Anonymous Ads - bitcoin advertising network [migration to v.2]

[Bro]

I will try it out on a few parked domains.

I suggest that you create a bitcoin category too, that is your best niche I suppose

[arsenische]

I will try it out on a few parked domains.

I suggest that you create a bitcoin category too, that is your best niche I suppose

Thank you! Are you talking about categorizing ads? (right now there is only a simple filtration to protect affiliates from inappropriate content.. but I agree, it would be nice to have categories for ads and to allow affiliates to subscribe for categories somehow)

[Bro]

Yeah that's what I'm saying.
Somehow both the publisher and the advertiser should be able to chose the niche they think will be more profitable to them. Then how your algorithm takes this information into account is up to you, you can still throw some randomness in the mix since it is your original idea.

[1QaZxSw2]

Yeah that's what I'm saying.
Somehow both the publisher and the advertiser should be able to chose the niche they think will be more profitable to them. Then how your algorithm takes this information into account is up to you, you can still throw some randomness in the mix since it is your original idea.

The following categories pay the best, since these are most likely to get clicks:

Medical: This includes diet pills and weird tip to cut belly fat
Dating: 6 hot women in your zip code want to **** you
Financial: 6 Hot stocks every day!
Gambling: enough said
Jobs: Looking for Java programmers!

Others who may pay moderately:
Local businesses (assuming you can do some Geo-IP)
Local real estate listings
Tourism

But the best idea is to let your advertisers pick or create a category and let affiliates subscribe. Keep the category very broad. "Medical" is better than "doctors, Arthritis, Rogaine", etc. This will keep your service very responsive.

You can of course, allow advertisers to request creation of new category. The reason is sometimes, they want "Pharma" differently from Medical and if you get say 4 votes on a request, you should consider creating it. Do not let affiliates create categories, or you will sink into a huge pile of highly specific categories resulting in very few matches. You dont have either the affiliates or the advertisers to handle it at this stage.

[Tritonio]

If you are going to do this you should consider some things:
1. Categories might increase the complexity. Keep then optional. For example let affiliates choose what they do NOT want to see. For example I know that my users are likely not interested in ads related to medicine.
2. You should probably make a system to edit ads and affiliates to compensate for new ad categories. If a new category is created I might need to exclude it from my affiliate. Maybe you should give an "edit URL" with some sort of random code in the GET parameters. Store the has of that code in the database and when I, or someone else, visit my affiliate link it should check whether the hash of the code in the URL matches with the one in the database, and if it does it should give me options to edit the ad. (just make sure you salt-hash it or the whole thing could be broken with a rainbow table if the codes are short)

[arsenische]

Thanks for this discussion. I see the following extremes:

The simplest solution would be to enhance the filtration functionality (i. e. replace "rating" with a limited set of checkboxes that represent basic categories). It seems to be relatively easy to implement, but it is not very flexible.

The ultimate solution would probably be to allow advertisers to specify tags that describe their ads, and to allow affiliates to exclude tags. But it breaks the whole idea of the system that has nothing to hide and raises security questions.

Maybe the idea with "edit URL" is not that bad. It is not secure at all, but maybe it is not that important for the system that doesn't contain sensitive data and allows to modify filtration settings only. Or is it?

So I am thinking about implementing the following features:
1) allow to specify billing cycle upon ad creation
2) enable some basic ip-based geo-targeting
3) enable filtration based on categories or tags (if based on tags, then enable advertisers to modify tags; show existing tags on affiliate's page and allow affiliates to exclude them)
4) improve graphic versions of ads (make them look like html versions; allow to upload different sizes; allow affiliates to ask for a specific size)
5) add more stats (basically I'd like to see some historical graphs for the whole system and for each affiliate and advertiser)
6) enable search of affiliate, ad or connection between them by bitcoin address, enable navigation through connections
7) show withdrawal history and transaction ids
8) support for smartfones (do they require any specific support?)
...


