Visualising the Effects of K-pop Happenings on Forum Activtiy (or: Let's Take a Look at a Year of AKP User Traffic)

--==See here for the 2024 update==--

On January 6th last year, I started a long-term project to document the number of AKP members and guests online at a given time. I wanted to see how major K-pop events impacted forum activity, and was curious about the general traffic patterns of the forum in general. To do this, I simply recorded member/guest counts periodically, and graphed the resulting numbers in the chart below. Note that member counts use the primary axis on the left, while guest counts use the secondary axis on the right:

AKP_User_Traffic.png

I've annotated what I believe are some significant events that have happened over the past year. Do you think that there's anything major that I missed?

As we can see, important comebacks and news definitely have a noticeable effect on forum activity. Particularly notable peaks of activity include Big Bang's comeback and Le Sserafim's reveal in April, and of course the big peak in August, when Blackpink, IVE, and Twice all had back-to-back comebacks.

However, what's intriguing is that forum traffic fell noticeably after this time period. What do you think caused this decline? I know there was a forum reorganization in August, so could that have been a contributing factor? Perhaps its effect was initially masked by the August girl group "bloodbath", but made itself known once that had passed.

Fortunately, user activity seems to have leveled out starting in November, and has actually recovered somewhat. Towards the end of December, the forum underwent another revamp, so I'm curious to see what effect that will have going forward.

Finally, I do have to note that, although the number of guests online does tend to track the number of members online, guest counts do seem more erratic. Look at the big shifts at the beginning of 2022, the start of September, and the start of November. It almost seems as if the method to track members online changed suddenly at these points.

Further Reading on This Topic:

AKP Forum Activity Versus Time of Day

AKP Forum Activity Versus Day of Week

Tracking the Number of Threads and Posts on the Forum

Forum Activity for 2023

Forum Activity for 2024

Complete Chart of Data from 2002 to 2004

Quote from desdesdes

Perhaps Jhonny might interested with this

Indeed. Though I suspect Johnny already has internal site metrics that show a similar situation.

Quote from thelightoftruth

Nice. Curious, what time did you capture the data? Peak time in Americas, Asia or Europe?

I captured throughout the day, at irregular intervals spaced generally a few hours apart. In fact, I'm actually putting together another chart that shows when the forum is the most active across a 24-hour period.

Quote from purepalette

You could feel in real-time after August the activity droppped. Maybe all the fan wars and chart discussions drove some people away

Yeah I felt that too after August. It's an interesting hypothesis, that sheer exhaustion drove people away. At least some of them seem to be returning though.

Quote from the101

I wonder if some of the Winter Wonderland activities helped somewhat stabilize in that November/December time period? Hunger Games, aesthetic competitions, secret Santa, Black Friday sales and watch parties were all popular threads and active events esp when there were no cbs to talk about.

To be quite frank, I wasn't even sure what Winter Wonderland entailed, as I didn't take part in it at all. But all those special events you mentioned could certainly have resulted in the rise in activity we saw at the end of 2022.

Quote from the101

I also think your chart is thinking about Kpop (external factors) but not site factors either. Obviously on a site with such a large Army presence, the slowdown in BTS activities would also have a corresponding slowdown, but it also isn’t taking into account what activity may have happened on the site. People come and go for Kpop comebacks but I also remember a spell last year where there were a sequence of highly controversial threads that resulted in banned users, while others fled the site. I don’t remember when it was tho, but it would be interesting to overlay that too and see if that impacts user stats. Or major fan wars that either drove traffic up or down?

This is a great point. I typically don't get involved in forum "drama", so unfortunately I don't have much insight into fan wars or problematic threads. I also never participated in the anon section, where I imagine at least some of this controversy must have taken place. And actually, now that I think of it, I wonder how much the removal of anons impacted forum activity.

Quote from the101

(I wonder what the long-term patterns on forum usage in general on the internet as communities use other spaces like Discord. I suppose this is a long term consideration for the site’s management - like if forum traffic across all Internet forums is down 10% and AKP is down 15%, let’s say, that’s also reflective of a general trend across the industry - where did those users go to get info now? Maybe the issue isn’t even AKP but the format of a forum in general?)

I'm glad you brought this up as well. Perhaps Internet forum activity in general is decreasing as users migrate to other platforms. I guess such macro factors would affect long-term AKP usage trends, but I'm still curious about the acute decrease in traffic this site saw in September.

Quote from K-Pop-Scientist

It's amazing how many more guests than members there are.... think AKP should be working out how to convert those numbers using this data lol

Yeah that was one of the things that jumped out at me too. So what are some ideas to convert guests to members?

Quote from Kreatin

Did you capture the visitors with a script or manually?

The recording of the number of users online was all done manually. Essentially, I would simply input member/guest counts (reported on the main forum page) into a spreadsheet when I visited this site.

