Network Shaving

Discussion in 'Mail Chat' started by airin, Dec 14, 2012.

  1. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    I have a feeling I am about to open a big ass can of worms with this thread, but oh well....

    Recently it has come to my attention that some reputable networks are doing some serious shaving. I have no intention of outing them here because for those of you who know me, know that is not my style... What I do propose is that we get a few volunteers to do some legitimate split testing and post the results. The volunteers should not disclose what networks or offers they are split testing until AFTER they have tested to ensure that pre-emptive measures cannot be taken by any networks. We then let the numbers speak for themselves.

    This is very frustrating because we constantly get hit up from our partners saying other networks are offering higher payouts when we know what they are doing and don't want to make enemies by calling them out. This has been happening a lot more than usual lately, including a network that offered a mailer a higher payout than we were giving that network on one of our AOR offers that we host the site on.

    A prime example of this is Auto Price Finder... we keep hearing how mailers are getting paid up to $2.75 elsewhere when we know that nobody is even getting close to that much! These same mailers tell us that they are converting about 20% at the higher payout... which makes it obvious that there is some serious shaving going on because our network wide average is 30% with some TLD mailers converting at almost 60%.

    I think that this should be an ongoing project as well because network shaving needs to stop, along with a lot of other shady practices. I am so over people not doing business the right way and screwing people over!
     
  2. roundabout

    roundabout VIP

    Joined:
    Feb 17, 2011
    Messages:
    2,713
    Likes Received:
    154
    Trophy Points:
    63
    Count me in.... very interested to see how this develops.
     
  3. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    Nice... who else is in?

    Round - I think you should coordinate this, seeing as it would be a conflict of interest for me to be involved anymore than getting the ball rolling with this post. =)
     
  4. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!

    We usually do end up getting most people to split test and they see the results, so it doesn't hurt us that much and in some cases does help. The point is not what it is doing to my network, it's that it is taking away from the amount of money being made by the mailers on this forum.

    I also have no plans to publicly humiliate anyone... I would just like numbers to speak for themselves.
     
  5. nickphx

    nickphx VIP

    Joined:
    Apr 2, 2011
    Messages:
    1,139
    Likes Received:
    363
    Trophy Points:
    83
    Gender:
    Male
    Location:
    guadalajara, chiuhuahua
    You're forgetting that advertisers also shave.. Probably more so than the networks. They all justify their thievery by claiming fraudulent traffic is high..
     
  6. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    I 100% agree this is bullshit too. We cut advertisers when we catch them doing shady stuff, especially since we have so little fraud in our network. We just had an advertiser try to claim fraud on legit email traffic... peace out!

    Shady advertisers are just as bad as shady publishers and networks. BS like this is the #1 reason why I started Crush... to do business the right way. /rant
     
  7. Daansen

    Daansen Member

    Joined:
    Aug 14, 2011
    Messages:
    84
    Likes Received:
    16
    Trophy Points:
    8
    I agree with SuperGenii. It is not uncommon practice when networks are showing the higher payouts attempting to attract the affiliates, simultaneously shaving in the way so that the final revenue is the same as it would be without these tricks.
    In your example, airin, we can see that 30% with $2 payout would be pretty similar to 20% with $2.75. Approximately.
    But anyway, it's unfair play, and it would be wise rather to avoid working with such networks, when possible.
     
  8. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    There is no way it's backing out the same... 30% is our network wide average with TLD guys in the 40 - 60% and GI mailers w/out a ton of clickbots in the 25 - 35% range. The 20% was coming from a TLD mailer, so even if we went with the lower range of 40% they would need to double our payout to be in the same range... if the split test comes back in the higher range that would be almost triple our payout.

    I fully understand that it's hard to run a profitable network on the standard margins, but it doesn't mean that it's okay to resort to tactics such as this. I also fully understand how hard it is to be profitable as a mailer... it's very expensive for all for all of the required hard resources and takes a lot of time. I used to be on that side. It pisses me off when I see mailers not getting what they rightfully earned, which is my main motive in starting this thread.

    Now who is in to do some testing? If the community as a whole does not come together against stuff like this it's never going to get better.
     
  9. DKPMO

    DKPMO VIP

    Joined:
    Mar 31, 2011
    Messages:
    1,452
    Likes Received:
    68
    Trophy Points:
    48
    Location:
    Elaborate Underground Base
    It would be fun to see some sort of report, but IMHO its primary value would be for entertainment purposes.

    The fact of the matter is any network or any advertisers can turn on or off any "lead acceptance" dials of any kind any time. You will only see it when you look / analyze and only when you are looking. Just because you find something in a random split testing sample does not mean that next day some other network that was not yet caught will do the same OR the advertiser will cut "lead acceptance" on their end just for that network SubID. But even if no foul play is involved a benign volatility in conversions or quality of tracking is also possible. Finally let's not forget the variations across offers and publishers - intentional or not.

    The only way to deal with shaving or volatility is to have technology that can catch this on the fly continuously and separate trends (EPC & eCPM) from noise in a large enough pool of ads.

    Of course the skill set required to do this properly goes way beyond what's typical for this industry.
     
    Last edited: Dec 14, 2012
  10. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    Definitely. The testing would have to be done privately and in a uniform manner. Something like this would work...

    1. A minimum of 3 mailers testing the same offers through the same 3 - 4 networks.
    2. The mailers would have to mailer to evenly split random segments of the same lists (not share lists, but on an individual basis)
    3. The mailers would have to send at relatively the same time
    4. The mails would have to send the drops from the same server (individual basis)

    It would definitely take effort to orchestrate this, but I think if you did something along these lines you would have a pretty good indication if the network is shaving and also if an advertiser is shaving. You would most likely see this if there were drops across all mailers on the offers at the same time, though advertisers do shave on individual publishers and networks.


    Do you have an idea for your next project now? =)
     
  11. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    Good to know, thanks!!

    I also do not want this to be "my" test, as I obviously have a bias. I would like for someone to run with the ball and keep me out of it until there are some results.
     
  12. DKPMO

    DKPMO VIP

    Joined:
    Mar 31, 2011
    Messages:
    1,452
    Likes Received:
    68
    Trophy Points:
    48
    Location:
    Elaborate Underground Base
    This is not a "next project", it is just a feature/subsystem of the one you are aware of. I am very happy to report that this part has been implemented and successfully passed basic simulations.

    I'll be happy to chat about this at ASW, but not really planning to release any further details on open forum for now.

    :top::top::top:
     
  13. airin

    airin Member

    Joined:
    Mar 22, 2011
    Messages:
    766
    Likes Received:
    20
    Trophy Points:
    18
    Location:
    Key Largo, FL!
    Yes, let's talk! I think the conversation will be more than we can have at the football event. Lunch on Monday? =)
     
  14. DKPMO

    DKPMO VIP

    Joined:
    Mar 31, 2011
    Messages:
    1,452
    Likes Received:
    68
    Trophy Points:
    48
    Location:
    Elaborate Underground Base
    Definitely. Let's chat on AIM/Skype in the next week or so. I have not even started scheduling meetings, but we definitely should talk further.
     

Share This Page