Probably I need to set up priorities and maybe find a team or an investor.. anyone interested? :)

[Tritonio]

The ultimate solution would probably be to allow advertisers to specify tags that describe their ads, and to allow affiliates to exclude tags. But it breaks the whole idea of the system that has nothing to hide and raises security questions.

What security questions?

Maybe the idea with "edit URL" is not that bad. It is not secure at all, but maybe it is not that important for the system that doesn't contain sensitive data and allows to modify filtration settings only. Or is it?

It's secure as long as affiliates don't loose their edit URL. The edit url isn't stored in the DB so there should be no way for someone to find it. But yes, it created a hackable point in the system while right now it has practically no hackable points.

So I am thinking about implementing the following features:
1) allow to specify billing cycle upon ad creation
2) enable some basic ip-based geo-targeting

MAYBE targeting based on browser language would be better? Not sure but I use a VPS and I am bored of seeing ads about girls in Amsterdam...

3) enable filtration based on categories or tags (if based on tags, then enable advertisers to modify tags; show existing tags on affiliate's page and allow affiliates to exclude them)

It's more important to let affiliates edit excluded tags.

4) improve graphic versions of ads (make them look like html versions; allow to upload different sizes; allow affiliates to ask for a specific size)

Also allow affiliates to block image ads. That would be really useful and it is offered by most ad networks. (and I miss this feature)

5) add more stats (basically I'd like to see some historical graphs for the whole system and for each affiliate and advertiser)
6) enable search of affiliate, ad or connection between them by bitcoin address, enable navigation through connections

Hahaha I was looking at the database yesterday to do that thing manually. I don't think most people need it though but it would be useful. :-)

7) show withdrawal history and transaction ids
support for smartfones (do they require any specific support?)

I don't think you have to do anything.

...


Probably I need to set up priorities and maybe find a team or an investor.. anyone interested?

I wish I could help you make these stuff but I've already got too much in my mind these days. :-|
I hope you find some investor. Maybe you should think about putting anonymous ads on GLBSE? (https://glbse.com/)

BTW I used your ads in a Chrome extension I made: https://chrome.google.com/webstore/detail/fblekdojabihjdhndhmloalbcnnejddl
This couldn't be done with Google Adsense for example. :-D You can't put Adsense in webpages that have no public URL and I had to put it a popup window.

[1QaZxSw2]

Excluding tags sounds important and maybe it is. But it gets in the way of efficient selection and working of the system. Selecting on a match (select where a = b) is far more efficient than selecting on mismatch (select where a not in ('x','y','z'))

Stop thinking in terms of the ad-sense model. It gets on an average 0.25% click rate and is a very inefficient system.

Your business model should not be to show ads. Your business model should be to show ads that the reader is interested in. Otherwise, when most ads are uninteresting, readers will scan the web page and automatically skip over ads. You don't want to display ads that more than 99% of the folks are uninterested in.

In theory, this is what adsense does by reading your page in realtime and categorizing it. Unfortunately, it sucks because if I use the word "sucks" to describe the market, ad-sense will show me ads for gay porn or breastfeeding pads. It is far more efficient to let content owners categorize their content.

Think about this: I have a financial blog. Yes I could say allow all ads except porn. But the truth is that ads for virus cleaners and wrinkle removers are likely to get very few hits and most users will not even glance at that space.

What is far more effective is for me to specify that my content is financial. Advertizers should jump on this because their ads would now be placed on a targeted site where readers are likely to be looking for financial solutions. The ads wont be intrusive and even an occasional interesting ad makes the reader more likely to read the ads on my site rather than scan past it. The click through rates should be far higher (10%+) provided ads are not repeats.

Btw, this is another thing you need to do. Round robin rotate ads shown to the same IP if you arent already doing this. If you have no more ads for the category/IP combination, consider not showing any ads for that IP. The usually clean site that occasionally shows an ad is far likely to get a click than one that always has an ad in the same space.

