This may be a newbie question, but I am not understanding the source of the following effect.
I have a campaign, and all conversions come from one OS, lets say the traffic and spend is approximately evenly distributed among 3 OSs, the one that converts is one of the major ones. There are also 2 very minor ones that do not convert. Total 5.
Based on what I am spending on each OS's targeting, by blacklisting the two major OS's that are not converting, I should be instantaneously profitable?
What I was expecting was that I would get similar or higher traffic to the converting OS now that I am not spending on the major non-converting OSs. Based on the numbers it should have been instantly profitable.
I leave the two minor OSs because they are not taking up much spend anyway.
Instead of the anticipated effect, my traffic drops to almost zero. After an hour my traffic source had only bid 5 times. I was worried I messed something up, so I unblacklist the non-converting OSs. I instantly get results, it bid several hundred times in a few minutes, and around 20% of these were on my targeted OS. The untargeted campaign bid more on my desired OS than the one targeted to it, in a fraction of the time.
So I guess I do not understand why for a given ecpm bid and daily spend, a targeted campaign will bid and spend less on a certain OS, than it will within a broadly targeted campaign? And why targeting a very broad category like a specific OS is killing my traffic. And of course, how to economically target this OS?
Thanks for your help in advance, any clarification on this effect is much appreciated.
update: after a day running, it has only bid 200 times, 80 wins, and one click.
Do people think this is a unique issue with my traffic source or something I did?
For me it means you hyper-targeted campaign and there isnt much inventory avaiable within your tageting
I would think about why some OS converts and why some not, maybe there's problem with your LP or offer that you can fix and make them profitable too.
But what I do not get is why I get more inventory of the same OS placement when I broadly target, than when I target that OS explicitly. My only targeting is mobile traffic for that specific OS, so I did not think that counted as hyper targeted?
I will look into why others are not converting, thanks!
The way your DSP bids for the inventory based on your targeting varies a lot. On some DSP's tight(er) targeting results in very little traffic despite high bids.
Talk to the DSP or test a different one to see. You now know which OS converts so it makes it much easier to do these tests