I’ve noticed on Plenty of Fish that my impression numbers do not change much, regardless of whether I bid the minimum CPM or a much higher CPM. So, I’ve decided to do a public case study on Plenty of Fish so we can all see if it matters and how much. The only thing I will be split testing is my CPM bid. I will bid 15c, 25c, 35c and 45c.
With some affiliates experiencing problems getting blacklisted from offers they promote on Plenty of Fish, I’ve decided to promote a dating offer where the quality isn’t expected to be top notch in the first place. With this in mind, I’ve picked the Singlesnet offer on Profit Kings Media that accepts traffic from users that are 21 – 24 years of age.
I will only be using one ad copy. The ad title will be: Single Available Women and the description will be: There are 986 sexy, single women online in {state:default}. Sign up and meet them today! The {state:default} will be replaced by the users state they registered in, so it will look more customized for them.
I will be targeting males in the United States that are 21 – 24 and are single, divorced, widowed or separated.
I’ve picked 6 random images from Justin Dupre’s image pack that he released on his blog. The ones I’ve choosen are below.
I will be automating the creation of these ads using Mr. Green‘s Plenty of Fish Ad Uploader. I suggest you buy a copy today!
I will report back next week on the numbers.
This is a follow-up to Split Testing on Facebook Case Study: Part 1 and Split Testing on Facebook Case Study: Part 2.
After having 2 sets of images have their CTR die out on me I got lazy and tested out another another ad copy and saw my conversions tank by 54% so I figure the ad copy had to be misleading. So I decided I wanted to try out 5 completely different sets of ad copy and see the difference between their conversion rates. I tried to relate each ad title with the description. I used the same images as my first go around so I know they can perform decently and since I haven’t ran them in awhile their CTR should at least be respectable now.
Here are the stats for the last 7 days according to Prosper.
- Ad Copy 1 (Original): 8280 Clicks, 438 Conversions, 5.29% Conversion Ratio
- Ad Copy 2: 12612 Clicks, 498 Conversions, 3.95% Conversation Ratio
- Ad Copy 3: 441 Clicks, 9 Conversions, 2.04% Conversion Ratio
- Ad Copy 4: 323 Clicks, 10 Conversions, 3.1% Conversion Ratio
- Ad Copy 5: 721 Clicks, 34 Conversions, 4.72% Conversion Ratio
I found it pretty crazy that my original ad copy converted best for me after the split testing I did.
So here’s concrete proof that split testing makes a huge difference! If you’ve never split tested your ad copy, DO IT NOW! I’ve never done much split testing of ad copy until I ran this campaign and have finally came to an amazing revelation, even though I should’ve known it.
On Facebook your CTR depends on the image you use in your ad copy. Your conversion ratio depends on the text in your ad copy! It is imperative to split test both!
This is a follow-up to Split Testing on Facebook Case Study: Part 1.
Okay so the CTR began to die out on the images I found in Split Testing on Facebook Case Study: Part 1. I was still making some pretty good money off this campaign and decided I needed to switch up the ad copy. Me being pretty lazy I went to the merchant’s landing page and used some of the lines that caught my attention in my ad copy and mind you since I was lazy I only used one variation and scaled this out to every demographic. After about 5 days of continually submitting ads I finally got my ads past the review team at Facebook. I just knew I was going to kill it again for a few days and wasn’t sure what I was going to do after that, probably try a different ad copy.
So I threw up this one ad copy using those images I used in the first go round. I figure they haven’t been shown in a couple weeks, they should have some juice left in them. Plus with some new ad copy I should be sitting pretty.
I ran it for a couple days and lost a nice chunk of change, mainly because my conversion ratio had tanked by 54%. I’ve included screenshots below. And that amount of data was more than enough to be statistically relevant. My first concern was since I haven’t ran this offer in a few days maybe the advertiser was up to no good so I hit up my affiliate manager and asked him if the conversion rate for the network as a whole changed and he said nope. So I was thinking since the only thing that had changed was the ad copy, maybe the ad copy was now a bit misleading because it did include that the user could win money.
Part 3 I will be split testing 5 different sets of ad copy to determine how it affects my conversion ratio.
So here I was dominating this niche on Facebook so I scaled it out to different demographics and was doing great with this offer. I was effing pumped and got kinda lazy for a few days and I knew my CTR would die out eventually. I thought no problem, I’ll find some more images that are similar to the ones I’m using now and just make a few more ads and keep banking on this offer.
I finally figured out something I could search for that would bring up similar images on Google image search and I found a couple I thought were going to kill it so I saved them. I was talking to a friend about needing some more images and he told me to use Bing’s image search. I checked it out and oh my god! Bing’s image search is killer! Try it out sometime if you haven’t already.
So I found 88 images I thought might work well and I used the same exact ad copy. The only thing I did was change the picture. I knew to be profitable in this particular demographic for this offer I would need a 0.10% CTR or higher with the amount I was bidding, so that’s what I was looking for. If you’ll look at the screenshots below you’ll see I found a handful of keepers.
On the next part of this case study I’ll switch up the ad copy and see how well the ads perform.