[1QaZxSw2]

1) allow to specify billing cycle upon ad creation
2) enable some basic ip-based geo-targeting
3) enable filtration based on categories or tags (if based on tags, then enable advertisers to modify tags; show existing tags on affiliate's page and allow affiliates to exclude them)
4) improve graphic versions of ads (make them look like html versions; allow to upload different sizes; allow affiliates to ask for a specific size)
5) add more stats (basically I'd like to see some historical graphs for the whole system and for each affiliate and advertiser)
6) enable search of affiliate, ad or connection between them by bitcoin address, enable navigation through connections
7) show withdrawal history and transaction ids
support for smartfones (do they require any specific support?)
Probably I need to set up priorities and maybe find a team or an investor.. anyone interested?

1) Set withdraw trigger instead.
2) great!
3) see prior post
4-7) great!
Nothing. Just keep ads compact in terms of pixels

I would be glad to help with product development with ideas fwiw (such as this). Investors will come looking for you once you get a few more advertisers.

[arsenische]

It's secure as long as affiliates don't loose their edit URL.

Not that secure:
1) web browsers often send HTTP Referer info to the server, so if you follow some link after you use your edit URL, then the owner of the site that link points to will be able to see in logs your edit URL.
2) nothing is secure over HTTP (and i don't have yet SSL certificate signed by CA to make https.. though probably should get one).

MAYBE targeting based on browser language would be better?

Good idea, need to think it over (maybe could be both)

Also allow affiliates to block image ads. That would be really useful and it is offered by most ad networks. (and I miss this feature)

Why do you miss it? If you use image ads, then it is always graphics; if you use JS or IFRAME version, then it is always text.

I hope you find some investor. Maybe you should think about putting anonymous ads on GLBSE? (https://glbse.com/)

Yes, I am thinking about it.

BTW I used your ads in a Chrome extension I made: https://chrome.google.com/webstore/detail/fblekdojabihjdhndhmloalbcnnejddl
This couldn't be done with Google Adsense for example. :-D You can't put Adsense in webpages that have no public URL and I had to put it a popup window.

Wow, great app, I like it!

If you know how to create browser extensions in other browsers too, you might be interested in LightRider's idea. Do you think it is possible to implement it so that it works in all major browsers? (if yes - maybe it is worth of creating a separate topic to discuss it).

[arsenische]

1) Set withdraw trigger instead.

Yes, withdraw trigger would be a useful additional feature for affiliates. But billing cycle is important too, because it allows to limit the duration of advertising campaign. Right now each advertiser's payment has duration of 30 days (after the last payment is made, the ads will be shown for 30 days, so advertisers have to pay monthly to continue advertising). I think many advertisers would prefer to pay daily or weekly, or to have an intensive but short advertising campaign for just few days (e. g. in case of some event or promotion).

3) see prior post

I don't have Round Robin algorithm there, but I like your idea. Maybe it could be an option for affiliate (to not show repeating ads).

1QaZxSw2 and Tritonio, thank you for your participation!

[Tritonio]

Also allow affiliates to block image ads. That would be really useful and it is offered by most ad networks. (and I miss this feature)

Why do you miss it? If you use image ads, then it is always graphics; if you use JS or IFRAME version, then it is always text.

For example in my extension I couldn't put a <script> tag to use the new ad format. I also can't put an iframe because it's too large to fit it in there. I am now using the <img> version of the ads because it's the perfect size for text ads but when the ad is an actual image it's again to large and I have to shrink it using CSS (max-height) to make it fit in the extension's popup. So it would be better if I could block image ads but use the <img> version. Anyway. I'm not asking that you do this since it is, perhaps, a problem that only I have to face while making Chrome extensions. I'll try a few more things first. I might be able to use the <script> version somehow. :-|

If you know how to create browser extensions in other browsers too, you might be interested in LightRider's idea. Do you think it is possible to implement it so that it works in all major browsers? (if yes - maybe it is worth of creating a separate topic to discuss it).

