Hi, thanks for your question. There are several factors that influence. The algorithm for calculation of CPM is pretty standard (except we calculate it based on globally unique impressions): 1000.0*<ad unit's income>/<ad unit's globally unique clicks>. The algorithm for money distribution is not trivial. Ad unit's income depends on a) targeting and budgets of advertising campaigns; b) competition for advertising budgets c) the relative volume and quality of ad unit's traffic (the 'quality' part is tricky and I am not going to explain all the nuances here).

Thanks for picking good examples, ad units #10539 and #67832 are quite similar. They both belong to Bitcoin and Earning online categories, they both have about the same amount of traffic, and I would also say that #10539 is embedded on a nicer site. Why does it earn ~7 times less than #67832? Let's have a closer look.

1) See the list of advertising campaigns that advertise on ad unit #67832:

(https://a-ads.com/ad_units/67832/campaigns)

Almost half of its income is generated by campaign #6468 that targets English-speaking non-US traffic.

Since ad unit #10539 has significant share of US traffic, it doesn't fit the targeting criteria of that campaign and earns significantly less.

2) See the stats of ad unit #10539:

(https://a-ads.com/ad_units/10539/stats)

You can see that it had ~1200 non-unique clicks (~400 unique clicks) a day until we deployed our solution to filter out bot clicks. When visitors click the banner then with a certain probability we display captcha and count the click only if the captcha is passed. If captcha is not passed, then the probability of captcha for this ad unit is increased.

At the moment ad unit #67832 displays captcha with probability of 1% since most of the clicks are made by humans, that's how it is supposed to work for every ad unit.

However ad unit #10539 displays captcha with probability of 99.9% since most clicks are probably made by bots. So the bot clicks are not counted and the traffic quality is considered low.

I hope 1) and 2) explain why ad unit #10539 earns much less than #67832 despite the fact that it has about the same amount of traffic and a nicer site.

Thanks for your idea, I implemented it (see http://bitcoinocracy.com/arguments/bitcoin-future-should-be-decided-by-miners).

Also added friendly URLs, social buttons and other minor improvements.


Thanks for your kind words. I hope to find volunteers to develop this project (to do programming, design, content creation). Still not sure whether it is useful.  Perhaps sometimes it could be an alternative to miners' voting or public discussions (that are often either censored, or have low signal-to-noise ratio).

If you want to add a new banner size to the existing campaign then you need to create a new ad, upload all the banner sizes to it and then link that ad to your campaign (you can do it in campaign's settings).

If you want to add a new advertising campaign, then you can use campaign creation wizard (it will create an ad and a campaign at the same time).

Ranlo is right, we don't have user accounts, but you can use the same refund address for your campaigns and then find them using our site-wide search. If you sign in with that bitcoin address, you will see them too.

I queried the database to see what kind of statistics we got and found out that over ~1K ad units generate > 0 captcha requests per day (daily estimates are calculated as: today_fraction*today_value_of_metric+(1-today_fraction)*yesterday_value_of_metric).

An average ad unit in this set has:
  captcha display rate = 47%
  captcha pass rate = 46%

If we assume that all ad units generate only real traffic, then 47%*(100%-46%)= 25.4% of visitors who clicked the ad on the average ad unit were stopped by captcha. That is the most pessimistic expectation.  Actual stats should look much nicer since we display captchas more often to ad units that rarely pass them and also many ad units are probably generating only bot visits thus distorting the general picture.

I've split this set of ad units into 5 groups:
A) ~42% with captcha pass rate = 0%
B) ~10% with captcha pass rate in (0..50%) range
C) ~5% with captcha pass rate = 50%
D) ~12% with captcha pass rate in (50..100%) range
E) ~31% with captcha pass rate = 100%

I think ad units of the group D are the most interesting since most of their traffic is confirmed to be human and they have more than 2 captcha requests per day.

An average ad unit in this group has:
  captcha display rate = ~1%
  captcha pass rate = ~75%

If we assume that all ad units in group D generate only real traffic, then 1%*(100%-75%)=0.25% of visitors who clicked the ad were stopped by captchas. This is less significant than daily fluctuations of traffic and our CPM (that we aim to improve by improving the quality of traffic our advertisers receive).

Some more stats (just to make more sense of the data):

~49% of all ad units that have click attempts, have captcha display rate of ~1%

~84% of all ad units that have successful captcha passes, have captcha display rate of ~1%

Ad units with > 10 daily unique clicks and captcha pass rate > 30%, have average captcha pass rate of ~74% and average captcha display rate of ~2%.

It doesn't appear in an ad, it appears in new browser tab with some probability after you click it (the probability depenends on past captcha performance of this ad unit and on campaign settings). So the network-wide click stats you see now should be more adequate (there are ~5K, not ~1M of real clicks per day).

