Hot Meta Posts: Allow for removal by moderators, and thoughts about future improvements

Question

We have just released a new feature for Moderators across the entire network: the ability to remove a Meta post from showing up in the Hot Meta Posts (HMP) section of the Community Bulletin on the parent site.

Some details about this feature:

• Moving forward, any meta posts that are added to HMP for the associated parent site will have a “Selected for Hot Meta Posts” history entry added (accessible on the timeline and revisions for the meta post). This will only be added once (the first time that the post shows up in HMP).
• Removal of the Meta post from HMP will be through the mod menu for the given post, where a “Remove from hot meta posts” option will be provided for any Meta post that has a “Selected for HMP” entry in its history (in the recent past) and has not yet been removed.
  • This will work identically to the way that the "Remove from hot network questions" button works on main sites.
• A Meta post that has been removed from HMP cannot be selected again by the system to appear in the Bulletin as a Hot Meta Post (it could still show up as Featured).
• When this function is called by a mod, it will:
  • Add a “Removed from HMP” entry to the post timeline (identifying the mod who removed the post).
  • Immediately refresh the cache for the Community Bulletin for the parent site.
• Similarly, when a post on HMP is deleted, the cache for the Community Bulletin on the parent site will now be refreshed right away.

As many of you may know, Hot Meta Posts had been removed from the Community Bulletin on Stack Overflow for over a year, before they were restored a couple of weeks ago. To quote Catija:

What we really want to avoid in the Hot Meta Posts are questions that call out, or are rude towards, specific users, moderators, or staff. Usually these questions don't get upvoted so our concerns are relatively low, but please understand that if we see these in the HMP we may act to reduce their visibility by changing the tags…

This new feature will provide moderators with a much more robust (and transparent) way to remove posts from HMP than has been available to date.

Additionally, now that we will be recording HMP selection (and removal) in post history, we will be better equipped to analyze the effectiveness of Hot Meta Posts in the future, and to make decisions about how we can improve this feature.

The current state of HMP network wide & future discovery

Hot Meta Posts is a very useful way to share important discussions that take place on Meta sites with the wider community. For many users who are looking to be more active in the Community, it can serve as their entry point for meta. That said, while for the main sites with more active and frequent Meta discussions there is almost always something to show in the HMP section of the Community Bulletin, for other sites this feature can go unused.

For example, right now (as I am writing this), Hot Meta Posts are only appearing on 32 out of 174 sites where they could potentially appear. On the other 142 sites, there simply are no recent meta discussions (or feature-requests, for sites where these are allowed on HMP) with a score of at least three that can be shown.

To quote Catija again:

We're planning to schedule more investigation for improvements to Hot Meta Posts network-wide as a project in Q4 2020. When we posted about this change last year, several answers proposed ideas of how to adjust the HMP feature along with questions about the impact of this change that we should consider and we'd like to investigate some of those, particularly thinking about the things that Bhargav wrote in his answer. These are important concerns we need to keep in mind, and I'm looking forward to seeing how we can improve and grow from this rather than just returning to where it was.

So, having the goal in mind of making interesting and useful meta discussions be more discoverable to users on the main sites, what types of content from meta sites would you like to see surfaced on the main sites that is not happening now? How might that content help to create new engagement among users who are more interested in getting involved in the community, in content curation, and in site governance? Are there any pain points that currently exist with this feature (even if it is only on site specific) that you would like to see addressed?

While we cannot make any promises or commitments at this point as to what changes will be made, and when: We are happy to hear thoughts from the wider Stack Exchange community about this, having in mind the effectiveness of this feature on both the sites that currently use it widely, and those who do not. Thank you all for your contributions to this discussion.

Answer 1

What types of content from meta sites would you like to see surfaced on the main sites that is not happening now

Something that cycles through a site's 'faq' tagged posts if there aren't any hot meta posts could be nice.

These posts aren't actively ongoing discussions, but I'd label them interesting and useful meta posts. I think they can provide valuable bits of information to newer users on a site. Especially if the site they're on may be a bit different than what they may be used to (Interpersonal Skills, for example). If such posts were automatically shown, we could get a lot more visibility for our citation expectations, good question and answer checklists and sandbox, for example.

