Cutting Banner or Pub First? (9)
Hi guys, I understand there are some guides to cutting Pubs, but my question would be, which would you prioritize first?
In the past, my focus is Pub only as I ran CPC campaigns (completely avoided CPM due to inexperience, but only realised how much I missed out recently. Also, couldn't get any traffic with CPM on adwords). I typically cut Pubs at 2-3X spend with 0 conversions. If there are conversions but making a loss, I will dig deeper to see if there are any converting handsets/manufacturers/carriers with this siteid and continue a further test before eliminating this site/app.
Now with CPM campaigns, in order to keep CPC low, I remove low performing CTR banners, around the 3k impressions mark as per the guide. However there is the hypothetical question: "what if this low CTR banner is performing very well in 1 or 2 publishers", but horrible at the rest.
So, banners first, or pubs first?
I'd recommend continuing the approach you're taking currently. Cut low-CTR banners but only if they're decidedly lower than the rest of the banners you're running - you really want to be cutting banners primarily on ROI. As for pubs, give them a bit more data - maybe 4x spend - before cutting on no conversions, because the increased number of variables on mobile means you need higher statistical significance.
There's always a danger that there's one tiny sub-combination - one banner-site-handset combo, or something similar - that works really well, but generally you're better to aim for things that work well overall rather than try to find those tiny pockets.
^^ yep. Forsure banners first, get rid of the bad ones.
Pubs are easier to knock off but you want to collect the data before you do.
To sum it up; you'll have banner data a lot faster then you'll have pub data.
Hmm okay, what about low CTR banners that convert? Keep those, or work soley based on CTR? I'm not sure if it is worth the trouble to look at banner ID - adds another layer of variability especially when during testing phase when volume is low and statistical significance is not established.
would we correlate banner to CR? I mean, a high CTR with 0 CR is quite useless right? Like the example on the CTR thread. A banner that says "SEX"! will probably have very high CTR but a total disconnect to the offer/lander.
The most straightforward way I suppose is to eliminate banners right at the traffic source without consideration for conversion? I haven't really used postback urls to traffic sources (for privacy purposes).
Why would you not use postback tracking? Tracking is a cornerstone of IM and you should be using it. I'm not sure what your privacy issue is... If it's some domain being revealed to the traffic source you could simply bounce a postback through another server/domain.
You should cut based on CTR but don't cut an ad that is currently profitable.
Profit matters more than CTR and always will.
If you are not tracking profit at the traffic source when available you are increasing your own workload nonsensically, basically wasting your own valuable time.
Yep, if you have a banner that has converted, only cut it on ROI - don't cut on CTR.
Originally Posted by zeno
Originally Posted by caurmen
Is this the norm? Most guys at RM don't do it, but indeed it is a lot of work looking at
That said, there's the benefit of ease of convenience. Well..
RM? Revived Media?
Tracking systems shouldn't be a lot of work.
They should reduce workload by making your data easier to digest and thus decision making more straightforward. They will certainly make this easier than your affiliate network platform will.
There is a reason most people use tracking systems.
As for posting back conversion data = traffic source has insight into campaigns...
Don't be so paranoid. Firstly, the traffic source can probably readily see what handsets, OS, placements etc. you are running volume to so you're not hiding shit by omitting conversion tracking.
Internal staffers running campaigns often happens. Will your stuff get jacked? Rarely.
I won't lie, this has been known to happen, but not so much with reliable, long established platforms.
Fact: If this happens and you have evidence, you can do irreparable damage to the reputation of the traffic source. Do you think a traffic source that has people spending a cumulative $xxx,xxx+ daily on it is going to risk that over your $xx-xxx per day campaign? I doubt it.
There will always be that one idiot who tries to exploit their position of power but it is no reason to handicap all your campaigns.
Furthermore, I would worry about the affiliate network more than the traffic source... Rogue AMs who can see the $$ are more likely to steal than traffic source employees.
Thanks Zeno,
I'm enlightened. I spoke to a friend about this thread and he said the same thing. It's true, the speed advantage > privacy issues, at least on the traffic source side. time to end this self-misery of going back n forth
Home > Paid Traffic Sources > Mobile