[Chart] Bitcointalk statistics on impression counts for ads

[Coin-1]

All charts are current for round 351 ending on July 29, 2022.

As I can see from the graph, the "Total daily impressions per slot" metric hasn't changed much, but the similar metric "Daily impressions from logged-in users" has significantly increased (actually doubled) from 11250 to 24400 in the last 6 months. Perhaps this means that people came to the forum to talk and not just read Bitcoin related topics. On the other hand, the "Unique IP addresses from logged-out visitors" index has dropped from 210K to 150K for one round. This looks strange.

[Coin-1]

I have updated all charts on the front page of this thread to the 358th round in order to reflect the new data as of February 27, 2023.

In general, all statistical indices look quite stable.


It should be noted that on September 20, 2022, theymos said:

Note that I've decided to suspend ad sales. These auctions will probably be permanently ended, and I'm not sure whether or in what form advertising in general will be continued. Maybe some ads will go up again fairly soon, or maybe there will never be any more ads; I haven't decided yet.

As far as I understand, rounds 353 and up did not advertise any service, but did show various dictums of the forum founder, Satoshi Nakamoto, and other Bitcoin-related expressions. Now each round lasts exactly 30 days.

[digaran]

That's bad, after seeing the charts it's obvious that forum has lost it's influence over time after 2018, I wouldn't rule out merit system as the culprit though. I guess we can't have it both ways, to achieve something you need to sacrifice another, but was it worth the try?

[logfiles]

That's bad, after seeing the charts it's obvious that forum has lost it's influence over time after 2018, I wouldn't rule out merit system as the culprit though. I guess we can't have it both ways, to achieve something you need to sacrifice another, but was it worth the try?

Forget about NFTs and DeFi... Bounty hunting was very lucrative during that time of ICOs so that's probably the time the forum received a big number of new accounts. This could be the reason why the impression counts for ads were high at the time.

Merit system may have had some impact too.

[joker_josue]

I have updated all charts on the front page of this thread to the 358th round in order to reflect the new data as of February 27, 2023.

In general, all statistical indices look quite stable.

I think the numbers are quite interesting, bearing in mind that we are going through a downturn in the market.
They are better than they were, between 2015-2017.

The peak of 2017-2018 was a bit out of the ordinary due to the huge wave of ICOs and the like that came to market. If we remove those years, we see that the numbers have risen consistently over the years.

[Coin-1]

I have updated all charts with new data as of November 24, 2023.

The minimal recent value for the "Daily impression counts for ads" parameter was recorded in May 2023 and averaged 100K per slot. Over the past months, the "Unique IP addresses" and "Unique logged-in users" parameters seem to be more stable.

The peak of 2017-2018 was a bit out of the ordinary due to the huge wave of ICOs and the like that came to market. If we remove those years, we see that the numbers have risen consistently over the years.

On the second post of this thread, I now show recent statistics starting from January 1, 2020, rather than January 1, 2017. As I can see from these charts, the latest peak in numbers occured in 2021. In my opinion, some parameters are quite volatile. They rise and fall depending on many circumstances.

[joker_josue]

On the second post of this thread, I now show recent statistics starting from January 1, 2020, rather than January 1, 2017. As I can see from these charts, the latest peak in numbers occured in 2021. In my opinion, some parameters are quite volatile. They rise and fall depending on many circumstances.

Thanks for the update.

Logically, there are always many factors involved in the variation of numbers. And many of these factors will be completely unknown, and it will be difficult to understand what motivated such a movement. But such movements are still interesting.

[Coin-1]

I haven't updated this thread in over 10 months, so I decided to renew all the charts to show statistics as of October 19, 2024, when the last 378th round ended.

As I can see from these graphs, in August 2024, the "Total daily impression counts for ads" index has surged from 200,000 to 350,000. The "Daily impressions from logged-in users" index also increased, but not as sharply. It is clear that the current state is more positive than it was before.

Logically, there are always many factors involved in the variation of numbers. And many of these factors will be completely unknown, and it will be difficult to understand what motivated such a movement. But such movements are still interesting.

