On every Stack Exchange site, there is a Community user bot that likes to randomly bump unanswered questions. Based on my observations, we got an almost consistent hourly bump every day, and that might dilute the "active" tab and the homepage from more recently active questions with very old questions that usually cannot be evaluated anymore due to a few reasons.

As per the FAQ about question bumping, 2 factors can be adjusted on this bot:

  • The maximum number of unanswered questions bumped per hour depends on the site: 4 per hour on Stack Overflow, 2 per hour on Super User and Server Fault, and 1 per hour on Meta Stack Exchange and all other sites.
  • Some sites have limits on how many auto-bumped questions can appear on the homepage. Once the limit is reached, no more questions will be bumped until one or more bumped questions disappears off the homepage.

As for the maximum number of unanswered questions bumped per hour, I have confirmed with the Community Manager team that the minimum is 1, so it is already the minimum possible number.

However, the other one about the limit on auto-bumped questions on the homepage is currently not set. Based on my observations, out of 50 questions, our homepage usually had at least 20+ from community bumping only. That was almost 40%-50% of old questions filling the homepage.

I recently tried to keep up with all the bumped questions to check if I can evaluate them. Unfortunately, in addition to the lack of knowledge, the fact that Android is so fragmented made it hard to evaluate, since one device/Android version may behave very differently from another.

So, should the community bot reduce the frequency of bumping unanswered questions?

This is somewhat related to a discussion in 2013, What should we do with older questions that just aren't relevant any more?. While we may revisit the discussion to tackle this issue in the future, existing questions will still be bumped by the bot.

  • 1
    My vote is in favour of reducing the frequency. I am inclined to reduce it to atmost "One" autobump on homepage. Until that is removed from the page, no more bumps. I'm interested in knowing how many of these auto-bumped questions get answered on a daily, weekly and monthly basis. Sadly, I don't know how to construct a query on SE's Data Explorer to get this data. :(
    – Firelord Mod
    Commented Dec 22, 2021 at 12:25
  • 1
    Eagle-eye :) Good find. 20 out of 50 is definitely overdoing it. I'd not go that extremely down as Firelord, but "more" isn't always better. Up to 5 I'd find OK (that's about 10% then).
    – Izzy Mod
    Commented Dec 22, 2021 at 12:35
  • 1
    That also would be great to reduce the frequency of Roomba: Decrease the required view count to prevent Roomba deletion when score = 0 on smaller sites, since many questions here have score = 0. Commented Dec 28, 2021 at 5:54

2 Answers 2


So, first of all, the number of posts you have bumped and active on the front page is certainly concerning and, for that reason, I'm willing to limit this but I want to talk with y'all a bit about why these posts are getting bumped and what that means for you as a community and your curation of the content here.

Back in October 2021 I wrote an answer on MSO about when and why the Community user bumps posts and how community members should respond to that. I'd like to invite you to read it through completely but I'll quote some of the key points here.

Why does Community Bump posts?

First, let's talk about why Community is bumping these posts in the first place:

Community bumping is something that I think has been confusing to people for a while. There's questions on MSE about it from 2011 and 2013 and even back in 2016, Shog9 started a discussion to ask How can we make the purpose of Community "bumping" more obvious? - in it he states:

I was discussing this with my esteemed colleague jmac the other day, and it occurred to me that we never actually hint at what we want folks to do when these questions are bumped.

To be clear, the intent here is to resurface questions that someone has attempted to answer, but which haven't yet attracted any votes to either confirm the usefulness or decry the worthlessness of the answer(s) that've been posted. Q&A that, above all, needs feedback.

...But we don't really say this anywhere. And I strongly suspect that an awful lot of folks viewing these questions just shrug and move on.

And... well... I think that's maybe part of the struggle I'm seeing in your question - you're frustrated (understandably) that you keep seeing posts being repeatedly bumped by Community without really understanding why we do this. Now, unfortunately, while there were some proposed solutions, I don't think we ever actually followed through on this.

When I visit your front page, I see what you do - a sea of older posts that have been bumped by the Community user and I agree - that's kinda scary.

Right now, y'all have 22,684 questions with no upvoted or accepted answers of which, 13,714 questions have no answers (and aren't eligible for bumping) which leaves up to 8970 questions eligible for bumping (there's likely some in this group that don't have any answers with a score of at least 0, so the list of eligible questions is likely somewhat lower than this).

That's... a lot. But what does it mean? We've talked about why we bump questions at all... what makes a question eligible for bumping? So...

What makes a question eligible for bumping?

It's likely that you specifically (Andrew T.) know the answer to this but, allow me to explain because it will help me outline what I think y'all should do moving forward.

From the MSO post:

First it checks the last activity date of open, non-deleted questions to see how recently Community has bumped the question (if ever). The question must be at least 30 days old and can not have been bumped (or otherwise modified) by Community in the last 120 days. This means that a question can only be bumped at most, three times per year.

From that, it selects the top 100 question by views that that meet the following requirements:

  • not recently active
  • not closed
  • not deleted
  • score of >=0
  • no accepted answer
  • is answered and the answer/s:
    • are not deleted
    • have a max score = 0 (meaning at least one answer must have a score of zero but no answers may have a score >0).

... and then it randomly selects the ones it's going to bump from these 100...

I know that y'all should be able to create an SEDE query using these limiters to identify exactly how many questions there are, if you like. I'm just somewhat useless at SQL.