I haven't made extensions for Firefox and I have no idea how different they are. I suppose they ain't that different... I'll read more about his idea although I have trouble finding a way to measure how much you "own" to each producer. I'll think about it. :-)

[Bro]

I received my first bitcent, nice 

[arsenische]

New feature: advertisers can specify billing cycle (in days) thus defining the pace and time limit of their advertising campaign.

"Bootstrap score" is now calculated as <amount of money deposited during last billing cycle> / <duration of billing cycle>.

Instead of amount paid to affiliates during last 30 days, "Reward score" is shown, that is calculated as <amount of money paid to affiliates during last billing cycle> / <duration of billing cycle>.

Not sure if concepts of "Bootstrap score" and "Reward score" make sense for anybody except me

[Tritonio]

New feature: advertisers can specify billing cycle (in days) thus defining the pace and time limit of their advertising campaign.

"Bootstrap score" is now calculated as <amount of money deposited during last billing cycle> / <duration of billing cycle>.

Instead of amount paid to affiliates during last 30 days, "Reward score" is shown, that is calculated as <amount of money paid to affiliates during last billing cycle> / <duration of billing cycle>.

Not sure if concepts of "Bootstrap score" and "Reward score" make sense for anybody except me

As I get it:

Bootstrap score = average amount paid in bootstrap address per day during the current cycle. Dictates how often you will be shown in that 10% of each affiliate's views. The higher this is, the more new affiliates you will try.

Reward score = average amount paid to affiliates per day, during the current billing cycle. Dictates how often you appear in affiliates in the other 90% of the pageviews. We are more interested in the per affiliate reward score since that dictates how often will each affiliate will show your ads.

Are you sure you really mean "last cycle" and not the currently running cycle? I suppose you meant "current cycle"

[arsenische]

Bootstrap score = average amount paid in bootstrap address per day during the current cycle. Dictates how often you will be shown in that 10% of each affiliate's views. The higher this is, the more new affiliates you will try.

Reward score = average amount paid to affiliates per day, during the current billing cycle. Dictates how often you appear in affiliates in the other 90% of the pageviews. We are more interested in the per affiliate reward score since that dictates how often will each affiliate will show your ads.

Are you sure you really mean "last cycle" and not the currently running cycle? I suppose you meant "current cycle"

Also, "Reward score" is different from "Bootstrap score": Bootstrap score is used directly in the ad selection algorithm for 10% of impressions, whereas Reward score is just an indicator, it shows de-facto speed of paying money to affiliates. For 90% impressions ads are selected according to their de-facto CPI for particular affiliates during the billing cycle (which is related to Reward score, but not exactly the same thing).

When I say "during last billing cycle", I mean "during last <billing cycle> days". Is it the same as "current billing cycle"? For example, if you have a billing cycle of 5 days, the history of bootstrap score changes could look like that:

Day	Deposited	Bootstrap score
1	0 btc	0 btc
2	1 btc	0.2 btc
3	0 btc	0.2 btc
4	0 btc	0.2 btc
5	3 btc	0.8 btc
6	0 btc	0.8 btc
7	0 btc	0.6 btc
8	0 btc	0.6 btc
9	0 btc	0.6 btc
10	0 btc	0.0 btc

Do you think if I replace "last billing cycle" with "current billing cycle", it will be clear? Or is the "Billing cycle" term confusing? Probably the "scores" are confusing too...  

