How traffic exchanges resell autosurfing and where from the network millions of bots
It was two years ago, I, as a hobby, was engaged in arbitration and tested traffic exchanges for various exchanges of traffic (and not only him). After talking with colleagues, I decided to pour traffic on gambling and casinos. Of course, I decided to try on the traffic, as the familiar language, respectively, is easier to make landing pages and it is easier to understand the audience.
I started by searching for offers, I found several casinos with affiliate programs. Affiliate programs usually have offers with ready-made landing pages, but the problem is that they have landing pages on servers, therefore conversion statistics and sources cannot be obtained from them. I understood that if one arbitrates seriously, then one should take a serious approach, with analytics. It was decided to copy the landing pages of affiliates, slightly modify them externally and, most importantly, fasten the goals and the webviewer metrics so that you can track all advertising platforms and conversions from them.
About a week later, all the landing pages were ready to be launched, the metric with the goals was set to “every sneeze” of the visitor and the decision was made to start the traffic.
I launch 10,000 popander traffic on one of the landing pages. Turning off pretty quickly, I set limits on the number of hits, but still it took about an hour for all 10k hits. By the way, 10k hits are about 2k visitors registered in the metric, the remaining 8k either closed the popander before the metric was loaded, or just bots (although on the exchanges you will be assured that the bots have exactly zero percent and the protection system is so super cool that already it becomes scary).
To be honest, I was very surprised when I saw that from these 2k "get" visitors there were 2 leads to register with the casino. The math was such that, having spent $ 10 per 10k of traffic, I got 2 leads. For statistics, not enough, so I went ahead and began to spend more.
Since my goals were set up on the Lands, I saw what users were doing when they hit the popander of my land. I do not know why I was not embarrassed that there were a lot of clicks on buttons, form fillings and other conversions that I set up, and users spent long minutes on sites. At the same time there was a clear dependence, the more traffic I buy at a time, the less real conversions I get. Having bought 10k of traffic, I get 2-3k users who have reached the metric load and 2 conversions on average, having bought 30k traffic, I get 6-8k users and 2-4 conversions. I could not understand what was happening and tried to track down good sites from which it is better to buy traffic.
First of all, I filtered the platforms inside the popander exchanges by regionality and the leads that were from them for secondary purposes which showed that the audience was interested in the offer.
I worked the system. Disconnected sites that did not pass at the cost of the lead or a secondary goal, the entire infa on conversions was taken from the metric. Disconnecting sites that exited at a cost of payback after a certain unscrewing of traffic, the exchange automatically included new, not yet tested, sites and continued to spend money on those sites where leads and secondary goals cost no more than limits.
It seems to be an ideal system that must have worked. Run!
It was decided to spend no more than $ 300 for the test. For 100 bucks a day.
After two days and $ 210 was unscrewed 210 000 traffic I, in anticipation of a brilliant result, went to check the leads.
9 registrations, it was a failure. There was no talk about payback. It is worth noting that they pay in these offers not for registration, but for the money made by the player, I had a little more than $ 30 in the account in the affiliate program, including all previous attempts.
I did not understand what was happening. In expectations, there were 80-120 registrations, taking into account the choice of only high-quality sites and an estimated conversion of 0.04% -0.06%.
It is useful to understand the metric. With user behavior, everything is fine, they sit on the Lands for more than 30 seconds on average, which seemed to me a very good indicator. Watch the video "how to beat the casino", click on internal links and so on in the same spirit.
Why are such excellent behavioral factors and so few conversions !? And this is despite the fact that the traffic is 10 times less than the conversion result was excellent.
I already wanted to drop everything and remove to hell, but curiosity stopped me. It became very interesting for me to find out who these people are from whom my popanders open up. I was sure that these were people because it was impossible to determine from the behavior that these were bots.
Since I have been in internet marketing for a long time, I knew that there is such a thing as social tracking. Social tracking - various ways to find out public information about users, for the most part I was interested in profiles in social networks. I found a working tracking, copied their ID identification system in VC and Od (copied, and did not buy access from them, because it is expensive to conduct research on 10k users directly through their system).
I made a simple interface, put it on one of the landings and launched an advertisement from one specific exchange. What was my next surprise when I saw that a person with the same profile in the social network comes to my site 10 times in a few minutes, while he has different user agents, ip, screen resolution ... As a result, the drain 1000 traffic, I realized that the site really was a very small number of people with a unique ID in the VC or OD, and 95% of them simulate excellent behavioral factors and visit the site 20-50 times per hour.
"What a wonderful bots!" - I thought. I decided to look at the profiles of these “people” in VK and ML and several times I came across advertisements of the same autosurf service. To my shame, I still did not know what it was. I was interested in these reposts, and, suspecting something was wrong, I was useful to study.
As a result, I registered on one of the autosurf exchanges and merged some traffic from there to my land with a stream of social tracking. Once again, to my surprise, there was no limit, I met several of the same "people" who imitated a lot of visits with different parameters, but all with the same ID from VC or OD. He began to pour even more traffic and even stumbled upon the very same “visitors” with the same social networks that came in with a particular stock ponder.
Having concluded that this popander exchange simply resells autosurf traffic, I decided to test the rest I worked with. I found a small amount of traffic autosurfing (about 30%) on one, up to 95% didn’t have a traceable ID in VC or OD on the rest, which is why I concluded that almost all traffic was bots, although they had excellent PF rates. It is worth noting that in the IDE recognition system I was extremely confident that on sites where contextual advertising went, it tracked visitors with a very high degree of probability.
Since then, I no longer work with exchanges where traffic is cheaper than $ 5 per 1000 impressions, since everything that was tested with a price of less than $ 5 (specifically in the popup, popander, clickander) was either bots or autosurfing.
