Tracking individual posts that contain certain content

Product Support

AI-generated summary

The discussion revolves around the challenge of tracking individual posts that contain certain content on a Discourse support forum. Perry_Eising [(/t/-/145884/1)] is struggling to count and analyze posts requesting the same staff intervention across multiple topics.

codinghorror [(/t/-/145884/2)] suggests using the Assign plugin to address this issue, but notes that a human assessment is necessary to identify repetitive support issues. They recommend setting up a FAQ topic and linking users to it.

sam [(/t/-/145884/3)] points out that most Discourse systems operate at a topic level, and suggests using post notices as a workaround. This would allow for data explorer queries to identify similar posts. However, making assign a per-post thing would require significant changes with limited benefits.

codinghorror [(/t/-/145884/4)] agrees that per-post assignment would incentivize megatopics, which is undesirable.

1

Hello all,

not 100% sure this belongs in the support category - there may be a technical solution for this, but also it’s a bit of a philosophical discussion.

We use discourse as a support forum. We use tags to tell whether a post needs an answer, or what its status is generally. This has been working well.

One thing we are trying to accomplish, and haven’t yet found a solution, is to easily keep track of requests (from our customers) to our team to perform a certain action for them they cannot accomplish on their own. Let’s say 5 people per month all request the same staff intervention that takes an average support engineer 10 minutes to complete.

We use data on these kinds of interactions to assess how important they are for the engineering team to dedicate work towards improving or implementing in a self-serve way through the UI. If its a request we only get rarely, we wouldn’t rush to fix it. If we got it all the time, we would.

If these requests are made individually as separate topics, they are fairly straightforward to count, but, if 10 users in 8 different threads all make the same request, we have no straightforward way to count the posts that ask for the same thing (except manually, which is prone to error.)

Or, if a thread starts as a conversation about something different and then suddenly a request for the staff action somehow surfaces.

If we had a way of counting, tagging, or otherwise singling out an individual post, we could find a way to somehow count that data.

I know this is maybe a niche scenario, but we can’t be the only ones wresting with this. I’d be intrigued by any suggestions as to how we can count/automate or otherwise capture this data!

many thanks.

2

I think the Assign plugin might possibly be a good choice here? Then you can assign the topic to a staff member for followup?

I think a human (a staff member) needs to assess the level of repetition in the support community and take action. What we typically do here on meta is set up a FAQ / howto topic for support issues that come up a lot, then link people to that topic.

The even deeper fix is to improve the product (Discourse!) so the questions don’t come up as often :wink:

3

We have very few systems that operate on a post level vs a topic level and are shareable.

Assign is at a topic level. Bookmark is not shareable.

One interesting trick you could possibly use is post notices.

image
image1319×377 47.6 KB

That operates at a post level and you could use a data explorer query to fish out all the posts with similar notices.

We have discussed ideas of making assign a per post thing, but there would be enormous amount of change needed with limited positive benefits.

4

This would indirectly incentivize megatopics as well, so it’s doubly bad. More topics = more betterer