Testing Campaigns - Stats that Matter

wiredniko

wiredniko

Jedi in training
Jul 20, 2010
712
26
0
New York
I keep working on new campaigns and I keep failing. I believe one of my problems is that I am looking at the wrong stats.

So when you are working on a campaign and trying to optimize it what do you pay attention to?

My last campaign was in PoF so I was focusing on my CTR, since in essence it lowers my eCPC. I figured the lower my eCPC the larger my profit margins.

Then I thought, eCPC is great but what is an ad gets clicked a lot BUT no one is converting....so I started looking at my eCPM.

Is this the right path?
 


wiredniko said:
I keep working on new campaigns and I keep failing. I believe one of my problems is that I am looking at the wrong stats.

So when you are working on a campaign and trying to optimize it what do you pay attention to?

My last campaign was in PoF so I was focusing on my CTR, since in essence it lowers my eCPC. I figured the lower my eCPC the larger my profit margins.

Then I thought, eCPC is great but what is an ad gets clicked a lot BUT no one is converting....so I started looking at my eCPM.

Is this the right path?
Click to expand...


First and foremost, watch your ROI. Great CTR and lower click costs for an ad doesn't always mean a great ROI for that ad. That ad could still convert like crap and go in the red. Likewise, ads with bad CTR and higher CPCs (compared to the rest of your ads) could still perform so well conversion-wise that they're doubling your money.

I assume you track everything down to the ad level using Tracking202 or similar? If you have a bunch of different ads you should be tracking them separately to compare their performance. Then keep an eye on each ad's ROI.
 
The ONLY stat that matters in the end is your CPA. As long as your CPA is lower than what your payout is your obviously making money. The other stats such as your CPC and CPM, ctr and engagement (how long someone actually stays on your page) are just factors you have to tweak and test, but in the end, I've had terrible ctr ads that have huge conversions and vice versa, at the end of the day any combination of your factors will make it work, you just have to test through em to see. For example: if your campaign has a bad lp ctr but high conversions, testing to raise your ctr may actually results in a lower conv rate on your backend, but if your CPA is lower as well, who gives a fuck? It's not rocket science its just simple testing and just looking at your stats to see which test performs best CPA wise...
 
I find it kind of amusing that I have three different opinions on the same question...yet they all sound right.
 
Obviously, the only thing that matters is ensuring that your ad spend is less than your revenue. The other stats are useful insofar as they can help you optimize. You can try all sorts of tricks to game POF users and increase your CTR and lower your CPC, but at the end of the day, it doesn't matter that tons of people are clicking your ad, it only matters if those people are the RIGHT people; i.e., people who will convert because they want what you are offering.

As an exercise, I would forget about stats for a second, and try to focus on understanding my demographic, and specifically how users are interacting with my ads, what their thought process is like when they click, and how they will respond to the offer page. Make sure everything is relevant to their interests, state of mind, and desire at a specific point in time.
 
wiredniko said:
I find it kind of amusing that I have three different opinions on the same question...yet they all sound right.
Click to expand...

That's because they're not really different opinions. They are three ways of saying the same thing: Making sure that each ad / landing page / creative / whatever you are testing, each gives you the most return for your buck. In other words, your ROI.

CPA, EPC, NEPC,...etc... these are all variables that in the end simply affect your ROI either way you look at it. It's just math really.

My point was that you gotta look at both sides of the equation, not just the input side of things (ad performance and cost), but also the output (offer conversion and return).
 
davidne said:
epc > cpc = $$
Click to expand...

well yea, of course.

But if you have a landing page, 100% of your clicks will not click-through. So you have to factor that in.

In the end, I just just look at ROI though. And eCPM between different ads
 
Garrett said:
well yea, of course.

But if you have a landing page, 100% of your clicks will not click-through. So you have to factor that in.

In the end, I just just look at ROI though. And eCPM between different ads
Click to expand...

true... if you are using LPs i guess you will have to look at ROI and etc.
 
You must log in or register to reply here.
Top Bottom