Home > Questions and Answers > Tracking Questions

About Bot Test (4)


05-16-2017 04:08 AM #1 sergioska (Member)
About Bot Test

Hi,

I have some doubt about bot test so i think to start this thread to ask some questions.
I'm sorry in advance about the length of this post, but i need to speak to explain my doubts.

I have a mobile campaign pop-traffic based with:

offer: mobidea
tracker: Adsbridge
traffic: popads

Offer

I have chosen an india offer (game cpa) with a low payout 0.74 €.

Tracker

Adsbridge allows to add an internal landing page so i have created two campaign to implement a bot trap following below flow.

landing page campaign -> internal landing page (bot trap) -> campaign with the real offer

To do so, i added in my landing page a javascript redirect after a timeout like explain on step by step tutorial about bot test by caurmen passing parameters to real campaign.

Click image for larger version. 

Name:	Screenshot 2017-05-16 04.52.39.png 
Views:	29 
Size:	39.6 KB 
ID:	15354

I have called first campaign bot landing and configure it using landing page feature like in screenshot below:

Click image for larger version. 

Name:	Screenshot 2017-05-16 05.07.18.png 
Views:	30 
Size:	67.9 KB 
ID:	15355

where landing page obviusly is landing page about bot trap which i have spoken.

So i have configured another campaign only with the real offer.

Click image for larger version. 

Name:	Screenshot 2017-05-16 05.15.55.png 
Views:	27 
Size:	63.0 KB 
ID:	15356

(please note direct link option set)

Traffic

Ok at this point i have started i campaign on popads configuring target, device, carrier, ecc ...
To start i have chosen a low bid offer (about 0.008 cpv) and i get something like this:

Click image for larger version. 

Name:	Screenshot 2017-05-16 05.21.39.png 
Views:	29 
Size:	69.7 KB 
ID:	15357

Anyway, after few hours i stopped traffic campaign to check tracker reports about bot test and this is what i saw:

Click image for larger version. 

Name:	Screenshot 2017-05-16 05.31.22.png 
Views:	22 
Size:	78.7 KB 
ID:	15391Click image for larger version. 

Name:	Screenshot 2017-05-16 05.31.43.png 
Views:	27 
Size:	83.3 KB 
ID:	15392

So i have compared report from bot trap campaign and real campaign to exclude shit placements. To do this i use an online tool to view diff about two file:

Click image for larger version. 

Name:	Screenshot 2017-05-16 05.34.18.png 
Views:	30 
Size:	288.8 KB 
ID:	15358

I have created a list of id to block on popads (row presents bot report side but not in real campaign side).
They should be bot traffic right?

This is my questions:

1) Watching reports i see that there are many ip address mapped on the same id, and i can see same id in both reports (bot landing and real traffic) so i think that if i exclude one of this i should exclude also some ip address good. I'm thinking right or i'm wrong?

2) Maybe i should count how many times an id is present on bot landing page and how on real campaign. So exclude this id If it is more present on bot landing page. Could be right?

3) If i'm thinking right how can i do to remove bot traffic to good traffic?

4) In the screenshot with campaign results i see 2 conversions and both has the same operator (vodafone in this case).

Click image for larger version. 

Name:	Screenshot 2017-05-16 06.00.27.png 
Views:	23 
Size:	44.8 KB 
ID:	15359

Considering that vodafone case is single opt-in while Idea case is double opt-in have sense to optimize exclude this carrier from traffic?

5) Anyone have any suggestion about this case study?


05-17-2017 10:13 AM #2 caurmen (Administrator)

You've put a lot of work into this and given lots of details here: well done.

Your link to the bot traffic report is broken (under "after few hours i stopped traffic campaign to check tracker reports about bot test and this is what i saw"). If I could see that I could answer you more precisely.

It looks like you're overcomplicating this a bit. There's no real need to run a second campaign in parallel to the bot campaign: just look for the low-CTR placements on the bot campaign and exclude them.

The second campaign, unless I'm missing something, doesn't really give you any additional data.

If you're seeing massively repeated IP addresses, yes, they're probably bots. It won't do any harm to exclude them from your campaigns.

So in summary: don't worry about the "real" campaign, just look for low CTR placements IDs on the bot campaign, and exclude those.

Hope that helps! If you think I've missed something about your approach, please do let me know and we can talk through it.


05-17-2017 01:29 PM #3 sergioska (Member)

Hi Caurmen,

Quote Originally Posted by caurmen View Post
You've put a lot of work into this and given lots of details here: well done.
thank you so much.

Your link to the bot traffic report is broken (under "after few hours i stopped traffic campaign to check tracker reports about bot test and this is what i saw"). If I could see that I could answer you more precisely.
i have fixed broken image on origin post.

It looks like you're overcomplicating this a bit. There's no real need to run a second campaign in parallel to the bot campaign: just look for the low-CTR placements on the bot campaign and exclude them.

The second campaign, unless I'm missing something, doesn't really give you any additional data.
probably you are right. I Know that there is no need second campaign but ...
I have complicated my flow with a second campaign, because following your post about bot detection i use link in screenshot below in javascript redirect

Click image for larger version. 

Name:	Screenshot_2017-05-18_06_16_52.jpg 
Views:	18 
Size:	45.5 KB 
ID:	15401

but in this way, when i test campaign i had an infinite redirect to self campaign.
Probably there is something that i don't understand about it.

So what url should use in javascript redirect to avoid this problem?

If you're seeing massively repeated IP addresses, yes, they're probably bots. It won't do any harm to exclude them from your campaigns.
i get some doubt about this point. Certainly repeated IP address are bot (specially if they don't generate payout) but how should i do to remove it? Popads doesn't allow to block by ip, but it allow block placements by websiteid that (please you correct me if i'm wrong) contains a group of ip.
For example in following screenshot i see 3 different ip mapped on the same websiteid

Click image for larger version. 

Name:	different_ip_same_websiteid.png 
Views:	20 
Size:	126.6 KB 
ID:	15396

I'm afraid of excluding it because in this way i could lost also good traffic.
So how can i do to optimize good traffic?

At this moment i'm following @vortex suggestion about how to optimize a campaign on this post and i see some feedback, but it's not enough.
From the following list i have blocked the first item because it has a spend at least 2 times higher than campaign payout. Right?

Click image for larger version. 

Name:	remove_bad_traffic.png 
Views:	18 
Size:	88.5 KB 
ID:	15397

but i would like to have a better way to optimize my campaigns.
I get some click, but i spend 10 to have 1. There is something that i don't know or that i don't understand

Could you give me any other suggestion about optimization?


05-18-2017 10:52 AM #4 caurmen (Administrator)

IPs:

If you can't block IPs at your traffic source, then don't worry about it. Just block the IDs that show up on the bot tracker.

Javascript redirect problems:

The infinite Javascript redirect - that's a problem! Unfortunately I don't know Adsbridge well enough to solve that, and it looks like an Adsbridge-specific problem.

I'd recommend getting in touch with Adsbridge support to ask them about that. Or perhaps one of the Adsbridge folks will chime in here?

Optimisation:

You're doing the right thing!

Don't worry about the fact that you're seeing a lot of unsuccessful placements: that's normal. The successful ones pay for all the unsuccessful ones.


Home > Questions and Answers > Tracking Questions