wow that's a lot of variables you're testing.
and unfortunately none of it is statistically significant. though it does appear that none of your creatives or LPs are particularly convincing. I would spy on pop sources for LPs. You'll see lots of antivirus. or maybe you won't. and that will tell you something too.
do you have some reason to believe that this offer is a performer?
creativity in the banners and landers certainly has it's place. but....... if your testing budget (and time) is limited you may want to focus your efforts elsewhere.
Those are some really low CTRs for AV.
Think about your average smartphone user.
Why would I care if an antivirus (which I don't have) detected a Google Play clone (which I probably don't have)?
Because it stole 1000 bank account details? "It won't happen to me". "I'm not that stupid"
I should be able to glance at your banner out of the corner of my eye and know what it's about, know why it's important to me, and know why I need it, RIGHT NOW!
Let's face it. You're interrupting my day. Why do I care about what you have to say? If I have to think about it to make sense of it, or double-take AT ALL, I'm probably going to ignore it. Gone. Merrily continuing along with my day. Wasted impression.
Think about a problem I might already have that you can solve. Don't try and pull something out of thin air.
Then bitch-slap me with it! Show me what you've got! Do it in 8 words or less! Do it in 4 words or less! Make sure it's stabbing me in the eye with a red-hot poker!
"I'm sick of my sack of shit $50 smartphone, a pregnant turtle could start a brothel faster than my Samsung can open Facebook!"
An antivirus helps de-clutter things, right? If you can rid my phone of the viruses from the filthy websites I visit in my spare time, maybe I can get Instagram open and stalk my old high school crush a little bit faster. I'm a happy chappy.
Do some spying. See what else is out there. If you see something around a lot, it's making someone money... so pay attention.
(also, perhaps consider another offer)

I usually run 3 angles x 3 different designs. I can pretty consistently get a display campaign up in under two hours. I have landers I know work, and banners I know have good CTR and are a good starting place. As hlyghst said, it's important to not test too many variables at the same time while starting out/low budget/no team to do the grunt work... it takes a little time to get your shit together when you have no baseline. So spying should be massive on your agenda to compensate, steal stuff shamelessly then improve upon it until you have your own little toolbox.
100 banners for an unproven offer, damn! I admire that determination. Keep in mind the 80/20 rule. 80% of the profit is gonna come from about 20% of the campaigns. Spy. Take inspiration. In order to innovate you must first imitate and assimilate.
If I'm giving everything a good go, (and I've determined through my own experience that the above method works for me), then I want to be churning as fast as I can, in a controlled fashion (usually one variable at a time), through geos, offers and ad copy until I find something that shows good promise straight off the bat.
Keep going, I'll be following along - was in your shoes not too long ago!
1. I don't really get your landing page. A good practice is to try to make the landing page repeat elements from your banner, right now the lander looks like it has come from another universe.
2. I can't really understand which individual banners and campaigns your stats refer to, but if I was just a betting man, I would probably say that banner 3 is likely to be more effective than the other 2
3. So if I was to just give you some thoughts (and I don't do anything in this space, so this is not on the ground tested advice at all) I would probably suggest testing slightly different versions of banner 3 with landers that reinforce the exact same message in that banner.
All these numbers are way to small, not representative enough to make any decisive conclusions. You need at least 100 clicks on every offer to say something about it.
Keep on testing. But if you wanna master affiliate marketing its gooing to cost your some $$$ learning money
Why weren't you direct linking at all? I'm pretty sure it was recommended to split test DL and LPs and different kinds of LPs.
-Your angles are too long.
-Your angles like like meh, no one gives a damn which the best a/v is and thats its not a scam.
-You need to scare in order to cause action or you need to appeal to other emotions/traits (finch wrote a kick ass guide about this)