In the past, we've been (ab)using the featured tag to get some of these posts in the sidebar. The problem with that is that if it's overused, it will reduce the amount of hot meta posts that can be shown. Something that automatically picks faq questions when there aren't any hot meta posts would help with that. It will also reduce some workload, users and moderators won't have to find and flag/tag these posts manually.

Answer 2

It might be interesting/useful if meta posts could "become hot" even when they're not new. For instance, the system could perhaps detect "bursts" of upvotes after a sufficiently long period of more-or-less stagnation. This could apply to either posts that were previously hot or posts that never made it to Hot Meta Posts.

Some examples of where this could be beneficial to the network:

• An old feature request that gets new attention due to being linked in a related post, generating new discussion/idea
• A post that gives good advice on a relevant topic (e.g., most of Shog9's old posts) getting new attention and sharing its wisdom with people who haven't seen it before, including those who may not actively browse that site's Meta (and thus are among the least likely to have seen it previously).

Answer 3

Firstly, I would like to thank you for this change. The turnaround time between me asking about this in the Stack Overflow moderator room (Aug 5th) and it being implemented (Aug 27th) is quite remarkable, and almost unprecedented in the recent past.

what types of content from meta sites would you like to see surfaced on the main sites that is not happening now?

There are some of the meta posts that have been posted long back but haven't had any significant views/votes. This might be because they posted it on a weekend when there was no activity or perhaps when there was too much activity and it slipped under the chaos. These posts might necessarily not be bad posts as such. Having a way to bring these to the forefront would be quite helpful. I understand that the community user bumps these regularly, but bumping them to the home page of the per site meta doesn't bring too many views as compared to them being visible on the main site. I don't want to use the "Hot Meta Posts" section for these, but perhaps another section on the Community Bulletin or somewhere else would be helpful.

Are there any pain points that currently exist with this feature (even if it is only on site specific) that you would like to see addressed?

As I mention in the answer you linked, Hot Meta Posts are geared towards posts that get upvotes/answers very quickly, and these might not be the best ones. A burnination request with a punny title would get a lot more votes than a post about how we can handle a particular type of off-topic questions, or something similar. Adding in some questions that are still new, or even those that are downvoted (as voting on meta is different) would perhaps be a nice experiment.

Answer 4

Hot Meta Posts are only appearing on 32 out of 174 sites where they could potentially appear.

Quite frequently, we see users asking questions on Meta Stack Exchange which are more (or only) appropriate on a per-site Meta. This is due to the fact that they don't know each site has its own Meta. The existing ways to discover per-site Metas (via the 'hamburger' menu, or the Feedback link in the footer) are harder to find than Hot Meta Posts, which are only one click and no scrolling away from the home page.

Wouldn't it be good to extend the period and/or lower the threshold for Hot Meta Posts, so that users are more likely to learn about the existence of per-site Metas? One could say those don't deserve the 'hot' label anymore, but it's not like the hot questions tab contains only stellar questions. Also, IMHO good support questions should be candidates for Hot Meta Posts; they might be about problems many other users may have too, and are useful to teach tips & tricks about using the website.

Answer 5

This is a good change - having accurate data should enable much more nuanced adjustment of this feature in the future.

Rotation and randomness

I have one question: how does this play with rotation? Previously, a question could be eligible for the list for a period of time, but only displayed intermittently during that time, as the set of questions displayed would be chosen randomly at regular periods from the set of eligible questions.

This is important, as it helps to avoid the list becoming static (and readers becoming blind to it). I am assuming that this behavior remains, with the "selected" record added the first time an eligible question is shown, but would be good to see that clarified.

Potential for the feedback loop

There's currently an age restriction on questions that can be "hot": 2 weeks for most sites, 3 days for SO. That restriction exists to constrain the positive feedback loop that normally goes along with the extra attention (question is considered hot based on votes, gets shown more, gets more votes, is considered "hotter", gets shown more, ...)

Constraining positive feedback loops is extremely important - however, the age of the question is only a rough proxy here, based on the assumption that questions will become "hot" soon after they are created. As Bhargav, fedorqui and others have noted, this assumption does not always hold - and relying on it can limit the attention that can be given to important discussions.

Now that you're tracking selection, there's less need to rely on heuristics to limit the feedback loop here: you can base the length of time a question is eligible on the time when it was first selected instead. This has the potential to greatly expand visibility and utility of crucial discussions across the network, regardless of site size and community behavior.

