Split Test Reminder: same offers vs. diff networks 35% -> 94% ROI by changing links! (18)
Split Test Reminder: same offers vs. diff networks 35% -> 94% ROI by changing links!
How the obvious tests can still surprise...
Been setting up a new campaign (mobile - 1 offer, 1 GEO, 1 traffic source) for over the past weeks: testing angles, ads, lp's, tweaking the backend funnel etc.. I got it up to around 35% green ROI after three weeks and was kinda out of new ideas to optimise further.
35% is something to keep going, sure, especially because this runs at good volume. However, last Monday or so I thought to finally split test networks. Details of this 'test':
- Both networks are 1st tier and have great footprint in the mobile space, having direct deals with merchants on this particular offer
- Same payout, highest in the market atm
- Hasoffers vs. Cake
The results shocked me as the ROI went up to 94% overnight, performing steady for over the past 6 days.
No idea at this point yet what is causing this huge difference, but obviously this gives me a lot more space on the frontend to raising my bids and killing some more competition. 
I know it's one of the basics and all and def something most of you guys do all the time, but don't get too comfortable with just 1 or 2 of your favourite networks. Remain really critical when it comes down to this stuff.
Wow that's significant, especially for such a simple change!
Possible causes:
- different clicks loss / technical differences
- Affiliate network scrubbing
- Advertiser scrubbing
from http://charlesngo.com/splittestnetworks/
How many networks are you in? I'm in 5 right now an none of the offer I've been running was in multiple networks for the same GEO.
I'm curious, is the HasOffers or the Cake one performing better?
Originally Posted by kepe95
Yes, my best bet here is network scrubbing. Although this particular (in my case, losing) network is one of the bigger ones out there, it has been acting (very) 'weird' lately in how it treats its pubs etc. So another reason for me to stay away from them as much as possible...
I have accounts on like 30 networks or so, but I'd say 95% of all my revenue comes from 3 main ones atm.
Originally Posted by andyvon
Scrubbing has a big impact on ROI.
Wow, that's nice increase. I have to try this for my current campaign ASAP as I'm getting suspicious
Originally Posted by erikgyepes
You probably know this already, but use offervault.com and affplus.com to your advantage in order to find other networks that have your offer listed.
First and foremost, I'll keep repeating this to death, always split test offers, even the same offer on different networks.
However From my experience most of the times it's not even the aff network scrubbing.
Nowdays the affiliate network scene is so competitive that very few networks would be so stupid as to scrub.
From what I've noticed it's mostly the advertiser scrubbing network wide due to network wide quality.
Also tracking platform and geo of tracking servers has a lot to do.
Originally Posted by bbrock32
with dating offers, lately ive been noticing a trend with offers dropping in CR a few days after launch.
every time I will split test it with an alternate aff ID on the same network after the CR drops. new aff ID always seems to perform excellent on launch once again, then drop as usual after a few days
every time I will split test it with an alternate aff ID on the same network after the CR drops. new aff ID always seems to perform excellent on launch once again, then drop as usual after a few days
^ I'm convinced that initial burst is the advertiser dressing to impress -- to get affiliates through the door.
Seen it so often.
One of the more frustrating parts of the adult sector.
Originally Posted by Finch
What I love the most is the always repeating "coincidence" of getting a lead within the first few clicks sent to the offer
seems to depend on the advertiser. with some adv, you can expect it with pretty much every offer they have.
I tried to see if they are skimming based on CR alone, so split test good traffic across 2 aff ID's with the 2nd diluted with poor converting traffic, to lower the CR
both aff ID's had their CR drop with the good traffic at nearly the same time
Thanks for shouting out my blog post.
As far what accounts for the difference, I've listed my theories.
These days I literally don't guess and don't care. I split test and let the numbers speak for themselves.
If a network keeps failing then they go on my shitlist.
On dating it's 99% the advertiser.
If you want to scale the best bet is go direct and pass as much info as you can to make it work for both.
Everything is a shortcut which will eventually fail.
Thanks for chiming in guys. I can see how CR can differ a lot, especially between dating offers. For the record, the offer I mentioned here was a carrier billing (MSISDN) type for a gaming content subscription.
Originally Posted by Finch
Home > General > Affiliate Marketing Forum