Forum participation data

Howdy all,

I don’t pay much attention to website analytics (probably to the detriment of our ability to make money), but I happened to notice that last month was our highest traffic to the forum (probably) ever with 1.9 million pageviews, which is nice. It’s nice to know folks are still finding what we’re doing useful, and it’s a confirmation that moving to Discourse was a great choice.

I qualified the above statement with “probably” because there’s an outlier on this chart that I can’t explain. Actually, a couple of outliers I can’t explain. One is March of 2022 with over 2 million pageviews. I’m suspicious of that because it is surrounded by months with dramatically fewer views. I think that indicates uncategorized crawlers (maybe “AI” crawlers, as they’ve been shown to be hiding the nature of their business), while this year we’ve seen steadily increasing visitors with nothing looking like a crazy jump. The other strange jump is from May to June of 2021. I don’t think that is mostly attributable to crawlers, more likely just a flurry of activity because of new releases of stuff, but maybe it was the beginning of the “AI” feeding frenzy.

I suspect the increase in crawler activity (purple) comes down to there actually being more crawlers (because “AI”) and also more crawlers being properly detected that were missed before.

I don’t have a great feel for how much of this is actual humans, but we can see that posts and interactions have been steadily, if slowly, increasing ever since we launched the Discourse forum in 2020, though much more slowly than the visitor count. Since I’m utterly ruthless about deleting LLM-generated posts (when I recognize them), and blocking users that use LLM-generated text, I think we’re seeing mostly human interactions. If you’re a human, and you’ve posted, thanks! If you’re not human, I’m gonna block you.

One reason I find this encouraging is that most StackOverflow sites have seen a dramatic decrease in usage since LLMs have come along, and I find that worrying, for a few reasons. One reason I’m worried about it is that LLMs are wrong so much of the time. I know from talking to customers and users who’ve tried to get answers out of ChatGPT, that it very often leads them down absolutely insane paths, getting them into a much deeper mess much faster than spending a few minutes reading documentation or asking here would have done. It sounds so confident, and is right some of the time, so it kind of overloads our brain into doing whatever it says. Like driving into a lake because the GPS told you to.

Anyway, the fact that we aren’t seeing a decline in real human users having actual human conversations, and in fact, those conversations have been consistently increasing over time, is heartening. We’re not just talking to ourselves here, we’re helping a lot of people. I’ve seen studies that indicate that orders of magnitude more people read forums than post to them, and these numbers bear that out. So, ask your questions, folks. It’ll probably help somebody else, maybe dozens or hundreds of somebodies. When you ask ChatGPT, it doesn’t help anybody else (and it might not help you, either, since you’re gonna have to fact-check it…you have to be cautious about stuff you get from forums, too, but at least in a forum, you’ll get some other opinions if someone is just plain wrong).

I don’t really have a point here. Just thought others might be interested in the state of the community. One more chart:

7 Likes

Regarding “human conversations” – I really appreciate this forum and the entire Virtualmin team. I only pay for one Virtualmin license per year but I always feel like your team treats me like a VIP client. And the “System Settings → Virtualmin Support” area providing “Submit Ticket” and “Remote Login For Virtualmin Staff” are such friendly features. It is also so nice to have a real person to ask emergency “non Virtualmin related” questions of when we’re in a panic. You all are so patient with us for putting you thru that. I wish VMware (now Broadcom) was as helpful as this. I will miss this kind of support when “AI” becomes commonplace and we all “talk” to our computers instead of typing on them.

Regarding forum interactions – I tend to post less now… probably my fear of being told I posted to the wrong group… but I did want to let the Virtualmin “Yo” Crew know that their support over the last decade+ has been much appreciated!

2 Likes

I don’t care at all if you post in the wrong category, as long as it’s clear what you’re talking about (the only frustrating ones are where the question is posted in Webmin category, and the question says “Webmin”, but then several comments in it becomes clear they’re actually talking about Virtualmin). It’s easy to recategorize if the category bugs somebody.

I know I can be a little grouchy sometimes, I’m not known for my social graces. I try.

3 Likes

To err is human, but to really foul things up you need a computer. Paul R. Ehrlich

1 Like

I appreciate that when you know your job/field and you are passionate that it can be VERY frustrating answering questions when people have not done the basic of searching the forum or doing a bit of googling. I am always grateful of your very detailed answers. I can get frustrated with my IT clients who no matter how many times you tell them not to use the same password will just ignore you and then you get a phone call where they tell you they have been hacked. :slight_smile:

This discourse forum is great, it pushes changes to topics and messages live which is brilliant. I am going to use it myself when I figure out docker and proxying (article to come).

I have found lots of answers to questions on this forum but I do find the search can be a bit flakey but it is not horrible.

Webmin + Virtualmin + Staff + this community = Awesome

Can this be mitigated? Can the IP ranges be blocked? Do they have customer User-Agents?

I have never used Chat-GPT but it might be fun for a few things, but I would never take it’s answers as golden.