If significant amount of users refuse to pass the captcha, advertisers' conversions should drop. If there are any advertisers in this thread - please let us know whether that is the case. I received 1 complaint from an advertiser from Russia, who noticed the decreased amount of clicks. But I'd be thankful for feedback from other advertisers as well (preferably English-speaking).

If conversions suffer significantly then we'll need to roll-back the change. Otherwise I think we should allow publishers to specify max captcha rate for their ad units. Publishers will be able to set it to 0, but if campaign's captcha rate is higher than ad unit's, then ad unit won't earn from that campaign. It may take couple of days to develop, test and deploy it.

How often do you think an average visitor clicks the ad? Do you think that clicking the checkbox is that annoying? Why do you think that it will have significant impact on your user loyalty?


That's true, our publishers are the key component of our business. We want to make them happy and we appreciate their loyalty despite the low revenue. We think that they deserve to earn more than they currently do.  Or do you think we are trying to solve non-existent problem? Publishers, are you happy with your current level of income from a-ads?

Publishers' earnings depend on amounts of money spent by advertisers. That's why we want to make our service look good for advertisers.

When advertisers realize that we have 50M impressions from 1M unique addresses and 2M clicks, many of them just leave because they understand that most of this traffic is fake and it doesn't worth it. Yes, they don't pay per click, but not all of them understand that. And they don't know how many real clicks they should expect to get when they set their advertising budgets. The numbers we showed so far (like hundreds of clicks for 0.01$ in btc) were too good to be true.

Also in our system fake traffic makes more damage to legit publishers than to advertisers because it reduces their share of the advertising budgets. So even if this feature doesn't increase advertising budgets, it would allow us to reduce the income of the fraudulent publishers and increase the earnings of the honest ones.

Why do you worry about captchas so much? We don't plan to display it every time user clicks an ad. Less than 1% of your visitors are likely to ever see it, and most of them will probably pass it by just clicking the checkbox. The negative effects are likely to be negligible in comparison with the possible gains.

Yes, that's somewhat similar to 90s except the no-captcha recaptcha is much more convenient and easy to solve.

An alternative solution would be to invade visitors' privacy with cookies and javascript. Perhaps we could change the money distribution principles and switch to some form of RTB network. I feel that it would be damaging to the original idea of user-agnostic service, but perhaps it would be a better experience and better monetization.

Do you think it would be better? Or is it better to stay in 90s with no javascript and no cookies?

We need to gather more feedback from publishers and to see whether the captchas influence advertising budgets and traffic conversion. If it doesn't solve the problems it is intended to solve - we'll probably rollback this change.

Are there any publishers that find my arguments reasonable, or is there a consensus that this solution is really, really bad?

Oh, sorry for missing your message. It is hard to say in advance how much you can earn. Much depends on the category of your site, and whether we've seen IP addresses of the audience of your site during the last 24 hours.

Our network-wide averages with CPM estimations are here: https://a-ads.com/stats -- but your earnings could be much more or much less than that. I'd suggest to try us our for several days and then decide whether it worths it.


What went wrong with CoinURL? I don't know their story.


To be frank with you, I don't know how to conduct a case study. It sounds hard, perhaps the easier solution could be to let publishers to disable captchas for their ad units. If they do so, they won't get money from the advertisers who want 100% human traffic. On the other hand other advertisers might benefit from it.

Thank you very much for the detailed explanation, but how likely is this scenario?

Let's say (X*100)% of your audience doesn't use AdBlock, (Y*100)% of them click the ad, (Z*100)% of them sees the captcha and (W*100)% of them gets annoyed in such a manner that it really damages your site (e. g. they planned to bookmark it, but due to the easy-to-pass captcha opened in the new browser tab upon their ad click they decided not to do so).

So publishers should expect to get R times higher revenue from advertisers at the cost of C=(X*Y*Z*W)*100.0% of annoyed customers.

What do you think are the realistic values of X, Y, Z, W?

According to my estimations, R is multiple orders of magnitude higher than C, doesn't it worth it? Advertisers can put their own captcha wall on their sites or advertise intersistial ads anyway (and some of them do).

PS: Advertisers that set captcha rate to 100% are potentially loosing (W*100)% of their conversions (or more, since some people won't get annoyed in a damaging way, but will just close the browser tab). Also the visitors might attribute the captcha to their sites as well (since it is opened in the new browser tab upon click on their ad).

Thanks for your feedback!

I hate captchas too and our original idea was that advertisers use analytics and distinguish/reward good publishers themselves.

But many advertisers complained that most of the clicks counted by our system either didn't reach their sites, or were just "pings" with ~100% bounce rate. Of course some of the clicks were real, but advertisers' frustration was so high that they didn't use the service long enough to get real customers and they withdrew their money. Thus publishers suffered from the decreased earnings.

