[Chart] Bitcointalk statistics on impression counts for ads

[Coin-1]

Since October 2011, ads are shown on the Bitcointalk forum. Each round lasts about 7-12 days. There are 10 rotating slots which are sequentially displayed for logged-in users and logged-out visitors. It should be noted that some users who are Hero Members or Legendaries have ads disabled in their profiles.

Since April 2012, the administrator theymos publishes statistics on impression counts for ads:
https://bitcointalk.org/adrotate.php?adstats


Each ad slot has 4 statistical indices:

• Total impressions
• Impressions from logged-in users
• Unique IPs from logged-out users
• Unique logged-in users

For example:

Round 271

From Mar 20 2019 15:49:07 UTC to Mar 30 2019 21:18:53 UTC (~10.2 days).

• Ad slot #1
• Total impressions: 3426361
• Total impressions from logged-in users: 1024362
• Unique IPs from logged-out users: 162236
• Unique logged-in users: 22912

There are no statistics of the first 25 rounds. The round 26 ended on April 4, 2012. According to the comment, the round 141 which lasted 15 days in January 2015 has an incomplete data. Since the round 243 which began on May 16, 2018, the "Unique IPs" index has been changed to "Unique IPs from logged-out users" that probably reduced values of this parameter.

It is weird that the round 74 which lasted a week in March 2013 has zero values for all indices. There are no notes related to this round. Perhaps this is an error.

• UPDATE: On September 20, 2022, the administrator decided to suspend ad sales, but these statistical indices are still provided.

The charts below illustrate the dynamics of all indices on average per slot.

Bitcointalk impression counts for ads

The following chart illustrates the daily statistics for the first and the second indices of ad slot.



The chart parameter which is illustrated in blue is calculated by the formula:

Code:

TOTAL_DAILY_IMPRESSIONS =
TOTAL_IMPRESSIONS / ROUND_DURATION_IN_DAYS

The chart parameter which is illustrated in green is calculated by the formula:

Code:

DAILY_IMPRESSIONS_FROM_LOGGED-IN_USERS =
IMPRESSIONS_FROM_LOGGED-IN_USERS / ROUND_DURATION_IN_DAYS

The chart parameter is calculated by the following formula:

Code:

UNIQUE_IPS_FROM_LOGGED-OUT_VISITORS_FOR_ONE_ROUND

The chart parameter is calculated by the following formula:

Code:

UNIQUE_LOGGED-IN_USERS_FOR_ONE_ROUND

Statistical data since the round 27 which began on April 4, 2012, till the round 378 which ended on October 19, 2024, have been parsed and analyzed to create these charts.

This post is relevant for October 20, 2024, and has been updated on November 12, 2024.

[Coin-1]

The following charts illustrate the same statistics for only the last 3-4 years in order to view the latest data.

Bitcointalk impression counts for ads

since 2020

This post is relevant for October 20, 2024, and has been updated on November 12, 2024.

[Coin-1]

Reserved.

[DdmrDdmr]

Nice stats you've created, thanks.

The first charts, related to the number of impressions, are more oriented towards potential advertisers and de-facto advertisers, and give a measure of potential impact in terms of máx. achievable views.

What’s interesting for a broader audience are the charts related to the unique IPs and users:

- The Unique Logged-in Users would be "us" I figure, where the latest reading on the stats (round 271)  indicate that nearly 23K distinct users logged-in during a 10 day period as of late. It would be interesting to know how many distinct IPs are behind, but that is just curiosity.
The charts show the evolution there, and I’m curious if they will rise with the BTC price evolution over these past couple of days (we’ll see that in current round 272).

- The unique IPs from logged-out users are I figure really users who have not logged-in to browse the site. These are nearly eightfold lately, in relation to logged-in users, although there are many variations under the hood (i.e. 1 user -> n IPs; 1 IP -> n Users; user can be logged-in and logged-out during the run, etc.). Even so, if we hypothesise the relation 1:8 to be real, there is still a large potential of users that could at some point proceed to log-in and post.


I’ts also curios to note that the stats indicate that only 265 users have ads disabled, although this option is only available to Hero, Legendary, Donators, VIPs, and moderators (roughly 6.500 accounts – that is 4% aprox. have adds disabled).

[LoyceV]

The charts show the evolution there, and I’m curious if they will rise with the BTC price evolution over these past couple of days (we’ll see that in current round 272).