This is great to know that the community is very active even if some are lurkers. As you know there are some very active members on this community that do a lot help and from whom I have had assistance (including Staff) which is appreciated.

Additional

I would love it if everyone who had their issue solved would post the answer and hit that solved button :smile:

I hope this community goes from strength to strength.

Because of the help from the staff I bought a Pro license even though I have not used it yet (I will do :smile: ). The GPL and the community helped me decide this was my platform to use.

Thanks for sharing these stats as we can see a slice of what you see.

1 Like

It’s true. I am one of these people. My native language is Spanish, but I read plenty of forums in English. The best way to learn about something. This forum with Discourse (with your answers @Joe, @Ilia and @Jamie… and the Virtualmin community), is the best way to learn about Webmin/Virtualmin :slight_smile: Thanks for all.

I bought the annual license for the first time this year, and, definitely, I am going to renew the following year. Great work!

2 Likes

One of my biggest teeth grinders is Stretchoid and Censy bots. I’m pretty sure there is more alike, but these two can really foul up your statistics. They are annoying to the part where I had to learn how to make my own ipset with firewalld and start rejecting them as I find them. These bots end up in fail2band on a constant bases as well. I’ve also noticed new bots within the past several months, because my logs are loading up with them.

I like to step aside for awhile even though your like “Hey I know the answer to this one!”
Everybody wants to help and you feel the enthusiasm. But, it seems we get stepping over each other. Then comes the arguing about who knows what and 90 post later we finally come up with a solution and at times hurt feelings. I don’t think for one minute that this is the only forum that acts like this, but I think we can do better at least attitude wise.

This forum is literally called “The Grumpy Old Men Club” in my office. I have clients with Pro versions who refuse to come back to this forum, so I am stuck asking all the questions that really don’t need direct support ticket. None the less, they stick with Webmin/Virtualmin because it has surpassed its basic duties and goes way beyond its competitions.

Yes we know, and appreciate your honesty :slight_smile:

1 Like

Topics can be moved.

Even if you don’t type, copious use of the “like” button can be just (if not more so) as effective as a “me too” post.

3 Likes

I admit I have used LLM a couple of times (on the repo, not on this forum), but it was really helpful because it presented my argument in proper neat English while following a certain standard template.

I hope such LLM use-cases would be more acceptable because they do make (certain) people more comprehensible (and presentable, as if one is wearing a suit).

Now it comes to my mind that what you meant by using LLM is to generate answers to questions posted in this forum without any actual effort. In such case, I’m totally with you :handshake:

For translation and helping you with English as a second language, I have no problem with AI-tools. There are many great things machine learning can do, but answering complicated questions without lying may not be one of them.

Note that I don’t like LLMs prose; it’s bad in a way akin to the most average, but wordy, flowery, and pretentious, writing by a student writer. But, I understand it can help with confidence when posting in a language you don’t know well. Just try to keep it concise, and don’t ever answer people’s questions with ChatGPT generated answers (I literally ban a couple of spambots using ChatGPT every day, I won’t ban someone I recognize and how genuinely contributes, but I will warn about posting LLM answers, and yes, I can usually tell immediately).

I spent some time analysing statistic on my own platforms (which doesn’t make traffics) and I was really surprise to see 99% of the few traffic generated was only crawlers. Or worst of all fake traffic (especially concerning specific urls, There are some platforms which will generate traffic on other site where their link is present). It can be hundreds per month for only 1 single page.

According to me here are the stats to check if you want to see if your community is growing:

  • The number of logged users
  • The number of member who create a post at least 1 Time per months (Maybe the “daily engaged user” ?)
  • And the number of server using Virtualmin (The most important if you want to get money)

I saw the first 30 318 logged users but I’m not sure for the 2 others.

So concerning anonymous users, there are so much platforms today generating so much fake traffic (and selling it cheap) that it’s hard to say if they are real or not. It completely depend how you count views.

Concerning the peak you highlighted:

  1. It may first be a YouTuber who did a video about Virtualmin (Or a journalist). Which will (Insanely) increase views (Due to the enormous amount of fake traffic on these platforms) and (slightly) Members and servers using Virtualmin (Because they hopefully also have a big community).

Watching the charts we can see there is around 15 (+30%) more signup on Mars 2022 and around 150 More posts (+30%).

  1. Or, if the number of member and servers using Virtualmin did NOT increase. It might simply be one of your fan posting a comments on social media and YouTube. This time it will simply generate more fake traffic (unfortunately few people will read or follow links in comment, but Bot/fake traffic usually follow every single link).

In June 2021 we can see 10 Less Signup (-20%), -25 Topics (-20%) and -150 Post (I would also say -20%).

There are other possibilities but I suggested those who seems to match your stats.

Concerning Generated answer, according to me they shall all be deleted because you stole the intellectual property of the one who answered first. Except if you mention the source. So you stole views (so money) and put all the money in the same pocket (Which is never good) which is exactly the purpose of chatgpt.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.