How to measure site traffic quality for legit & bot-traffic with site data?

[Navid]

Hey guys.

I'm new to affiliate marketing & Media buying and I'm doing some website research to find some good websites to run some specific offers on.

I'm using Similarweb for website ranking and get a "taste" of how/where they get their traffic from. The site I'm sharing the data with you is an adult site, so traffic would mainly be Direct, Referral, Search and small % social (or none).

So basically reason I do it is that previously I burned my budget on some sites that almost everything was bot-traffic and I want to determine if the site gets its traffic mainly from bot-traffic or the traffic is mainly legit.


These are some images from the data I got:

Based on these data, is it anyway it can determine if the majority of their traffic is bot-traffic or legit?

Appreciate any help!

 

[tyoussef]

i think the best thing is RON traffic, not to put your budget in one website and praying for the traffic to be good .
am i right ?.

 

[Navid]

i think the best thing is RON traffic, not to put your budget in one website and praying for the traffic to be good .
am i right ?.

Been praying a lot and what I received was clearly bot-traffic lol

 

[Navid]

I would do a forensic analysis of the advertising target's webserver access.log.

Resolve a random and representative sample of the IPs delivered by the advertising.

Supposition is a waste of time and inconclusive.

Low bounce rate and 3+ pages per user is not much bot traffic (as a guess).

However, if you buy advertising based on that site's profile -- why would it matter? You could just bot those ads and never enter that website. Without server logs you would never know ... 2¢

Thanks a alot man

 

[JamesLeeOnline]

Been praying a lot and what I received was clearly bot-traffic lol

hhaha.. so sad to hear for that mate

 

[towhidzaman]

Pretty unfortunate. Have you tried checking the keywords they ranked for? Since most of the traffic coming from search, there has to be keywords for it.

 

[Graybeard]

I would do a forensic analysis of the advertising target's webserver access.log.

Resolve a random and representative sample of the IPs delivered by the advertising.

Supposition is a waste of time and inconclusive.

Low bounce rate and 3+ pages per user is not much bot traffic (as a guess).

However, if you buy advertising based on that site's profile -- why would it matter? You could just bot those ads and never enter that website. Without server logs you would never know ... 2 cents

Graybeard, Dec 11, 2017

Navid likes this

 

[Graybeard]

Almost all HTTP/1.0 traffic is not human ...

Graybeard, Jan 4, 2018

Navid likes this