The Most Active and Friendliest
Affiliate Marketing Community Online!

“Adavice”/  “1Win

Guide: How to clean the campaign from bot traffic

Affmy

Affiliate Manager
Affiliate Manager
Affmy
PXNbMCP.jpg





How to clean the campaign from bot traffic



A single bot is just a computer programm, which is often rather simple. Still, it can cause a lot of harm to your campaign and lead to big losses. So, it would be a great idea first to identify them, and then to cut these waste traffic. The whole issue isn't as complicated as it looks.



Tips to detect bots



A great many bots are so simple that they don't have the means to operate undetected. Use it to detect these harmful programs. First, they often do not run Javascript, while most of your real visitors do. Then, bots might download random pages without any analysis of info they contain.



Second, they are a piece of code and don’t need time to look through the page. The third issue is that humans have a longer reaction time when reading the pages. Bots are much faster, if they don’t copy human’s behavior.



The mentioned principles of detecting the “campaign pests” are simple, but they do work. If you think your traffic is full of bots, just follow them to unveil these harmful programs. For example, when you’re getting a lot of traffic without any conversion even after optimization, it is the first signal to start “bot detection”.



You should also understand that there are much more complicated and smart programs and you won’t detect them using the mentioned methods. However, they’ll help you to clean your campaign from a great variety of simple bots.




How to understand that your landing page is a dwelling for bots?



There is a method that will help you to see, if there is any waste traffic on your lander. Note: it won’t show the source of bots, just the fact they are here. To get the results you should run a very simple test that will not harm your landing page.



The first thing you need is to create a transparent 1x1 format GIF. Then just upload it to your server or any image site. The next step is a personal account on Bit.ly, which will give you the statistics. Also, create a shortened link of your image. After that you should copy the following code and paste it before the tag “body” of your landing page.



<script type='text/javascript'>

function changeimage()<

document.getElementById("myimage").src="Your Bit.ly Link";

>

</script>



The part "Your Bit.ly Link" must be changed for the actual link you’ve already created. The second part of the code should be added directly to the tag “body”.



onload="setTimeout(changeimage, 300);"



At last, the third section must be copied and pasted right to the bottom of the landing page.



<img src="broken" id="myimage">



Before all the preparations you need to check the current number of visitors. Then wait for another 100 visitors after the test was launched. In the end, visit Bit.ly to get the “click” statistics.


To be continued…
 
PXNbMCP.jpg




What does it all mean?


The mentioned test uses the difference between human reaction time and the one for bots. It uses Javascript and downloads all the images in 300 ms and does it when the page is already downloaded. Individuals are not as fast as the bots, they most likely won't close the page before the test is finished. Also, there are not many internet users that disable Javascript.


If you take the above items and the descriptions in the above paragraphs, it is a 99% possibility, that when the page was downloaded, but the image not, it was a bot. It’s very easy to calculate the percentage of waste traffic using this approach. The formula looks in the following way:



Only page downloads without the image/the total number of visits*100


This method is great to show you, if there is a bot traffic you receive, but it won’t tell you the source of such traffic. So, if you want to find it, you’ll have to use another approach.


How to identify “clean” placements?


The next test can be launched even with the campaign that is already running. It can show you, what is what is worth targeting on a new traffic source and what is not. There are four simple steps to launch the test. First, create a new text file and copy the following code there:


<script type='text/javascript'>

function redir(){

window.location.replace('CTA_Link_of_your_tracker');

}

</script>

</head>

<body onload="setTimeout(redir, 300);">

</body>


Then replace the 'CTA_Link_of_your_tracker' part with the actual CTA link of the tracker you’d always use for the lander. Second, save the file with the name like project.html and upload it to the hosting of your landing page. Third, start a campaign on any source of traffic. No matter which offer you’ll use, but set project.html as the only lander.


At last, launch the campaign and observe. Whenever any placement gets 60 clicks, go to another one and pause the previous. You’ll have to test a bit more than 20 placements. Why 60 clicks? Well, just because it’s enough data.


What can the test tell you?


The mentioned code does a very interesting and simple thing. It redirects the visitors to the lander only when 300 ms have passed (using Javascript). Then it comes to the CTR. Real humans don't have the fast reaction time of a bot, so they’ll be “clickthrough”, while the bots won’t. Actually, anyone who didn’t make a click is probably not a real person.


The test you’ve used is very versatile. You can apply it for countries, carriers, phone models, etc. The approach will be the same for all the cases. Different is only the thing you’re testing.


What to do with the traffic from bots


If the tests shows your placements are full of waste traffic, you should pause them immediately. Then some analysis is needed. The question is about the percentage of bot traffic and the possibility of making any money on these placements. If 20% or more of your visitors are not real humans, it’s unlikely that you’ll receive any profits on such placements. However, there might be some exceptions, but still, be sure it’s worth to deal with them.


The other thing you should definitely do is try to get a refund. In case it’s possible to show you have bot traffic using IP logs, contact you source of traffic for that. You can be sure to ask them for a refund. You’re paying for real traffic, not for the one from bots. It won’t solve the problem, but you will at least get some money back.
 
MI
Back