Yes, these trends are interesting. I think people see the growth of the cryptocurrency market, so they come to the Bitcointalk forum as one of the most reliable sources of information.

[fillippone]

I haven't updated this thread in over 10 months, so I decided to renew all the charts to show statistics as of October 19, 2024, when the last 378th round ended.

As I can see from these graphs, in August 2024, the "Total daily impression counts for ads" index has surged from 200,000 to 350,000. The "Daily impressions from logged-in users" index also increased, but not as sharply. It is clear that the current state is more positive than it was before.

Logically, there are always many factors involved in the variation of numbers. And many of these factors will be completely unknown, and it will be difficult to understand what motivated such a movement. But such movements are still interesting.

Yes, these trends are interesting. I think people see the growth of the cryptocurrency market, so they come to the Bitcointalk forum as one of the most reliable sources of information.

I couldn't reconcile the surge in the impression count with the decreased number of users.
What does it mean? The user number decreased, but they are actually more active on the forum, so that the impression number actually increased?

[joker_josue]

I couldn't reconcile the surge in the impression count with the decreased number of users.
What does it mean? The user number decreased, but they are actually more active on the forum, so that the impression number actually increased?

Yes, that could be it.
In other words, whenever you open a topic, it counts as a view/impressions. If you open this topic 10 times, it's 1 user = 10 impressions.

Another thing that could be happening is more unregistered people (guests) viewing the forum.

[fillippone]

I couldn't reconcile the surge in the impression count with the decreased number of users.
What does it mean? The user number decreased, but they are actually more active on the forum, so that the impression number actually increased?

Yes, that could be it.
In other words, whenever you open a topic, it counts as a view/impressions. If you open this topic 10 times, it's 1 user = 10 impressions.

Another thing that could be happening is more unregistered people (guests) viewing the forum.

I see no reason why there should be more unregistered users.
I, on the other hand, understand why someone who registered to a forum from the year 10s feels compelled to open many threads and read many posts.
Anyway, given the price action, I would have expected the number of users to grow a bit more sharply.
We are so early!
 

[BitcoinGirl.Club]

I haven't updated this thread in over 10 months, so I decided to renew all the charts to show statistics as of October 19, 2024, when the last 378th round ended.

As I can see from these graphs, in August 2024, the "Total daily impression counts for ads" index has surged from 200,000 to 350,000. The "Daily impressions from logged-in users" index also increased, but not as sharply. It is clear that the current state is more positive than it was before.

I feel stupid and I am sure I am missing something. The forum is not providing any ad service [the top banners] then how it's generating the impressions on the ad stat page?

[nutildah]

I see no reason why there should be more unregistered users.

Because the forum is an SEO powerhouse. Whenever someone searches for something related to Bitcoin, a link or two to a post/thread in the forum will appear in the top 10 Google search results. According to SimilarWeb, 39% of all traffic comes from "organic search" and 50.1% is "direct", which is us + people who just copy/paste a forum link. Referrals make up 8.76%

Whats kind of interesting is ChatGPT accounts for 3.69% of referrals, which is 0.32% of all forum traffic. I expect the number of AI-driven referrals to get bigger over the next few years. I even get hits on my website from ChatGPT these days. Exciting times.

[joker_josue]

I feel stupid and I am sure I am missing something. The forum is not providing any ad service [the top banners] then how it's generating the impressions on the ad stat page?

Basically this page was created to present the forum's statistical data, with advertisements in mind.
In other words, the page is configured to display forum statistics, related to each announcement cycle.

Ultimately, it's not about the statistics of the ad itself, but about access to the forum.

[BitcoinGirl.Club]

I feel stupid and I am sure I am missing something. The forum is not providing any ad service [the top banners] then how it's generating the impressions on the ad stat page?

Basically this page was created to present the forum's statistical data, with advertisements in mind.
In other words, the page is configured to display forum statistics, related to each announcement cycle.

Ultimately, it's not about the statistics of the ad itself, but about access to the forum.