Some stats

When I was digging in internal SEDE to find a query that could identify a list of the questions that qualify for bumping (which I didn't find), I did find a query that Shog wrote that identifies the impact of bumping on a question - how many bump events happened, what percentage of those bumps were "effective" (meaning they led to an upvote on an answer or new answer), and how many received a vote or answer.

I ran the query network-wide and identified that there's 21 sites sitting about where you are in posts bumped. Based on math, it seems that the most number of posts that can be bumped in the course of a year is 8760 (365*24) unless that site has more than one post being bumped per hour. So, out of that max, you've had posts bumped 8196 times in the last year.

So, again, y'all are being inundated and, unlike many of the other sites, your questions per day is likely lower, so the number of questions being bumped is significantly more visible. As far as effectiveness of bumps goes - since we don't adequately explain why we bump, it's generally low everywhere - for sites in the 8.1k-8.3k/year they range in effectiveness from 1.4% to 20.5%. Y'all sit at 4.6% which is only higher than six others, so you're on the low end of that spread.

That said, on sites with significantly fewer bumps, the percentages of them getting some engagement is much higher, often in the 20-50% range, which indicates that reducing the number of bumps might increase engagement in the posts that are bumped.

My hope is that by understanding why we bump questions we might be able to work towards a solution that will address reducing the number of eligible questions and the number of bumps happening so that, eventually, y'all can remove the restriction entirely and still not feel overwhelmed by bumped questions.


I can't promise that any solution will absolutely fix this issue but I can definitely make some suggestions. I'll start with the guidance that I posted on MSO because, even though there may be some special considerations here based on this question and the linked one, I think it's still a good place to start and something anyone reviewing bumped questions should keep in mind.

What I'm getting out of this all is that Community bumps are... a sort of unofficial, poorly-explained review queue. They're a way we've come up with to get people to look at older content and see if it has value - so, knowing that, what should you do?

The first thing worth considering is whether you have the domain expertise to judge the questions and answers - if not, then it's probably best to leave it for someone else to review. Let's assume you do have that expertise:

  1. Look at the question first - is it a good question and not a duplicate?
    • Yes! (go to 2)
    • No! You have two options, you can do one or both of them:
      • Close - closed questions will not be bumped. (requires more than one person)
      • Downvote - negatively-scoring questions will not be bumped. (most effective if post has a score of 0)
  2. Look at the zero-score answers one at a time and repeat as necessary.
    • If you can confirm the answer is good and correct, upvote - if at least one answer has a score of >= 1, the question will not be bumped.
    • If you can confirm the answer is low quality or incorrect, downvote - if all answers have a score <0, the question will not be bumped.
    • If you are unsure, skip. Best not to vote if you can't adequately judge the answer.

This is all well and good for standard issues with bumped posts. And I'd guess that it might be enough for some of your posts. To some degree, this may be an issue of reduced voting behaviors or (as you indicate in the question) the content is often niche enough that it's difficult to validate whether answers are correct or not.

While I don't know that I have a solution for that situation, what I would offer is a solution for the situation described in the linked meta question. The answer seems to indicate that the old, irrelevant questions are bad but that there's no appropriate close reason for them, so nothing should be done... why don't we do something?

Right now you have three community-specific close reasons which is the most we make available by default. If y'all can start a discussion around requesting a new close reason for obsolete questions that can't be answered any more, that's something I'm willing to consider. Y'all don't use the Android-independent close reason that much (10k rep link) - only 10 times in the last 90 days - but you do likely use it enough that it's worth keeping.

This sort of close reason could be somewhat controversial but if it's honestly the case that the questions are about an Android device or version of Android so old that it renders the questions unanswerable, I think I'm willing to go along with it - assuming that y'all feel like it's a good option. Since closing these questions will make them ineligible for bumping, this will allow some cleanup of content that may not be useful, even from a historical standpoint.

The most similar close reason that I'm aware of is one on Ask Ubuntu called "End of standard support or end of life release" - essentially, they don't support questions that are about Ubuntu if the version is no longer supported. They leave existing questions open but close new questions about that version if they are asked on the site.

What y'all could do - since I know that a strict cutoff may be difficult, is work on making some rules around how these questions can be closed - eventually. For example, if a question is asked about a device or OS version that's no longer supported, leave it for some amount of time - say, a year - and if it doesn't get an answer or if an answer isn't upvoted or the question is inactive, close the question. At that point it's unlikely the asker will still find value in an answer to the question and, if the question score and views are low, it's unlikely to be useful to anyone else.

That said, I want to emphasize, I don't recommend closing older questions that have good, validated answers - only those that are of little value due to having no, or no good answers.

While some of these may then be deleted by Roomba once they're closed, they'll stick around if the question or one of the answers has a score of at least 1.

Anyway, for now, you're getting a reprieve - I've lowered the setting to 5 max on the front page but please - please - consider finding a way to get rid of this old content that's likely not benefitting anyone.


Considering the data from 2021: a year in moderation where the Community bot bumped 8223 in 2021, it means that on average, around 22.5 questions were bumped every day, which I think is excessive and polluting the homepage and "Active" questions tab.

Until today, I've still been trying to check every bumped question every day, but nothing much that I personally can do other than doing a copyedit (which bumped the same question more), deleting apps-only answers that are already obsolete, or cleaning up the comments; which doesn't really prevent the Community bot from bumping the same questions over and over. And if I don't have much time, I may ignore these bumped questions again; which, again, doesn't solve the issue.

As such, for a trial, I'm proposing to reduce the threshold to 5 for now. If it's still overwhelming, then we may revisit this discussion in the future.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .