open to ideas to fund here.


So, what are you working on?
(Ed Fry) #16

@richard_millington @HAWK We use HubSpot on the backend (we kinda have to given we’re funded by them, but it’s actually alright hehe!) - but you can use any good email tool. I’d look seriously at Intercom.io

The segments you can create are all from data in our database which we just sync across. We can also look at

But the technology isn’t really the hard or interesting part. The messaging is.

Once you get to sending personal, contextual emails on a larger scale you can start to optimise the messaging based on the topic, the type of person (are they a regular contributor? Are they a C-suite employee? What timezone are they in?) and the question you want them to answer.


(Alessio Fattorini) #17

@Bas_van_Leeuwen has already developed a tagging system on his analytics, it looks handy to put and pull data


(Sarah Hawk) #18

The only bit that isn’t happening is the automated creation and sending of the emails, which would be cool.


(Bas van Leeuwen) #19

Thats a good idea! Though currently we don’t offer the results as an exportable CSV, probablly should add that :slight_smile:


(Richard Millington) #20

Does this mean I could send an email to everyone that has participated in platform discussions tomorrow?

I’m curious about the technical side of this. What would be involved in setting it up on discourse.

Do we have to manually tag everyone? What do you do @edfryed ?