Probably! The visitors graph looks a lot like the Bitcoin price graph, you can quite literally see all the historic price spikes in there.

I’ts also curios to note that the stats indicate that only 265 users have ads disabled, although this option is only available to Hero, Legendary, Donators, VIPs, and moderators (roughly 6.500 accounts – that is 4% aprox. have adds disabled).

Bitcointalk has one of the least intrusive ads I've seen, I too see no reason to disable them.

Coin-1, I merited your post, then realized you only needed 11 more to become Hero Member. Congrats, you deserve it!

[Coin-1]

- The unique IPs from logged-out users are I figure really users who have not logged-in to browse the site. These are nearly eightfold lately, in relation to logged-in users, although there are many variations under the hood (i.e. 1 user -> n IPs; 1 IP -> n Users; user can be logged-in and logged-out during the run, etc.). Even so, if we hypothesise the relation 1:8 to be real, there is still a large potential of users that could at some point proceed to log-in and post.

I think that "logged-out users" means "visitors". In other words, they probably do not have an account on the Bitcointalk forum and come here just to read discussions about crypto currencies, etc.

- The Unique Logged-in Users would be "us" I figure, where the latest reading on the stats (round 271)  indicate that nearly 23K distinct users logged-in during a 10 day period as of late. It would be interesting to know how many distinct IPs are behind, but that is just curiosity.

This is true. We are "unique logged-in users". The All-Time-High of this index has reached on the day the merit system was implemented, on January 24, 2018, and as I can see, the number of active members decreases every month.

I’ts also curios to note that the stats indicate that only 265 users have ads disabled, although this option is only available to Hero, Legendary, Donators, VIPs, and moderators (roughly 6.500 accounts – that is 4% aprox. have adds disabled).

Bitcointalk has one of the least intrusive ads I've seen, I too see no reason to disable them.

I agree. All ads on Bitcointalk look beautiful and are not annoying, so not many high-ranking members have such banners disabled.

Congrats, you deserve it!

Thanks! I have really ranked up. I will endeavor to contribute to the forum.

[Coin-1]

Since another 4 rounds have been completed, I parsed the ad stats and updated all the charts up to the 275th round.

As I can see, in February 2019 there was a dip of impressions. According to the latest data, the indexes look stable, especially "Unique IPs" and "Unique users".

[Coin-1]

All charts have been updated up to the 279th round which ended on June 26, 2019.

Interestingly, impression counts for ads in the 277th round exceeded 5 million, that is, in the beginning of June 2019 about 410 thousands impressions were counted per day. The number of unique IP addresses from logged-out users and the number of unique logged-in users did not significantly increased.

[Coin-1]

Another 5 ad rounds are over, so I updated all the charts again.

Daily impression counts for ads look stable enough. The number of unique IP addresses from logged-out users for one round has slightly decreased. About 18500 unique logged-in users have visited the forum for the last ad round (~10 days).

[Coin-1]

I have updated all the graphs up to the 289th round which ended on October 16, 2019.

In general, all displayed indices are not volatile. Total ads impressions is about 300 thousand (per day). Unique IP-addresses from logged-out users was reduced slightly over the past 4 rounds, but there are no significant changes in the trends.

[Coin-1]

All charts are now updated to the 293th round which ended on December 22, 2019.

As I can see, the number of daily ad impressions has significantly decreased, especially from logged-in users. According to the chart, in October 2019 there was a short-term dip. However, the number of unique IP addresses and unique logged-in users looks more stable.


As far as I understand, during the 10th anniversary art contest there was no advertising on the Bitcointalk forum, so theymos divided the 291st round into 3 periods:

Round 291.6
From Nov 18 2019 14:13:35 UTC to Nov 29 2019 18:34:58 UTC (~11.2 days).
~

Round 291.3
From Nov 7 2019 17:28:22 UTC to Nov 18 2019 14:13:35 UTC (~10.9 days).
~

Round 291
From Oct 27 2019 17:26:14 UTC to Nov 7 2019 17:28:22 UTC (~11 days).

[Coin-1]

Since another 5 rounds ended, I updated all the charts up to the 298th round.

The daily number of ad impressions was slightly reduced to 210 thousand. The number of unique IP addresses from logged-out visitors and the number of unique logged-in users have not changed much over the past two months. About 13500 users view forum threads for each round (~10 days).

[famososMuertos]

The CTRs, are there such statistics that of the clicks generated by each impression.