Let's analyze the data from the round 379. We have
379   Oct 19 2024 16:00:00 UTC   Nov 18 2024 16:00:00 UTC   30   10   9867673   1877155   391358   6903

This means we have 10x9867673 impressions for the slots?
98,676,730 that's above 98 millions of page loads per month with 6903 unique logged in users.
Surely there are so many bots are running by different organizations to consume the resources from the forum host.

[joker_josue]

Let's analyze the data from the round 379. We have
379   Oct 19 2024 16:00:00 UTC   Nov 18 2024 16:00:00 UTC   30   10   9867673   1877155   391358   6903

This means we have 10x9867673 impressions for the slots?
98,676,730 that's above 98 millions of page loads per month with 6903 unique logged in users.
Surely there are so many bots are running by different organizations to consume the resources from the forum host.

Look, it's not that difficult a number to achieve.

Note, each of us for each topic we reply to generates 3 pages loaded. At the very least.
It's the page when you enter the topic and read it, it's the page where you write the post, and it's again the page Topic.

Now it starts to multiply that by the remaining pages and thousands of users.

[fillippone]

Let's analyze the data from the round 379. We have
379   Oct 19 2024 16:00:00 UTC   Nov 18 2024 16:00:00 UTC   30   10   9867673   1877155   391358   6903

This means we have 10x9867673 impressions for the slots?
98,676,730 that's above 98 millions of page loads per month with 6903 unique logged in users.
Surely there are so many bots are running by different organizations to consume the resources from the forum host.

Look, it's not that difficult a number to achieve.

Note, each of us for each topic we reply to generates 3 pages loaded. At the very least.
It's the page when you enter the topic and read it, it's the page where you write the post, and it's again the page Topic.

Now it starts to multiply that by the remaining pages and thousands of users.

If every reply generates 3-page loads, this means 4,700 "replies" per registered user monthly.
I guess that makes you understand the amount of traffic generated by unregistered users.
It would be nice to streamline this amount of pages more profitably for the forum.

[joker_josue]

If every reply generates 3-page loads, this means 4,700 "replies" per registered user monthly.
I guess that makes you understand the amount of traffic generated by unregistered users.
It would be nice to streamline this amount of pages more profitably for the forum.

Now I don't understand your calculations.

In the last 30 days registered, there were 9 867 673 impressions/views.
The same data says that only 1 877 155 users logged in.
This gives more or less 6 impressions/views per registered user for 30 days.

If for each post you make, it gives at least 2 impressions, as the banner space does not record the page where the post was created (contrary to the number I indicated before). So, this means that the number of impressions/views is very low compared to logged users. The value should be higher.

[BitcoinGirl.Club]

Look, it's not that difficult a number to achieve.

Note, each of us for each topic we reply to generates 3 pages loaded. At the very least.
It's the page when you enter the topic and read it, it's the page where you write the post, and it's again the page Topic.

Now it starts to multiply that by the remaining pages and thousands of users.

I can see the holes LOL.
Before making a post I preview them, not one or twice but many times if the reply is long. All these count too.

[fillippone]

If every reply generates 3-page loads, this means 4,700 "replies" per registered user monthly.
I guess that makes you understand the amount of traffic generated by unregistered users.
It would be nice to streamline this amount of pages more profitably for the forum.

Now I don't understand your calculations.

In the last 30 days registered, there were 9 867 673 impressions/views.
The same data says that only 1 877 155 users logged in.
This gives more or less 6 impressions/views per registered user for 30 days.

If for each post you make, it gives at least 2 impressions, as the banner space does not record the page where the post was created (contrary to the number I indicated before). So, this means that the number of impressions/views is very low compared to logged users. The value should be higher.

Ah, now I understand, and it makes much more sense.
MY value was too unrealistically high and based on the wrong number of users.
Of course, the number of users logged remains low, but the "bot" (either users casually browsing or proper AI agent reading posts) activity seems to be the real reason for this.
I'm not sure if and how to account for or react to this bot-proliferation.