11% CTR to Offer & 0 Conversions - Mobile Pop - Any advice appreciated

[Anthony.]

Hi guys, I've been running a bot trap campaign (Image Above) and I can now see the real CTR% from human traffic which goes to my lead-gen offer.
As you can see, there are 21.34% bots. But the thing which bothers me the most is the fact that I have 0 conversions. Even if I got a conversion now it's around 1% cr.
Previous to this bot test, I've tested 5 different offers from that GEO and had ZERO leads.

I know that a certain % of human visitors who see my LP must just close, but that's another focus after I figure out my offers problem.

Can a GEO just not perform? I am targeting Singapore, beginning to wonder..
Any advice would be a great help, this is a great forum and I am the type of person to NEVER EVER give up, it's not an option, so your help / advice will not be wasted.
TIA


NOTE: Ignore the costs, messed up since doing the bot test using 2 offers.

EDIT: After saving this thread, now it's even 400 clicks, this is to one offer, the payout is $1.80 (image below)

 

[Anthony.]

Your traffic source is bad. Ad networks live off bot traffic. This is revealed by many studies for years.
What you are seeing is not good or you are not scrubbing legit bots from your server access logs and analyzing the data.
Use a funnel or redirector -- never send bulk ads direct to a sponsor. At least you will know the quality of the ad buy this way

Yeah I've been bot trapping and only 20% was caught out, so I decided to test direct. But the examples above are not direct..
Traffic source --> Blank page Javascript redirect 200ms > Real LP > offer
I'm allover at the moment, hard to make a plan and stick to it.
Any advice what to do next? I'm not bothered about profit (yet) just improving and learning to progress/improve a campaign. Lately however, they don't look like any improvement can be made!

 

[Anthony.]

function getQueryString() {
  var args = new Object();
  var query = location.search.substring(1);
  var pairs = query.split("&");
  var counter = 0;
  for(var i = 0; i < pairs.length; i++) {
    var pos = pairs[i].indexOf('=');
    if (pos == -1) continue;
    var argname = pairs[i].substring(0,pos);
    var value = pairs[i].substring(pos+1);
    if (!args[argname]) {
      args[argname] = unescape(value);
    }
    else {
      args[argname] += ("&" + argname + "=" + unescape(value));
    }
  }
 
  return args; // Return the Object
}

var oQS = getQueryString();
var mof= oQS['model'];
      var futdate = new Date()        //Get the current time and date
      var expdate = futdate.getTime()  //Get the milliseconds since Jan 1, 1970
      expdate += 3600*24*13*1000  //expires in 1 hour(milliseconds)
      futdate.setTime(expdate)
      var newCookie="model=yes; path=/;domain=.livejasmin.ca;"    //Set the new cookie values up
      newCookie += " expires=" + futdate.toGMTString()
      window.document.cookie=newCookie
if(document.cookie.indexOf("model=") == -1) {
    alert("Cookies  are required to view this offer.");}
else {
    parent.location = "http:/ ... you know the rest of the drill ...

most bots will not take a cookie ... This is old code I would want to write a log too

Ok great, so I use this on a blank page before my real LP and send humans there from the parent location at the bottom?

So if my CTR is 100% to the LP, that means they are all human. Or will some bots still bypass this?

Thanks for the code

 

[Anthony.]

scrapers will croak too, No cookie no laundry (<period>)

Great thanks, so just to make sure:
Traffic source --> blank LP with this redirect code --> real LP --> offer

Also, if Voluum says ex: 1.7k clicks... but my aff network says 800 clicks, was the difference bot clicks?

 

[Anthony.]

They were bots or users that did not accept cookies.
You can deduce that difference only.

If you want the user agent (if any), IP, headers 200 OK then 30x or nanda then you need a PHP (other) redirection script with logging.

Bots, scrapers and *a few real users* will not accept cookies. Users in private mode will accept this cookie only for the browser's private session. It's also possible to track this cookie as a return cookie (of sorts) but it is maybe only 70% or so accurate.

I have the expire set day*13*1000 13,000 days for that reason ... long ago ...

Ok I was meaning before using this script but thanks a lot for the info!
All setup now, looks like its working and due to what you said I can closely guarantee the CTR% on the actual offer is human now (I hope!) Appreciate the help, thanks!