Us can know (link) "publicly" this impression data in each of the communities.
(?)
Regards.

[Coin-1]

All charts have been updated to the 305th round which ended on May 1, 2020.

As I can see, the number of daily impressions increased from 200K to 300K over the past two months.

The CTRs, are there such statistics that of the clicks generated by each impression.

These statistics are only on ad impressions, not clicks. I assume that advertisers should log the activity on their services and estimate the profit from promotion on Bitcointalk.

As far as I understand, an advertiser should hire a designer here and create an attractive and "eye catching" banner to promote a website on the forum. Afterwards the advertiser should go to the Auctions section and post the bid. Website owners will compete with each other, so the advertiser should offer a price that is high enough to buy at least one ad space (slot) for one round (~10 days). I am sure that the website traffic will increase significantly.

Us can know (link) "publicly" this impression data in each of the communities.
(?)

I guess that you are asking about advertising services on local boards. No, these are common statistics for all sections of the forum.

[Coin-1]

I've updated all statistics on impression counts for ads up to the 312nd round (inclusive) which ended on July 17, 2020.

The number of daily impressions has rebounded to 250000 in the past three months. Visitors read the forum and see ads from approximately 80000 unique IP addresses. 11000 unique members log into their accounts during one round. In general, the current statistics have a positive trend.

[Coin-1]

Almost 3 months have passed, so I decided to update all the charts to reflect the data as of the 320th round which ended on October 17, 2020.

In August 2020, daily impression counts dropped to 190K (on average per ad slot) and were relatively low during that month, but in September 2020, this index has soared again to 300K. Perhaps the reason for the recovery in statistics is the bull run on the cryptocurrency market which occurred at the beginning of last month. People were actively talking about Bitcoin and altcoins on the forum.

[Coin-1]

Since it's been over 4 months, I have actualized all the charts up to the 332nd round which ended on February 26, 2021.

I noticed that in early 2021, the number of unique IP addresses from logged-out users increased significantly, as well as the number of unique users who logged-in for one round. Perhaps people from all over the world wondered why the Bitcoin price skyrocketed to $58K, so they came to the Bitcointalk forum as the first and best place to discuss cryptocurrencies.

[Coin-1]

I have updated all the charts again.

I noticed that theymos recently simplified the structure of the HTML page that lists impressions for ads.

The stats for the 271st round mentioned in the first post of this thread are now displayed in the following format:

Round number	Start	Stop	Duration (days)	Slots	Impressions per slot (all users)	Impressions per slot (logged-in users)	Unique IPs from logged-out users	Unique logged-in users
271	Mar 20 2019 15:49:07 UTC	Mar 30 2019 21:18:53 UTC	10.2	10	3429084	1026072	161241	22787

So I modified my parser and redrawn all the graphs.


On March 29, 2021, theymos also said that ad rounds will now last for ~30 days rather than ~10 days, as it was before:

So your ads will now be up for 29 days instead of ~11 days, and I'm giving more precision about the other times. Future rounds will (according to my current thinking, at least) not always be 29 days long exactly, but will be 30 days plus or minus a few days. The end time for the auction is still arbitrary, but I'm giving a "rough target time": sort of like the center of a probability density function.

Therefore I decided to remove the chart that illustrates the number of ad impressions per one round. I will only maintain daily statistics for the first and second indices.

[DdmrDdmr]

<…>

Ok, that’s important to state, since the first two charts are on the same scale before and after the change in duration from 10 days to 30 days (the both are averaged to a daily scale), whilst the latter two relating to the IPs are not, and the spikes shown over the last few months are due to the cumulous over divergent lengths of time.

Note: Conceptually, the latter two could be averaged likewise by dividing the number of unique IPs by the round’s duration, or separated into before and after the change charts.

[Coin-1]

Another half year has elapsed, so all charts have been updated as of the 346th ad round finished on February 8, 2022.

The current total impressions per day is about 160K, the number of daily impressions from logged-in users is 11250 that is relatively low but looks very stable over the passed 6 months. The other two indices also remained virtually unchanged.

Note: Conceptually, the latter two could be averaged likewise by dividing the number of unique IPs by the round’s duration, or separated into before and after the change charts.

My guess is that most users visit the Bitcointalk forum on a regular basis, at least once a week, so after the ad round has lengthened, the indices illustrated in the latter two charts have no spikes as against the first chart. Therefore, I don't think it's necessary to split the data on unique logged-in users and IP addresses into four charts.