Answer 6

Posts are hot and then become hotter by appearing in the sidebar. However, in small sites this initial hotness is difficult to get, so it is common that a debate never shows in the sidebar.

What criteria are used to select the links that appear in the community bulletin sidebar block? explains:

The rules are as follows:

• ...
• If there are still under four items, the rest of the space is filled with hot discussion questions not marked with any moderator status tag, scoring at least 3 and posted within the past two weeks. These are picked semi-randomly. (Note that the timing, score and tags can be adjusted per-site to suit the needs of each community).

In small sites, the problem comes with the scoring at least 3 and posted within the past two weeks part: it can be rare that a post reaches this score of 3, since not many users visit their Meta. So I would suggest allowing to decrease this number so it is easier for a post to make it to the sidebar. Some comments in the question should signal a debate, so they could be used as a parameter to take into account.

Answer 7

First of all I want to thank you for your engagement for doing changes at HMPs. The features added seem good and helpful.

"Are there any pain points that currently exist with this feature (even if it is only on site specific) that you would like to see addressed?"

One big pain point for me is that there are only such less posts addressed as HMP in the Community Bulletin (facing specially Stack Overflow but might be elsewhere, too).

I think there can only be max. 2 or 3 posts marked as HMP simultaneously which stay there then for about 1 or 2 weeks (my guess), which is IMHO problematic. Many posts which deserve being marked as HMP will get lost on the Meta site and many user who are not confirm about Meta or are just not so interested in Meta in general have no ability to share their opinion on these posts, which is bad.

I do not say I want hundreds of HMPs but just a bit more (best would be 10, fine would be at least 5 at the same time).

If the HMPs section in the Community Bulletin has too less space available, maybe think about to implement a special page where all HMPs are listed.

Or another solution as mentioned by @Trilarion in the comments would be to have a list of HMPs and just show 3 of them, cycling through the list each time the page is refreshed.

Any of such an improvement on that point would be a big improvement in my eyes.

Answer 8

Today I discovered an edge-case that you might want to take into account in future updates.

When a post that is currently a HMP gets migrated the cache of the HMP isn't refreshed automatically. This causes a period of time in which the post is a HMP, but no longer on the site-specific meta (in this example).

Perhaps you can remove the post from HMP automatically on migration.

Answer 9

"So, having the goal in mind of making interesting and useful meta discussions be more discoverable to users on the main sites, what types of content from meta sites would you like to see surfaced on the main sites that is not happening now?"

It is happening right now, but IMHO to less and rarely. feature-requests.

I think at a specific threshold of upvotes, a feature request should become a HMP.

"How might that content help to create new engagement among users who are more interested in getting involved in the community, in content curation, and in site governance?"

Providing user the ability to see more that they actually can have an impact on the site's features would I think have a good influence that they participate on discussions on the Meta sites more regularly.

And more experience and more opinions make also the discussion about adding a specific feature and - at the end of the day - the site better.

Answer 10

support status-bydesign

I'm seeing a grayed out 'Remove from hot meta posts' in the mod menu for this post, even though it is a Hot Meta Post right now:

Answer 11

All around a good change, I think. Certainly a good change that HMP was re-enabled on SO. Thanks again for that.

This is sort of a question out of curiosity mostly. As mentioned here about down-votes:

If downvoting is made more difficult, then upvoting would need to be made correspondingly more difficult.

The point being here is that there is that the two actions should be reciprocated.

I'm sure discussing down-votes here seems totally out of place here, but let me try to articulate my point. If Moderators have the ability to remove posts from being "hot," should they have the ability to do the opposite action? That is, have you considered giving them the ability to make a post hot that otherwise wouldn't be hot (for whatever reason)?

I'm sure most moderators have better things to do with their time than look at every single discussion on meta and decide whether it is correctly tagged as hot/not hot, but I'd imagine it'd be a useful tool to have if the need to use it ever did arise.

TL;DR: Have you considered the possibility of giving moderators the ability to do the reverse action? Would you mind sharing your thoughts on it?

Answer 12

with a score of at least three that can be shown.

Is the selection criteria "to only allow the hottest posts"? AFAIK, some randomness was introduced to prevent the posts from going stale and only two-three topics dominating the community bulletin, but I'm not seeing that on Stack Overflow. The same 2-3 posts have been static in the past 7 days. Can we introduce some randomness network wide?