Our way to the very first profitable campaign (7)
Some nice green already! Congrats on that!
A couple of points that need to be made:
1)You can't conclude that LP4 is the best lander - there are simply too few conversions. Here's the method you should be using when comparing landers:
http://stmforum.com/forum/showthread...Banners-Part-1
Also - why did the landers receive different amounts of traffic? You should be rotating them equally in the same
2)How many landers have you tested in total? Just those 4, or have you tested more before? And where did the landers come from?
Do you have a subscription to Adplexity? If not you can get a discount from here:
https://stmforum.com/forum/showthrea...e-STM-Discount
Go rip and fix up 10+ of the most aggressive landers you can find - set filter to "received most traffic" and date range to something recent like 30 days, and download ones that look completely different. You can set country filters too - just select all the countries that speak the language you need.
Start with those instead of reinventing the wheel. You'll then have a solid idea on what the CTR range should be, and which type of lander converts the best. After that, you can either just run the winning lander, or rip/create variations on the winning lander and test some more. (And remember to use the cutting method I've linked to above to cut landers down to a winner.)
It's usually a much better idea to start with ripped landers to establish a benchmark, and THEN optionally create your own to try to beat your benchmark. There isn't as much value in creating truly original landers these days - with all the spy tools around that everyone has access to, your custom lander will get ripped and used in a matter of hours. So unless you have a way of keeping your landers out of the spy tools, I would suggest to stick with the approach above, i.e. test very different-looking ripped landers first, then testing variations of the winner.
Testing variations is a sneakier way of staying under the radar - there may be hundreds of landers on Adplexity that look very similar, but people won't know which one performs the best until they've tested them. Whereas if you come up with your own lander that looks completely original, it would really stick out and would get ripped right away.
3)Regarding CTR - as you've seen, a high CTR will not necessarily translate into a higher CR, and it's the latter that's important and not the former.
However - a low CTR can be an indication of issues that can be prevented/fixed, such as lander loading speed. Are you using a CDN? If not that should be arranged. And have you checked page load times? Ideally the pages should load under 1s, and the faster the better if you're running mobile traffic.
4)You'll find a bunch of relevant and helpful posts on landers in the "On Landing Pages" section of this FAQ post:
https://stmforum.com/forum/showthrea...-AM-FAQ-Thread!
Have fun testing! And do let us know your results! 
Amy
Originally Posted by vortex
Originally Posted by vortex
Originally Posted by vortex
), 4 to 5 more landers (9/10), of course same GEO: Italy. I have a question for you about bidding: we now have bid quite low 1,3 to 1,8 $ / 1.000 impressions. Is it too low? Is this the reason way we did not collect enough traffic? Originally Posted by vortex
We are using a CDN and the loading speed for our landers is optimized: according to the kind it varies between 1.200 and 600 ms.
Originally Posted by vortex
Originally Posted by vortex
Originally Posted by ciscotd
), 4 to 5 more landers (9/10), of course same GEO: Italy. I have a question for you about bidding: we now have bid quite low 1,3 to 1,8 $ / 1.000 impressions. Is it too low? Is this the reason way we did not collect enough traffic?
The easiest way to run aggressive without cloaking, is to 1)only run offers that don't require lander approval, and 2)only run on traffic sources that allow aggressive landers.
Of course there's value in your approach as well, i.e. running on a traffic source that is strict about landers, by tweaking the wording until the landers are accepted. I'd be really curious to see if your current landers are good enough to make offers in other geos profitable when you translate them.
Regarding the bid: Every traffic source has different bids for different targeting at different times so I can't say for sure, but 1.3-1.8 cpm sounds OK. You'll need to test bids anyways to see which one will give you the most profits. You can either clone your camp a few times and set them to different bids, and run them at the same time to see which one is more profitable. Or change the bid every day, and run it for a day at each bid, and observe the results.
It would be good to wait until you've finished split-testing landers and offers, before testing bids. Trying to split-test landers/offers while testing bids at the same time will make things very messy and skew your data - too many moving parts can't be good.
For example, I could stick a piece of code on your best lander right now to automatically send the visitor to the offer after 2 seconds. The CTR would skyrocket, but the CR could be much lower because the visitor isn't being pre-sold properly - unless the offer page alone is enough to do all the selling. But if the offer page alone is working better than pre-selling with your lander, than you should be direct-linking in the first place.
The important thing to remember is to (almost) always base lander-cutting decisions on CR and not CTR. Exception would be when the CTR is so low that the CR would have to be impossibly high just to breakeven - in that case you can cut it based on CTR alone.
Amy
Yes your bid is "low" for carrier traffic in Italy. But it could be as well that your traffic source just doesn't have a big inventory for carrier IT. If increasing bid does not help, you can check some other traffic sources. Exoclick you can spend easily a couple hundred $/day for carrier IT.
Originally Posted by sebastian_r
Ok, yesterday night we were able to rip&fix other 6/7 LPs. We set up and started a camp with 11 different landings. All landing have a loading time < 1 sec some of them 0.6, some 0.9. Let's say in average between 0.8 and 0.7. That's should be fine, considering that last month I did not know what <div> meant, right?
I think it'll take a couple of days ( or maybe a week 
) to gather enough data, even if we increased a bit the bid and traffic is running slightly faster than before.
I am looking forward to update you and million thanks for the advice!!
Originally Posted by ciscotd
I think it'll take a couple of days ( or maybe a week
) to gather enough data, even if we increased a bit the bid and traffic is running slightly faster than before. I am looking forward to update you and million thanks for the advice!!
A week to gather data? Which source are you running at?
Amy
Home > Mobile > Follow-along Campaigns