Landing page while testing (5)
I've heard people say that you want to have your landing page optimized before making a decision on whether to cut ads/audiences. Does anyone do this? It seems inefficient to me since you'll lose money while waiting to cut ads and audiences and the ads still get a fair chance against each other since they are all going against the same landing page variants. What do you think?
The more parts of your funnel are optimized when testing some element, the better. Even with optimized landers, you would still test multiple ads at the same time, so they would still be able to prove themselves. And with a good LP, you can kind of "trust" the data more. To put it simple, some Ads might show poor performance with a specific LP but a good one might work much better.
The problem is, when starting something fresh nothing is optimized and you cannot optimize all at once. So proceed step by step, optimize one part first, then move to the next one.
Originally Posted by loicstdenis
If you're wanting to split-test accurately, according to statistics, then I would suggest this approach:
-Throw everything into a campaign - let it run until you see either 1)a promising lander emerge, and/or 2)a promising ad emerge. "Promising" can be subjective though - basically you want to find something that LOOKS LIKE it does a decent job in converting.
-Next, use your promising ad to test all your lander variations to find the best lander (by cutting with the help of a statistical calculator), and then use the best lander to re-test all the ads again (because there may be multiple winning ads - you just haven't run enough traffic to them yet).
If you're going to do this though, make sure all the ads and landers have the same angle. This is because my approach above does NOT test ALL ad+lander combinations. The approach above assumes that the "winning lander" will convert well with every ad - which isn't necessarily an accurate assumption, especially if you're testing landers and ads that have different angles.
Your suggestion of cutting ads while testing multiple landing pages COULD work - just be aware that it may not be an accurate way of testing/cutting. The devil is in the details. It would be easier if you could show a specific example.
Amy
Originally Posted by vortex
- - - Updated - - -
Originally Posted by matuloo
thank you!
Originally Posted by loicstdenis
Some ads have high CTR / low CPC, but completely flops in conversion rate.
Then there are ads that have lower CTR / higher CPC, that have a good conversion rate that more than compensates.
It depends on how good a job the ad is doing at preselling the offer, and how smooth the transition is from ad to offer to checkout.
My favorite example: I once tested some viral gifs from imgur.com that were eye-catching and attention-grabbing, that were only marginally related to the offer. Some of them were a big hit with FB due to the high CTRs / low CPCs - but I got zero conversions. Although there was preselling in the ad text, I suspect many people just clicked through expecting to see the "whole video", without even reading the ad text or even realizing it was promoting an offer, and not an entertaining video.
Of course that's an extreme example, but I'm sure you get my point.
So generally speaking, I would judge an ad by its overall CR and ROI, and not just by CTR/CPC.
Having said that: If you have experience experience with the niche/vertical and the particular audience, i.e. you know what benchmark stats to look for in a good vs. bad ad, then you may be able to take shortcuts and judge an ad based on CTR/CPC alone. Without knowing the CR and ROI, you'd still be guessing, but if you have enough knowledge and experience to guess correctly more often than not, then you could still be saving money overall.
Amy
Home > Paid Traffic Sources > Facebook & Instagram