Pooling method running mobile pop questions (4)
Hey all,
I've been trying to setup a proper optimization method as outlined by Ruby in his POP tutorial and Bigeasy123's pooling method.
I'm running POP's currently on PopAds primarily and
Campaign is an app-install.
This is my understanding:
- Start with 1 campaign, and setup a RON campaign with pop source. Bid kind of low, but enough to get a good amount of impressions. (Keep in mind the payout and CPM relationship)
- Once you’ve got some data, around 8-12x payout x landing pages, make 3 separate campaigns
- 1 campaign is a "profit/whitelist” campaign (20% higher bid). 1 "test pool" (low-ish bid as specified above), and 1 "lower bid" whitelist campaign.
- So now I can take the profitable landers/placements and place them in the “profit/whitelist” campaign and “lowerbid” campaign and let it run.
-Take the converting landers/placements off "test pool" campaign so it’s a “negative” campaign to test remaining placements.
-Cut placements that don’t go positive after 2-3x payout on the test pool campaign.
-Cut landers that don’t go positive after 8-12x payout.
Question 1a: Do the "placements" I'm testing include OS, ISP, Category, Device Brand/Model, Browser, etc? (I just want to be sure I have an actual understanding of what placements are in relation to pop traffic.)
Question 1b: Do I cut landers from the test pool campaign also?
If I remove all profitable placements then how am I supposed to test the remainders?
One removal of the highest traffic ISP completely changes the entire way a campaign works. This would make the test pool worthless at that point once multiple +ROI placements(OS, Browser, ISP, etc) are removed.
On top of that, cutting one profitable lander out of.. say 6 will also completely change the test campaign.
Question 2: When replacing and trying new landers, do I just add them straight away to the whitelist campaign or re-created a RON/pooling type campaign with all placements again?
Question 3: For the whitelist campaign, if placements go negative, do I put that one back into the test pool campaign or just trash it all together?
You don't understand what a placement is... in the case of Pops it's a website, in the case of MDSP it can be a Site or an App (so 1a: Noooo it doesn't include OS, ISP, etc !!!)
Placement = where your "creatives" (pops, banners, videos, text ads, etc) are displayed, so = Apps, Websites
Placements are listed in Campaigns > Quick Menu > Quick report by Ws. in popads
Don't know why you want to take off the good landers from test pool... of course if you only leave the shit placements + the shit landers in the test pool you'll get shit results... I would run good/unknown landers on the "test/unknown" placements.
1b Well if you're sure you have a crappy lander on the best placements, why not kill it everywhere ?
Question 2: I would add it to the whitelist as I know it should convert if it's a good lander (1 variable less to test, you know the placements are good in this campaign)
Question 3: I guess if it turns back red after being in the whitelist for a while it probably means it won't come back in the green, just leave it in the red long enough to be sure it's not just because it's thanksgiving (or whatever environmental variable could make your placement turn red a few days) and ditch it if it doesn't come back (or find a new offer to try on it... people might have gotten sick of your offer)
Cheers,
The problem with that strategy is the test pool. Everybody who runs pop on some volume will probably confirm that low bid traffic is completely different to high bid traffic. By only bidding on the low end for the traffic, you'll actually miss out a lot of targets depending on which source you're running and hardly get any 1st impressions in the queue.
Personally, if I see some potential bidding somewhere in the middle (I tend to go in above average bids just to see if there's some movement), I always give it a quick shot by bidding high to see how this effects the placements the traffic is coming from and my overall ROI.
No golden formula to make this work, I guess everybody has a bit of his own style when it comes to testing. One thing you might try is splitting up in wifi/3g traffic (assuming a source allows you to do so) as 3g is significantly more expensive than wifi traffic, especially in the tier 1 countries but still has quite some volume.
@wiifmdude, thank you for the explanation! That makes so much more sense now. What about categories? Do you cant those as placements?
With Popads, I'm seeing the ability to target via website id and also categories.
As far as targets (OS, Browser, Device, etc), do you have a process of how you test them? Or do you just focus on removing bad placements (website/app) and landers?
That is probably where I am going completely wrong. I tried only whitelisting landers/OS/Browser/Device, etc.
@fjk87, Thanks for the input! Very helpful advice 
Home > Paid Traffic Sources > Mobile