Quote from ranpo

in akp's case, they only allow guests to read a few posts in a thread and if they want to read the rest, they have to make an account

Ah this must be a new feature, since previously, you could read entire threads without making an account (which is what I did prior to signing up as a member).

Quote from motorp77

in the summer, a lot of people are on holidays so more people are here, in September people start working again (most of the time even more than the beginning of the year) so there are less people here.

That's an interesting hypothesis, and actually the opposite of what I was initially thinking.

I would've thought that forum traffic would be less in the summer, since people are more likely to be out traveling and having fun. On the other hand, people might be more tempted to browse the forum when they're bored at work.

Quote from bartkun

Forum needs more articles on main site which will be linking to threads.

For example let's say in next days there will be new Baby Monster teaser then at the end of article there should be link to forum thread and some words: What do you think about the teaser? Come to our forum and join us in a discussion about it

Indeed. I've always thought it a bit odd that comments on news articles on the main AKP site are completely separate from the forum.

As stated in my response to thelightoftruth, over the course of a year, I captured user activity numbers throughout each day, at irregular intervals generally spaced a few hours apart. Thus, I was able to generate the chart below, which shows when the forum is the most active across a typical 24-hour period. Again, member counts use the primary axis on the left, while guest counts use the secondary axis on the right:

AKP_User_Traffic_TimeOfDay.png

If you're not familiar with UTC (Coordinated Universal Time), it's essentially equivalent to GMT (Greenwich Mean Time) for our purposes.

As we can clearly see, forum activity is the highest around 16:00 UTC. Given that AKP is based in the United States, my assumption going in was that site activity would be highest during normal US business hours, and this appears to be the case. A typical "9:00 to 5:00" work day in the EST time zone corresponds to 14:00 UTC to 22:00 UTC, and this roughly correlates to the period of highest forum activity. Would you agree that these results imply that a large portion of AKP members are based in the US?

Now the above applies to AKP members. However, it's also interesting to note that the number of *guests* online does not seem to depend nearly as much on time of day, as there is no clear peak period of guest activity. Why do you think this would be the case?

Quote from K-Pop-Scientist

This is interesting, I noticed that uploading to my Youtube and Tiktok especially at 16:00 GMT gets most views especially on Tiktok vs other times....

Maybe at this time most of the West is awake or waking up (USA) maybe?

Yes, I believe that time frame corresponds to when people in the US/Canada are the most active. But it's also the afternoon in Europe, which is also an active time period.

Take a look at this thread, which asks AKP members which country they're from. The US is by far the most represented country, but there are just as many users from Europe.

Quote from Help1

And there is a dip between 6:00 to 11:00 AM UTC which is 1AM-6AM EST (10PM-3AM PST) so that could support that too

Absolutely. I feel pretty comfortable declaring that this site's traffic is primarily driven by US-based users.

Since I already had all the data, I decided to put together a chart showing how the average number of users online varies by the day of week:

AKP_User_Traffic_DayOfWeek.png

It turns out that there's actually not *that* much of difference in site activity depending on the day.

Member activity is lower on the weekends, as I initially expected, though strangely, it's also lower on Wednesday and Thursday. What could be the reason for that?

In April last year, along with the number of members/guests online, I also began capturing the number of threads and posts on the forum, as we see in the chart below:

AKP_ThreadPosts_Count.png

From looking at this chart, it's clear that both threads and posts are being deleted periodically, which surprised me a bit. This begs the questions:

(1) Why are threads/posts being removed in the first place?

(2) What are the criteria for removal?

I've put together a new chart that shows user activity for 2023, as shown below:

AKP_User_Traffic_2023.gif

Some thoughts and observations:

• On average, forum activity was much lower in 2023 compared to 2022. It looks like the decline we saw in September 2022 (shown in the initial post of this thread) is the new normal.
• Forum traffic was also less "peaky" in terms of steep fluctuations. In other words, it was more difficult for news events to cause spikes in forum activity. Even "big" girl group comebacks failed to move the needle much. Is this a case of girl group saturation?
  • Similarly, with the exception of Jungkook, BTS solo debuts didn't appear to have a major effect on forum activity either.
• In August 2023, we see what appears to be a noticeable decline from 125 members online on average to a new, lower steady state of about 100 members. What could have caused this?
• Also around August 2023, the number of guests online rose dramatically. This persisted for the rest of the year. Any idea what happened?
• What does the future hold for AKP? Will activity continue at the current levels, or is more decline expected? Can anything be done to reinvigorate and revitalise the forum?
  • There appears to be a large number of guests visiting AKP. How can they be converted to participating members?

Quote from Catgirl

Didn't NJ, BP and SNSD have comebacks in August?

NewJeans released Get Up in July 2023, which is shown on the chart. Meanwhile, BP had The Girls, a promotional single, while SNSD didn't have any comeback in that time frame as far as I know.