Hello Guys,
I'm working with native ads for few months and now looking for effective way to filter out bot traffic fast .
Could you tell me please what methods are working the best for you? Which services you're using?
Thanks!
Lots of ways including testing for scroll depth, loading javascript, and load images...though these all require you put your own HTML/javascript on the page. Other ideas could include testing for cookies, looking at user agents, or keeping track of bad IPs/ISPs. (And sometimes you can tell your network to block IPs from getting further traffic.)
Some people use their cloakers as a way to filter bot traffic too. And I've heard good things about http://trafficarmor.com.
I think what most people do is just automatically block placements when they have either too high of a CTR or too low of a CTR after it reaches a certain number of visits. (The theory being that bots either never click, or always click.) It's more crude, but it works.
BTW I'm working on a tool that auto-blocks native widgets/placements when they reach thresholds of CTRs - monitors via the APIs 24 hours a day. (Among other optimizations like auto-bidding, blacklists, etc.) Hit me up on Skype marcel.katz@driveads.co we are running a private beta and looking for feedback.
Hey,
Thx for detailed reply!
Jus have added you on skype.
Here's the approach I generally use on any traffic source - more or less what happycoder recommends as one approach. It's not super-detailed and there are definitely more layers you can add, but it should give you reasonably good initial results.
https://stmforum.com/forum/showthrea...n-Any-Campaign
Definitely worth also looking at ISPs ("Gosh, those guys from Digital Ocean really like my ads") and useragents.
Most native networks are not blind, means you get referrer data. Best way to get rid of bad placements is checking the URLs manually or hiring somebody to do it. The competitive environment on the page will tell you all you need to know. For example all garbage abitrage -> kill it.
A nice idea would be to have some sort of STM-wide blacklist of bad widgets/sources from various networks.
A list that we would built by us the users collaboratively and that would be open to everyone.