An experiment in adding a live activity panel to the question ask page

Interesting hypothesis, but I can easily see this leaving a negative impression with newer users as opposed to a positive one for two reasons:

1. Site activity is not the same as question views, so users may be left wondering why nobody is viewing their question.
2. In the same line as number one, there's no correlation between site activity and expertise, so a lot of users online doesn't mean that anyone can actually answer the question.

Furthermore, as has fairly persistently been pointed out elsewhere, writing a quality answer takes time, practically when someone has a well written question that points to an esoteric problem. It seems to me that SO is trying to address the wrong problem - instead of possibly giving the impression that a question may be answer quickly why not be more explicit that "Quality answers take time!"

Remember, Stack Overflow has a large number of tags, and activity levels vary significantly across them.

You might find it helpful to include in your experiment the observation of the effects of not showing or showing the most active tags.

This is the wrong place for that information

New users frequently report to us that they view Stack Overflow as a slow place to get answers to problems that they face.

First off, thank you for taking feedback from the previous post and making sure to describe the experiment in advance this time.

However, I want to make sure we're on the same page regarding the validity of trying something like this.

Quoting myself from a comment there, with minor editing:

We aren't offended if new users view Stack Overflow as "slow" — nor do we feel disrespected. However, we don't care if users have that perception, because being "fast" in the sense they're talking about - i.e., responding promptly to questions and ensuring that they get a personally relevant answer quickly (and without additional time expenditure on their part) - is explicitly not part of what we do.

We achieve speed by being a) direct and b) searchable. The way to address new users' perception of slowness is to get them to not ask a question unless really necessary.

Which is to say: we do new users a service when we close their questions as duplicates. We link them directly to an existing answer, which is even faster than writing one; we can do this without insisting on polishing the question to meet other standards and avoid closure (since the question does get closed anyway); and we do it without any more drama than necessary. (In the most egregious cases of laziness, one might comment along the lines of "I copied and pasted your question title into a search engine and found the linked duplicate as the first result", assuming that's true. For me, it has been true many times.)

The way that new users are intended to get fast results is to find an existing question, not to ask their own. The primary purpose of asking a new question is not to get a direct, personalized response (since we after all explicitly don't offer "help" and feel no sense of urgency). Rather, it's to expand the supply of high-quality questions with high-quality answers, specifically so that searching will work better in the future.

A new question is a kind of bug report against the site - an implicit claim that the question being posed doesn't appear to be there already, and furthermore that it should be. This is, of course, actionable only when the underlying problem has been correctly identified, the corresponding question has been properly scoped, any necessary context is provided, and the absence of prior questions has been evaluated appropriately.

And in 2025, on a site which has over three times as many publicly visible questions about programming as Wikipedia has articles about literally anything notable, it is vanishingly unlikely that a beginner to programming has a question that actually matches that description.

I also want to pick at your phrasing a bit, and make sure it's clear: "problems" cannot be answered; only questions can. The distinction is vital because a question is what you end up with after analyzing the problem you've encountered, so as to:

1. Figure out whether you need to ask about how to do something or about why something you don't understand is the way it is;

2. Figure out what that thing is based on the problem you encountered;

3. Phrase a question in a way that has enough context to make sense, but is still clearly a question that could be asked by others who go through the same process.

Therefore, if the goal is to convince new users that Stack Overflow is full of recent activity, the information about online user count etc. should appear all the time, not just on the Ask Question page - because new users should usually not see that page. (And you could perhaps also make a show of, for example, how many curation actions were taken in the last hour or so. Yes, including duplicate closures. If you can't put a positive spin on that, you haven't "gotten it" yet.)

And if the goal is to convince new users (especially ones not already skilled at programming) that Stack Overflow can provide quick answers, the right approach is to openly discourage asking a new question, and instead prominently encourage them to:

1. Follow standard debugging advice, if they are trying to fix a problem in existing code;

2. Break complex tasks down into logical steps, if they are trying to figure out how to do something;

3. Either way, make sure they can identify specific, atomic things to ask about (and if they need to learn more than one thing, they should be prepared to have to worry about each one separately);

4. Put effort into making sure they understand and can properly use the relevant terms and jargon (because this is essential to searching existing Q&A and also important for clearly posing new questions);

5. Search the site both with the site's internal search and with external search engines (and explicitly include useful tips, such as adding site:stackoverflow.com to external searches).

Because every programmer needs these skills anyway, no matter what; and having learned them, they are permanent — and, in my experience, highly transferable to every other aspect of one's life.

Certainly they are necessary to "get help" from Stack Overflow - because they're part of the process of finding existing Q&A, and also (a smaller, but still essential) part of the process of asking new questions.

It kind of reminds me of Reddit's activity thing for subreddits. I wonder if you'd want to lean into that and factor in the aspect of tags somehow.

My one minor gripe is that it shifts the guidance down. Is there some way to make that not happen? There's a lot of basically unused space at the top right of the page.

I think this communicates the opposite, this indidicates an expectation of 'fastness' that the site does not (and cannot) offer.

On many social media sites with algorithmic feeds that people are used to your post gets attention in the first hour or even ten minutes or it is buried. Advice for getting that initial momentum usually includes making your post at the right time of the right weekday, and piggy-backing on tags of some form that are 'trending' in some way -- This activity window gives users the impression that they need to do this on here, too.

But both of these techniques are useless and unnecessary on this site.

We don't want newbies to rush out posts or leave them waiting to hit some specific astrological time window, we don't want newbies to choose the tags of their question by what gets attention, we don't want newbies to look for short-term trends in tag-activity.

We don't want to give new users the impression that a new users lack of answers is due to not playing the algorithm and attention game correctly.

The simple reason that people are not getting 'fast' answers is that new questions nowadays are either bad (duplicate, low-quality, unanswerable, lacking understanding) or hard (niche, deep). The bad ones are not answered out of principle, the hard ones just take time naturally. The questions that are both easy and good are mostly gone, because they already exist on here.

A few minor notes on the UI of the box:

• I'm unsure why it uses @Svg.SpeechBubbleRightStar for '36 answers shared' as opposed to using @Svg.Answer, especially considering the questions one uses @Svg.Question.
• It's not wrong, per se, but the blue color on last 1 hr feels like it should be a link (maybe to let the user select a different time span, although I don't know that such a dropdown would be useful to 1-rep users)
• Should 'Community Activity' be 'Community activity' for consistency with 'More helpful links'
• The horizontal rules between the stats are fine, but it might be good to remove the one under the heading