POF Targeting and Testing

9 replies
Hey everyone,

I got another two questions on POF split testing.

1) What should the testing process look like?

- Do you first test let's say 15 different images with the same 'loose' targeting (e.g. 15 images shown to males age 30-50)?
By that you could probably see what age groups etc. click the most and THEN adjust the targeting, right? Or doesn't POF (or Prosper) show exact statistics what age groups etc. clicked?

- OR do you target very specifically from the beginning?
Let's say 15 images, images 1-5 shown to males age group 30-35, images 6-10 shown to males age group 35-40, images 11-15 shown to males age group 40-45 (plus maybe more targeting options like body type etc.)?

Please be precise when answering!



2) For how long do you test one campaign? I read on the POF blog that you should test it for 7 whole days- But that means a huge amount spent just one testing, doesn't it? Or do you just say I'm giving every ad 5,000 impressions, doesn't matter if it's all spend in one day or in a couple of days?

Thanks a lot!
#pof #targeting
  • Profile picture of the author Greedy
    1) What should the testing process look like?

    Do more of way #1. You want the 15 to be tested to everyone, not split up across different ages. When testing you need to make things as constant as possiable

    Test until you get an image with a strong CTR .12+.

    Then start playing with the ad copy to boost your CTR even more, ad copy effects conversions a lot as well. So make sure you have tracking setup to see how things convert. You need to pay attention to your CTR and conversions closely especially when tinkering with your ad copy.

    Lots of people will say shoot for .15+ CTR. Because this will keep you getting you the volume you need long term.

    2) For how long do you test one campaign? I read on the POF blog that you should test it for 7 whole days- But that means a huge amount spent just one testing, doesn't it? Or do you just say I'm giving every ad 5,000 impressions, doesn't matter if it's all spend in one day or in a couple of days?


    Spend 1.5x the offer's payout testing each image in each campaign.

    Don't worry about going for 7 days straight. Just make sure you test your ads all at the same time. You should test them all on one day. If you can't get enough traffic to test on one day play with your targeting so you can get more volume. Evenings are usually better for dating.

    Start with targeting set very tightly, you can always expand later.

    Here is free guide I posted. If you want to check it out....

    How I Made $1,000 off Plenty of Fish

    Good luck, be persistant you'll crack things!
    {{ DiscussionBoard.errors[7571429].message }}
  • Profile picture of the author Piren
    Testing during a week will be better. Users usually click the ads more active on the weekend and evenings. Of course, everything can be different with some offers, but usually the longer the test, the more clearer the result would be and then you will be more sure what works the best and bring you more profit.
    {{ DiscussionBoard.errors[7571590].message }}
  • Profile picture of the author seregap
    im sure everyone has their own way. it really depends on the offer for me for example.

    if it's a general dating offer i would start with a bunch of images and generic copy and run high volume in short period of time (with wide targeting) to find the winner images without worrying about conversions. then you can start breaking down your audience into smaller groups working on conversions by adjusting ad copy and adding landing pages to suit your smaller groups, introducing angles in other words. this way you ideally end up with the same high volume, but broken down into smaller converting campaigns.

    with a very niche offer on the other hand you cant start with high exposure because of small audience and higher cpms, so i would think of a very tightly targeted angle within the niche and try and make it convert straight away, then move on to add another angle or target and so on. this way you limit exposure and make your overall campaign last longer hopefully.
    {{ DiscussionBoard.errors[7571603].message }}
  • Profile picture of the author PaulBaker
    Originally Posted by saphiry View Post

    with the same 'loose' targeting (e.g. 15 images shown to males age 30-50)?
    hi Saphiry,

    I would break those age groups down into 5 years, there can be a huge difference between 36-40 and 40-44 year olds for example. You wont know which group is or isn't performing if you run them all together. You can track age with dynamic tags and subids through prosper but personally I find smaller target groups make it easier to get a snapshot of ctrs and cpcs etc. Hope that makes sense.

    For how long do you test one campaign? I read on the POF blog that you should test it for 7 whole days
    I would be hesitant on testing your campaigns for one day only. If you can't afford to test for a full week or the campaign isn't showing enough promise to warrant it, I would suggest at least running it again a few days later. Lots of factors can effect the day to day performance of a campaign, day of the week, weather, competition etc. You want to make sure your campaign didn't fail due to just really bad timing.
    Signature

    Put Your Money Where Your Mouth is: CLICK HERE

    {{ DiscussionBoard.errors[7571722].message }}
  • Profile picture of the author ChrisBa
    Originally Posted by saphiry View Post

    Hey everyone,

    I got another two questions on POF split testing.

    1) What should the testing process look like?

    - Do you first test let's say 15 different images with the same 'loose' targeting (e.g. 15 images shown to males age 30-50)?
    By that you could probably see what age groups etc. click the most and THEN adjust the targeting, right? Or doesn't POF (or Prosper) show exact statistics what age groups etc. clicked?

    - OR do you target very specifically from the beginning?
    Let's say 15 images, images 1-5 shown to males age group 30-35, images 6-10 shown to males age group 35-40, images 11-15 shown to males age group 40-45 (plus maybe more targeting options like body type etc.)?

    Please be precise when answering!



    2) For how long do you test one campaign? I read on the POF blog that you should test it for 7 whole days- But that means a huge amount spent just one testing, doesn't it? Or do you just say I'm giving every ad 5,000 impressions, doesn't matter if it's all spend in one day or in a couple of days?

    Thanks a lot!
    I test a variety of different images, different ads and different age groups. It's a lot of work to set this up and cost a fair bit of cash but its the proper way to test IMO.

    As for how long to test, it depends on how much the offer pays and how much traffic you get
    {{ DiscussionBoard.errors[7572119].message }}
  • Profile picture of the author saphiry
    Thanks a lot, this was really helpful

    One additional question: I'm a little bit confused about the bidding. I read several case studies and guides- In some they recommend to start bidding very high (more than the maximum recommended bid of 0.80$), in others they recommend to find the golden middle between 0.45 and 0.65$.

    What is the best stragety? And when can I start to decrease my bid?
    {{ DiscussionBoard.errors[7574253].message }}
  • Profile picture of the author seregap
    there's no bulletproof strategy.
    as a general rule you bid lower on broad targeting and higher on niche targeting.
    but it all depends on your epc and competition.

    you might find an angle with low competition where a drop in bid from .80 to .40 will result in the same traffic volume and quality, while there might be targets where a drop from .60 to .50 will cut your traffic in half and reduce conversions to 0

    while testing ads, images and lps i usually keep my bids quite high to have the best quality traffic and good volume.

    then base your starting bid on expected epc from your testing keeping it profitable and go lowering by .02-.05 every now and then while monitoring traffic volume and quality to find your golden middle
    {{ DiscussionBoard.errors[7574363].message }}
    • Profile picture of the author calciferol
      yeah it depends on the offer pay out and your current cash flow situation and how long ur willing to find that one holy grail campaign which covers losses of all the campaign failed
      {{ DiscussionBoard.errors[7582744].message }}
  • Profile picture of the author Greedy
    What is the best stragety? And when can I start to decrease my bid?
    Bid high and start lowering it once you get good traffic, and conversions.

    I usually leave my bids around .55 for testing. But every campaign is different.
    {{ DiscussionBoard.errors[7583650].message }}

Trending Topics