Testing order? (3)
Lets say my testing order is as follow:
1) Traffic sources
2) Different offers
3) The angle
4) The type of landing page
5) The components of the landing page
6) The banners / ads
When should i tests the following?
- connection types
- device types
- Geo
between offers / angles, angle / type landing pages or somewhere else?
I like to hear your experiences with this.
I would say you mixed it up a bit, some of the things you lined up are not really varaibles that you need to test during the optimization process.
For example GEO is pretty much the first thing you need to decide on, as all your creatives, landing pages and even offer selection will depend on. So its no the same type of variable for testing as banner or a landing page.
Lets split this into 2 parts, first one is planning, then its testing/optimizing.
Part 1 : decide what kinda vertical you want to run in and pick the GEO to start with and later on, you can scale the same vertical into other GEOs. Or do it the other way around, maybe you live in a large country and want to test your native market first, so pick the GEO first and then decide what kinda vertical you want to test in it. Once you get some traction with the GEO, you can test more verticals/offers in it.
So now you have the first GEO and the first offer to test, do some spying and find out how people promote it. Grab some LPs, tweak them a bit so you dont start as a lazy copycat 
and also prepare some banners if its needed. During the spying process, you will also get an idea about the traffic networks that you could try.
This is the end of part 1 - planning. Now you have an offer, creatives, LPs and a few traffic networks to test.
Part 2 : Start to run in 1 traffic source, with low amount of banners and LPs. Buy some traffic and wait for the data. Now you need to analyze it and your next step will depend on what the data tells you. There is number of situations that you can run into. First you need to figure out what is your biggest problem at this stage.
- getting no traffic at all? This is a problem related to the traffic source so figure out what the problem is - low bid? To much targeting? Wrong keyword selection? No traffic for the GEO? ... Fix what is needed.
- getting traffic that is too expensive? - Is it low banner CTR? Bid too high? Maybe its just some sources that are driving the cost up? ...
- not enough clicks through the landing page? - Is it working at all? Is the copy translated properly? Is it aggresive enough ...
...
I can go on and on but to simplify what Im trying to say : there is always some problem that is basically the bottleneck and unless you get rid of this problem, the campaign wont get nowhere. You need to find this problem first, it would be pointless to follow any set order of optimization steps. There is no point in cutting device types for example when you dont have the LP optimized to display correctly on all of them. Or cutting banners when the offer isnt converting, thats pointless too.
Sooner or later, you will develop your own testing routine and it might be totally different from what Im telling you now, but this is how I do it - Im always looking for the weakest part, and correct that one first Once you get familiar with your market, it will all be easier as you will know how to make proper funnels fast. You will know what kinda performance to expect so you will be able to optimize everything faster as you will limit the number of week points right at the begining of launching a new campaign.
Test creatives first. Maybe 5 angles ass using the same 5 banner templates so in turn youll have 25 banners. Or 5 banners per angle.
Then 2 landers
Once you spent 100$ look at data. First pick best banner that bought most conversions.. then let it run another day.. pick best lp... then look at placements.. block ones that spent offer payout x2 and 0 conversions...
Home > Paid Traffic Sources > Mobile