[Gary13579]

Deposited 30.49293495 btc thus gaining ~60955 impressions, CPM = 0.50025076 btc
Paid 10.65445235 btc to affiliates thus gaining ~548597 impressions, CPM = 0.01942125 btc

am I right in reading this to mean that 30 BTC has been paid for bootstrapping, while only 10 BTC has been paid to affiliates? I know/understand that AA needs bootstrap BTC for income, but these numbers seem waaaay too out of whack. AA getting 75% of the BTC seems a bit much, doesn't it? I understand that currently you are using bootstrap money to pay affiliates (according to one of your previous posts in this thread), but if the 3:1 trend continues, do you plan on continuing to pump this money into affiliates? And I think there should be more transparency as to where those coins go... if I look at the affiliate section I see:

Brought 609553 impressions, 8028 clicks, CTR = 0.01317
Earned: 10.65445235 btc
Withdrew: 10.04922624 btc
Paid fees: 0.32200000 btc
Left to withdraw: 0.28322611 btc

Which to me seems to imply we have only earned what advertisers have paid us, and that we've gotten none of the 30 BTC from bootstrapping.

[arsenische]

Deposited 30.49293495 btc thus gaining ~60955 impressions, CPM = 0.50025076 btc
Paid 10.65445235 btc to affiliates thus gaining ~548597 impressions, CPM = 0.01942125 btc

am I right in reading this to mean that 30 BTC has been paid for bootstrapping, while only 10 BTC has been paid to affiliates? I know/understand that AA needs bootstrap BTC for income, but these numbers seem waaaay too out of whack. AA getting 75% of the BTC seems a bit much, doesn't it?

Thank you for your question and feedback. The system is yet in its infancy, it is too early to judge.

Just several days ago statistics looked very different. Amount deposited to bootstrap ads was about the same as amount paid to affiliates. Most of the money paid to affiliates were paid by me on behalf of advertisers (they show up in statistics as if they were paid by advertisers).

But several days ago 20 btc was deposited to bootstrap the ad of Dragon's tale, and it influenced statistics a lot. I am thankful to this advertiser because I have some spare btc now to support affiliates and to promote the service. Bootstrap deposits work for the whole billing cycle (30 days by default), but affiliates need to be supported on the daily basis.

Also it seems that advertisers started paying to affiliates themselves. For example, the ad of SteamCoin already paid to affiliates more than to anonymous ads. And we can see that it gets more unique impressions for less money.

It is good for affiliates, because I was paying them only ~0.3 btc daily, whereas advertisers pay much more.

I think that generally it is not a good idea for me to support affiliates on behalf of advertisers, because:
1) I don't know what visits are successful, only advertisers know it and can pay for it... If I pay, then I give incentive for fake clicks.
2) I don't want to interfere with advertisers that do it themselves and to reduce efficiency of their advertising campaign.

So I expect that in the long run affiliates will get much more than Anonymous Ads, and I will not have to support affiliates on behalf of advertisers at all.

Which to me seems to imply we have only earned what advertisers have paid us, and that we've gotten none of the 30 BTC from bootstrapping.

Bootstrapping money are generally not supposed to go to affiliates (except 10% referral payment to affiliate that attracted the advertiser). But it is not correct to say that affiliates didn't get anything from 30 BTC. I used this money to support them and will be using if advertisers won't.

[Gary13579]

Thanks for your response. I still think there needs to be more, or at least more clear, documentation on how it works. I don't see why Dragon's Tale would have dumped 20 BTC into bootstrapping when they could have spent that money on affiliates and gotten more impressions/clicks for those BTC. My thoughts are that they didn't entirely understand how the system works?

It all seems rather complicated when I first approached it, and it took a while to wrap my head around it, as this network is so vastly different compared to existing ad networks.

[arsenische]

Thanks for your response. I still think there needs to be more, or at least more clear, documentation on how it works. I don't see why Dragon's Tale would have dumped 20 BTC into bootstrapping when they could have spent that money on affiliates and gotten more impressions/clicks for those BTC. My thoughts are that they didn't entirely understand how the system works?

It all seems rather complicated when I first approached it, and it took a while to wrap my head around it, as this network is so vastly different compared to existing ad networks.

Yes, I agree with you. I'd like either to make a better documentation, or to simplify the system, or both. Not sure how to do it yet.

Regarding Dragon's Tale, I will ensure that it receives an adequate traffic for its money so that its anonymous advertiser is not offended, and affiliates get paid.