I think that with the help of such a traffic you can try to wind up the PF for search engines :)
Training
I started by searching for offers, I found several casinos with affiliate programs. Affiliate programs usually have offers with ready-made landing pages, but the problem is that they have landing pages on servers, therefore conversion statistics and sources cannot be obtained from them. I understood that if one arbitrates seriously, then one should take a serious approach, with analytics. It was decided to copy the landing pages of affiliates, slightly modify them externally and, most importantly, fasten the goals and the webviewer metrics so that you can track all advertising platforms and conversions from them.
About a week later, all the landing pages were ready to be launched, the metric with the goals was set to “every sneeze” of the visitor and the decision was made to start the traffic.
Launch
I launch 10,000 popander traffic on one of the landing pages. Turning off pretty quickly, I set limits on the number of hits, but still it took about an hour for all 10k hits. By the way, 10k hits are about 2k visitors registered in the metric, the remaining 8k either closed the popander before the metric was loaded, or just bots (although on the exchanges you will be assured that the bots have exactly zero percent and the protection system is so super cool that already it becomes scary).
To be honest, I was very surprised when I saw that from these 2k "get" visitors there were 2 leads to register with the casino. The math was such that, having spent $ 10 per 10k of traffic, I got 2 leads. For statistics, not enough, so I went ahead and began to spend more.
How did the tests
Since my goals were set up on the Lands, I saw what users were doing when they hit the popander of my land. I do not know why I was not embarrassed that there were a lot of clicks on buttons, form fillings and other conversions that I set up, and users spent long minutes on sites. At the same time there was a clear dependence, the more traffic I buy at a time, the less real conversions I get. Having bought 10k of traffic, I get 2-3k users who have reached the metric load and 2 conversions on average, having bought 30k traffic, I get 6-8k users and 2-4 conversions. I could not understand what was happening and tried to track down good sites from which it is better to buy traffic.
First of all, I filtered the platforms inside the popander exchanges by regionality and the leads that were from them for secondary purposes which showed that the audience was interested in the offer.
I worked the system. Disconnected sites that did not pass at the cost of the lead or a secondary goal, the entire infa on conversions was taken from the metric. Disconnecting sites that exited at a cost of payback after a certain unscrewing of traffic, the exchange automatically included new, not yet tested, sites and continued to spend money on those sites where leads and secondary goals cost no more than limits.
It seems to be an ideal system that must have worked. Run!
It was decided to spend no more than $ 300 for the test. For 100 bucks a day.
After two days and $ 210 was unscrewed 210 000 traffic I, in anticipation of a brilliant result, went to check the leads.
9 registrations, it was a failure. There was no talk about payback. It is worth noting that they pay in these offers not for registration, but for the money made by the player, I had a little more than $ 30 in the account in the affiliate program, including all previous attempts.
Where are so many bots !?
I did not understand what was happening. In expectations, there were 80-120 registrations, taking into account the choice of only high-quality sites and an estimated conversion of 0.04% -0.06%.
It is useful to understand the metric. With user behavior, everything is fine, they sit on the Lands for more than 30 seconds on average, which seemed to me a very good indicator. Watch the video "how to beat the casino", click on internal links and so on in the same spirit.
Why are such excellent behavioral factors and so few conversions !? And this is despite the fact that the traffic is 10 times less than the conversion result was excellent.
I already wanted to drop everything and remove to hell, but curiosity stopped me. It became very interesting for me to find out who these people are from whom my popanders open up. I was sure that these were people because it was impossible to determine from the behavior that these were bots.
Since I have been in internet marketing for a long time, I knew that there is such a thing as social tracking. Social tracking - various ways to find out public information about users, for the most part I was interested in profiles in social networks. I found a working tracking, copied their ID identification system in VC and Od (copied, and did not buy access from them, because it is expensive to conduct research on 10k users directly through their system).
I made a simple interface, put it on one of the landings and launched an advertisement from one specific exchange. What was my next surprise when I saw that a person with the same profile in the social network comes to my site 10 times in a few minutes, while he has different user agents, ip, screen resolution ... As a result, the drain 1000 traffic, I realized that the site really was a very small number of people with a unique ID in the VC or OD, and 95% of them simulate excellent behavioral factors and visit the site 20-50 times per hour.
"What a wonderful bots!" - I thought. I decided to look at the profiles of these “people” in VK and ML and several times I came across advertisements of the same autosurf service. To my shame, I still did not know what it was. I was interested in these reposts, and, suspecting something was wrong, I was useful to study.
Autosurfing
As a result, I registered on one of the autosurf exchanges and merged some traffic from there to my land with a stream of social tracking. Once again, to my surprise, there was no limit, I met several of the same "people" who imitated a lot of visits with different parameters, but all with the same ID from VC or OD. He began to pour even more traffic and even stumbled upon the very same “visitors” with the same social networks that came in with a particular stock ponder.
Having concluded that this popander exchange simply resells autosurf traffic, I decided to test the rest I worked with. I found a small amount of traffic autosurfing (about 30%) on one, up to 95% didn’t have a traceable ID in VC or OD on the rest, which is why I concluded that almost all traffic was bots, although they had excellent PF rates. It is worth noting that in the IDE recognition system I was extremely confident that on sites where contextual advertising went, it tracked visitors with a very high degree of probability.
Since then, I no longer work with exchanges where traffic is cheaper than $ 5 per 1000 impressions, since everything that was tested with a price of less than $ 5 (specifically in the popup, popander, clickander) was either bots or autosurfing.
I think that with the help of such a traffic you can try to wind up the PF for search engines :)