 

[Anthony.]

@Graybeard Am I right in thinking this way?

 

[tyoussef]

Traffic source --> Blank page Javascript redirect 200ms > Real LP > offer

some bots read JavascriptCodes .
make sure you use invisible link methos as well .
and start blacklisting.
in the beginning try to increase the bid so you can get traffic from good placements.
make sure you pay attention to the placements that sends boot traffic .

 

[Emli cristina]

what is this network?

 

[Anthony.]

What is a click bot going to do after reading the .js? Do you have any proof of a click bot's accepting and returning cookies -- empirical evidence? I have a phantomjs bot that scrapes Google search just fine but when it hits a Cloudflare client Cloudflare locks out the SERP content (the SERP's URL). But then I scraped 40,000 Twitter accounts in a few days (random timed [request 16s to 28s apart]) with a common mobile user-agent forged and Twitter never stopped me ...

Cloudflare might be a good filtering alternative.

As far as your question -- net clicks out/gross clicks in= users that returned cookies (in this context it means ALL users).

I spec'd a nosql real time access log analyzer API. Elasticsearch and Kibana on our own servers that ended up being used to appraise the click fraud of the 50K - 70K per/day ad buy and affiliate clicks we received. After analysis of our ad buys we got the (*suspected) fraud down to 38% or so by monitoring the ad networks and using the better ones. We used medians and statistical deviations to identify suspicious *click bot* activities (inductive logic patterns and predictive argument). NicInfo is a smart, command-line RDAP client to help identify your clicks if they are humans or from servers in datacenters. However, any computer on any assumed non-hostile IP (home or business), could be a bot too. It is a cat and mouse game. I assume at least 40% of the ads I am buying are click fraud -- bots.

However, it is a value received equation. Maybe I got less fraud from AdWords or Bing Search ads but they cost a lot more than other placements of campaigns (bulk pops mainly).

Thanks for the reply! I put that js code in a html file, inside script tags, was I supposed to use a .js file?
Traffic source --> website.com/botfilter.js --> real lp --> offer?

What I did was:
Traffic source --> website.com/bt.html --> real lp --> offer

 

[Anthony.]

some bots read JavascriptCodes .
make sure you use invisible link methos as well .
and start blacklisting.
in the beginning try to increase the bid so you can get traffic from good placements.
make sure you pay attention to the placements that sends boot traffic .

Yeah good point, to basically "double check". So ex:
Traffic source --> js filter(Check1) --> Real LP(Check2 - with tiny link) --> Offer

Just started running a fresh campaign.
I have done the tiny image link method.

Top row is bot link, bottom is my offer, seems to me like the bots are clicking my offer instead of the trap. Maybe I might name the image some trigger words or make it bigger and position different etc

 

[Anthony.]

You have to first set a cookie then retrieve it. I used it as a SSI include on the receiving page.
I set the cookie on the home (or entry) page then when the user clicked a link that was the target page, the target page retrieved the cookie and did the JavaScript redirect. I would look into using Cloudflare for your landing page domain. For low volume domains they have a free service: DNS anti ddos (detects most bots) and they offer free HTTPS/TLS service.

Ok I will look into that, I've used cloudflare in the past didn't know about that. thanks

 

[tyoussef]

Traffic source --> js filter(Check1) --> Real LP(Check2 - with tiny link) --> Offer

Perfect That's Good .
Keep Going and let's see what's Going to happen.
increase the bid so you get some good placements.

 

[Anthony.]

Ok so I just realised a BIG mistake I've done.
I had my robots.txt disallow bots.... Whilst trying to identify the bad placements...

So I've been paying for the views, but not able to catch the bot clicks and know bad placements because the bots couldn't access my site.
FFS ah well, learn by mistakes! Good pointer for anyone else doing this.

EDIT: Seem to be getting little to no clicks on the small link only humans would see...

 

[Anthony.]

robots.txt is only read, and it's directives honored, by legitimate search engine indexing robots.
These *bots* are not to be confused with click bots are not human, internet parasites that attempt to eat up your CPC advertising budget, or influence your SEO by creating a high bounce rate or scraping your pages and duplicating them.

Any <a href ="url">...</a> that appears in the source code can be read and ''clicked'' by a bot.

What about Javascript onclick, some bots can't read that right?