Small GEO Campaigns With Lots Of Placements, How Would You Set It Up? (9)
Small GEO Campaigns With Lots Of Placements, How Would You Set It Up?
So I've been running a bunch of small(ish) GEO campaigns on TJ and Exo for a little over a month now and started wondering if my setup might be sub-optimal...
My CPV Lab tracking is set up as follows:
- 1 campaign per GEO and Traffic Source
- placement is passed via keyword (e.g. PORNHUB_VA)
- banner ID is passed via additional token
- cost is the average TJ/Exo CPC of all placements together
I am running GEOs on up to 9 placements respectively with volume currently ranging from $5.00 - $ 70.00 per day.
While this setup keeps everything grouped together pretty nicely, it comes with a couple of limitations:
- CPVLab cost stats for individual banners are pretty much useless as the actual CPC varies significantly between different placements, so they have to be cross checked with TJ/Exo stats individually (huge PITA).
- decisions on LP / offer performance are made based on overall campaign data, not specific placements.
Which brings me to my questions:
- How do you guys set up campaigns like this? Would it be better to split it up into individual campaigns for each placement?
- If yes, would you do so right from the beginning of the campaign or rather decide on a top performing offer / LP first with all placements together and split up afterwards? (could take a while to get statistically significant data on low volume placements)
Any additional suggestions on how to improve this setup are highly appreciated, I feel like I spend way too much time each day filtering data manually...
If you don't want to create every permutation (which results in a ton of campaigns), you should stick with the current setup.
What you need for a temporary fix is an Excel doc that you can paste in exported .csv's that generates an overall banner by banner report on their individual performance.
You need the same Excel doc for sites (which translate to their exact placements since you're only targeting 1 ad format in a campaign).
So, in the end you're able to drill down to the performance metrics of each site and each banner. Cut the crappiest first, obviously, let it run, and do another round of cutting, and so forth. Don't cut all poor performing ones right away because, like you said, you are not creating every permutation, so it could have been some really bad sites that drag down the performance of otherwise good ads.
If you're not good with Excel, we've made a bunch of tutorials on it:
http://ipyxel.com/category/tutorials/excel-tutorials/
I can also take a look at your data, if you're up for that, and help you create one. Just contact me directly.
Last tip: sign up for www.AdultAdsPro.com, which will solve all your problems when it arrives 
Thanks Tom, I will definitely check out your Excel tutorials, it is something that I meant to improve for a long time now.
Already signed up for AdultAdsPro, really looking forward to it! 
FYI: Tom Fang is awesome!
He just helped me out big time by creating a kickass Excel spreadsheet which makes running my daily reports way easier.
Thanks again man, this will save me a bunch of time in the future!
That sounds awesome, I've been trying to work out a spreadsheet myself lately. Tom, does the offer go for anyone? Is this something that needs to be custom done or could you guys post one up to use?
Here you go: https://www.dropbox.com/s/ijbczdzvza...-Template.xlsx
You probably have to do a bit of customization depending on your campaign setup and I can't help you with that cause I suck at Excel
What you need to do:
- In the CPVLab tab, you paste the target performance CSV from CPVLab
- In the TJ Tab, you paste the campaign performance CSV from TJ
- In the UID Tab, you assign the TJ campaign ID to the corresponding target (in my case, 1 target = 1 TJ campaign)
If you understand Excel, you can tweak what @andyvon put up, but there are some custom things specific to him in there, so it likely won't work out of the box.
If you really need something like this, feel free to add me on Skype. However, I'd be taking my own time to do this, so please, no newbies! I'll try to help some of you if you really need this.
Have your reports from CPVLab/Prosper and TJ/Exo (a campaign performance report) ready.
And you better be opted into www.AdultAdsPro.com already!
Thanks andyvon, appreciate the share. That is a beast of a file, kept choking my excel, could be my ancient macbook too.
Thanks Tom, I think I might take you up on your offer, very generous of you. I had already opted into Pro. I'll hit you up on skype to see when you have a few minutes free.
As my EPCs on those campaigns pretty much went to shit over the last week or so due to various reasons (drastically increased bids, "bad run" with new banners etc.), I decided to take a step back and reevaluate my whole process which still involves way more throwing shit against the wall than I would like to admit
To start with, I think I will split the campaigns up into separate ones for each placement in order to have more control over individual performance.
There are a couple new offers in my GEOs that others seem to do well with and different LP styles I'd like to test, so I will more or less start from the beginning.
I will follow Maynzie's suggestion and go Offer=>LP=>Banners, which brings me to my questions:
- Should I decide on top offers for each placement on it's own or overall GEO performance?
- Same thing for LP, each placement on it's own or overall GEO stats?
I know that in a vacuum, deciding for each placement on it's own is a no-brainer in both cases, but some placements don't receive that much volume and it could take quite a while to get a statistically significant sample.
Home > Paid Traffic Sources > Adult Traffic (NSFW)