A friend who runs a substantial amount of POF traffic came back from ASW and mentioned that the first few session depths for the female demographic might not be worth bidding for.  The rationale being that female users, who tend to get a lot of messages, head immediately to their inboxes and do not notice the ads.

I decided to put this theory to the test with a case study testing the performance of a campaign with no session depth targeting and with session depth targeting > 5.  The campaign that I used targets females who are ages 24-29, and I set my bid at $0.87.  I won’t go into too much detail on the rest of the targeting as I still want to try to run this for a bit longer 😉

I had been running the campaign for a bit already before the start of the case study and have weeded out the poor performing ads before starting the test.  I ran the 2 variations of the campaign each for 5 days.  From variation to variation, I only changed the session depth targeting in order to run a controlled experiment.

I am going to provide the creatives below that I ran for this test.  This angle is doing quite well right now, and if you feel up to it, put your own spin on it and run it as well.  I highly recommend you do not just straight up copy these ads as I’m sure multiple people are starting to run them as we speak.

Case Study Creatives

Female Session Depth Case Study

So what are the results?

5 Day Performance With No Session Depth Targeting

Female Session Depth Case Study

5 Day Performance With > 5 Session Depth Targeting

Female Session Depth Case Study

What conclusions can we draw from this.  Clearly, while it’s only 5 days of data for each case, the amount of traffic was affected by limiting the session depth.  Performance of the case where we limited session depth seemed to perform slightly worse overall, while the conversion rate seemed to be slightly higher.  I didn’t think the CTR would dip as much as it did, but part of it may be due to volatility. The study seems to still be on the side of lower session depth being higher quality traffic.

It seems that the theory might not be true after all.  However, please take this case study with a grain of salt, as it is only for one targeting.  We also only collected a few hundred thousand impressions, so the data might have said something different if we had collected a few million impressions.  I encourage you all do perform your own tests once you are at the point where you are finding profitable campaigns.  Not only do you have hard data backing your theory after completing the tests, you will have learned a great deal just carrying out your plan.

Hope you guys enjoyed this one!  I wonder how much longer this campaign will last…

  • Great CS, many thanks for sharing. Did you use a general dating offer for this angle or something niche specific? Always wondering how to close the gap with a lander and a broad dating offer…

    • It was a general dating offer. The soldier angle is what I would call a gimmick. It’s purely to play into the user’s fantasy and get them to click. Once they are at the landing page, it would be time to reel back the gimmick. Obviously, there is no site that caters to soldier dating (that would be weird). It’s important now to make the user believe that this site is real and down to earth, and that they can realistically gain the benefits promised. I try to mix in some average looking veterans and other types of people too so things are very believable.

    • affm

      ur a jerk

  • Thanks a bunch Tom, appreciate your time.

  • J.S

    Cool Case study…Surprised that such ad copy works, specially with females. The headline and first part of the copy is basically the same question repeated…plus ”Join Free” is also said 2 times! lol Got an explanation for that?

    • I didn’t think it would work that well to be honest (and the CTR’s aren’t THAT great). But I wanted to see if I jam one unique proposition and “FREE” in there twice what would happen. I guess it did ok, but I’m no psychologist so I can’t tell you why it performed the way it did.

    • J.S

       had you used any login count with these campaigns? Or you left it blank so it wouldn’t affect the session depth test?

    • I think this was either <50 or <100 login count.

    • J.S

       And you really needed a bid of 0.87$ to get clicks from that younger female demo? Its quite surprising you’re not just breaking-even with such a high bid & quite low ctr’s…