Please note that advertisers don't pay per click, so when they enable captchas, they are potentially decreasing their own conversion, not the publishers' income.

Publishers don't earn per click but if their CTR is too low then they get a smaller share of the advertising budget. I expect that the earnings of fraudulent publishers will decrease but earnings of legit ones will increase.

We are collecting statistics about publishers' ad units that allows us to show captchas more frequently to the ones that fake traffic and to better distribute the funds among publishers.

In fact, this feature is introduced to increase monetization of real traffic and thus is supposed to be good for legit publishers. But why do you think that publishers suffer?

Deployed a new feature: Captcha rate for advertising campaigns.

Since we are not pay-per-click advertising network, fake clicks don't affect advertisers' expenses. But they distort advertisers' statistics, that is why we implemented a new feature that enables our advertisers to filter out bot generated clicks via captcha.

Advertisers can specify the Captcha rate for their campaigns. The default value is 0, but if you want to check all the visitors before they arrive to your site, you can set it to 100%.

Although some people might refuse to pass the captcha, we don't expect significant decrease of conversion since this kind of captcha is usually easy to pass.



We expect a significant decrease of unique clicks registered by our servers because they will be filtered out by captchas.

http://blog.anonymousads.com/2015/08/new-feature-captcha-rate-for-advertisers.html

That's how it works now, it caches the balance in our database and if you spend your funds after making a vote, it won't decrease the weight of the vote.

I consider it to be a bug that should be fixed, because it makes it trivial to manipulate voting results by moving your funds from one bitcoin addresses to another inside your own wallet and thus generate unlimited amount of "valuable" signatures.

It is important to fetch up-to-date data from the Blockchain on a regular basis (and it is easy to implement).

That's true, but if the issue is important then another semi-whale should join the party, or a huge mass of people will try to oppose with their small balances. That would be interesting to watch.

Though any voter can easily split one's balance and pretend to be a crowd of voters, that's why the balance is the only reliable measure, not the amount of bitcoin signatures.



Money have something to say, I'd like to hear that (though probably we hear it all the time from the mass media, but not in such a transparent and accurate manner).


Thanks for you suggestion! Do you suggest to display: "9.12659522 Ƀ Believe |  0.0001 Ƀ Doubt" as "9.12659522 Ƀ (0.998904313%) Believe | 0.0001 Ƀ (0.001095687%) Doubt" ?


Thanks for your feedback, I totally agree with you. Someone (preferably a native speaker) should create description of how Bitcoinocracy works and how to make Bitcoin signatures. Long ago I made a brief explanation of how to sign a message with popular bitcoin clients (Bitcoin-QT, Multi-Bit, Electrum, Blockchain.info) for other service, probably we need something similar for Bitcoinocracy.


Yes, it is supposed to change as the balance changes and it is easy to implement, just didn't do it yet.

There are a lot of things to be done. Friendly URLs, social sharing buttons, bitcoin balance updater task, create content, etc.

The current state of the project is just proof of concept.  Feel free to join the project and contribute in any form

Sign "I believe that Block size limit should be increased to 8 mb as soon as possible" with your bitcoin address to show your support. Or click "Doubt" and sign "I doubt that Block size limit should be increased to 8 mb as soon as possible". We'll see what part of economy supports this statement.



The results are verifiable, here is the list of signatures for and against the argument.

Bitcoinocracy is an opensource project. If you think it is useful - please join or use it so that I feel interest and find time to develop it.

Hi, thanks for your message. Please let me know your campaign id.

Oh, sorry for inconvenience. Could you please tell me your campaign id and/or bitcoin transaction id?

Thanks for your message. Sorry for inconvenience. I think this function should work properly in most browsers. The problem was probably related to the changes in configuration of our server and if you clear your cache and cookies and start again, it should work from your device as well. If not - please let me know the version of your browser, I'll try to reproduce the problem so that we can fix and test it.

Thanks for reporting! Unfortunately I don't know how to reproduce it. Are there any errors in javascript console? Perhaps your browser cached the invalid server response. Could you please clear the cache/cookies and try again, or perhaps try from a different browser/PC? Does anybody else experience this problem? Many thanks!


Thank you very much, sorry for the inconvenience!

Thanks for your message! I think it is fixed, please try again.

Sale of traffic! The same traffic is currently 2-3 times cheaper than it was a couple of weeks ago. Hurry up while it is cheap, create your campaign without registration in less than a minute! http://a-ads.com/campaigns/new

Hi, thank you very much for your message. We see some error messages and are working to fix them. Please PM me the exact page URL / action that fails. Many thanks!

Update: I think the bugs have been fixed, please try again.