19 {{ upvoteCount | shortNum }}
19 {{ upvoteCount | shortNum }}

Why affiliates don't test enough LPs?

by crysper
17 replies
I know most affiliates test only 2-3 LP versions...even if they pay for the traffic, and in PPV, mobile or even adult, they get enough impressions to get statistically significant data in a short period of time, for a high number of variations.

Is this laziness, lack of patience or not having tools to automate the stuff?

If you check successful affiliates, they all test like mad... i know that newbies should stick to normal 3-5 lps, but the other ones who still get traffic why do you waste it?

I heard that people are afraid of testing multiple variations because they are afraid that will pay too much to get significant data....but that's bullshit, which sounds like an excuse, BECAUSE even if you test 10 variations giving 1000 impressions to each, you'll get enough data to stop the ones the have very low CR/CTR...and you'll mostly reach to top 3-5 who might be converting the same at the beginning, so you'll still be earning more than losing... and is more likely to get 3 good variations starting from 10 than starting from only...3.

So what is it?

Newbie guys who are getting less than 1k impression per day or who are not profitable, please don't respond Thanks
#affiliates #landing page #lps #testing
  • Profile picture of the author Silentkiller1
    Laziness.. I guess...
    Signature
    {{ DiscussionBoard.errors[8870600].message }}
    • Profile picture of the author crysper
      Yeah, that's the main reason in my opinion as well, other guys?
      {{ DiscussionBoard.errors[8870624].message }}
  • Profile picture of the author vask
    Are you talking about testing multiple landing pages at the same time? Or over the lifetime of a campaign?
    {{ DiscussionBoard.errors[8870671].message }}
  • Profile picture of the author Greedy
    I'm testing much more, but not from the start
    That was my vote
    Signature
    {{ DiscussionBoard.errors[8870672].message }}
    • Profile picture of the author crysper
      Originally Posted by vask View Post

      Are you talking about testing multiple landing pages at the same time? Or over the lifetime of a campaign?
      At the same time mostly.

      Originally Posted by Greedy View Post

      That was my vote
      But what strategy do you guys use? You test 3 LPs, get the first best one and you make another 3 variations based on that...or you test AGAIN 3 than other 3, but not at a time? If they are totally different, why not at all? If they are variations from the preceding best performing one, have you tried multivariate testing?
      {{ DiscussionBoard.errors[8870757].message }}
  • Profile picture of the author Inferno272
    every LP you test to start out with makes your opening budget that much bigger...not everyone can afford to test 10 lps

    i mean hell, if its a 20 dollar payout, then you gotta test 20X2X10

    thats a 400 dollar test budget on a 20 dollar product.

    that doesnt even include testing ad copy (unless by LP you do mean the ad copy....PPV sorta combines ad copy and actual landing pages...in media buying, youd have to test both)...so if youre testing say 10 ads as well, thats a 4 thousand dollar test budget for a 20 dollar pay out that may never convert.
    {{ DiscussionBoard.errors[8870856].message }}
    • Profile picture of the author crysper
      Originally Posted by Inferno272 View Post

      every LP you test to start out with makes your opening budget that much bigger...not everyone can afford to test 10 lps

      i mean hell, if its a 20 dollar payout, then you gotta test 20X2X10

      thats a 400 dollar test budget on a 20 dollar product.

      that doesnt even include testing ad copy (unless by LP you do mean the ad copy....PPV sorta combines ad copy and actual landing pages...in media buying, youd have to test both)...so if youre testing say 10 ads as well, thats a 4 thousand dollar test budget for a 20 dollar pay out that may never convert.
      I got what you said, but I don't mean testing if the offer converts etc...

      I mean when you know that offer ALREADY converts(previously tested) and you know also the right targets...I refer only to LP testing, that's all. The thing is you already paid for the traffic, why not getting the most of it?

      Most of the cases, affiliates test 3 lps even if they get traffic that could make decisions for 10 different LPs and on PPV it happens very common. Of course, I'm not talking about noobs or beginners, i'm talking about the established affiliates who know the stuff.

      An example is that you are already getting some kind of profitable campaign on ppv, you start pushing more and more traffic, but you keep your LP testing at only 3 lps when you can test 10.
      {{ DiscussionBoard.errors[8870984].message }}
  • Profile picture of the author tallchase
    For me (and I'm only getting 5k impressions a day so small) it's a time 80/20 thing. I get a campaign profitable and go through several rounds of iterative optimization. The amount of work to squeeze 20% more profitability out through much more testing may ( or may not) have a better roi on my time vs starting a new campaign. That is, specifically, making new lps takes time. My .02
    {{ DiscussionBoard.errors[8871927].message }}
    • Profile picture of the author crysper
      Originally Posted by tallchase View Post

      For me (and I'm only getting 5k impressions a day so small) it's a time 80/20 thing. I get a campaign profitable and go through several rounds of iterative optimization. The amount of work to squeeze 20% more profitability out through much more testing may ( or may not) have a better roi on my time vs starting a new campaign. That is, specifically, making new lps takes time. My .02
      Yea, it's more safe to squeeze profits from an existing campaigns with targeting and offer set. But do you use something like 70% goes to winner LP and 30% to test other versions...or no strategy? Oh, I just see you get 5k/day
      {{ DiscussionBoard.errors[8872050].message }}
  • Profile picture of the author xrcv
    Testing 10 LPs will usually take a few thousand dollars, considering you are testing a worthwhile product and not some BS zip submit. If you don't see conversions upfront it's difficult to continue investing. Not everyone has a 10k budget for testing.
    {{ DiscussionBoard.errors[8872602].message }}
    • Profile picture of the author crysper
      Originally Posted by xrcv View Post

      Testing 10 LPs will usually take a few thousand dollars, considering you are testing a worthwhile product and not some BS zip submit. If you don't see conversions upfront it's difficult to continue investing. Not everyone has a 10k budget for testing.
      It depends on the conversion rate and how fast you kill the non-performing ones. Statistically, you don't need to burn 10k to test 10 pages because if the differences are huge compared to the best converting one, you should remove those LPs from the test from the early start.

      What I understand from this thread so far, is that people don't test at the beginning because they are afraid the offer doesn't convert. After they get more trust in the offer/traffic, they start to test more variations. On this phase, how much variations do they test is related to their return on time invested to create new variations.
      {{ DiscussionBoard.errors[8872765].message }}
  • Profile picture of the author Greedy
    But what strategy do you guys use? You test 3 LPs, get the first best one and you make another 3 variations based on that...or you test AGAIN 3 than other 3, but not at a time? If they are totally different, why not at all? If they are variations from the preceding best performing one, have you tried multivariate testing?
    Test 3, find best, test 3 based of it, pick best, test more based off of it, and maybe some new wildcard ones.
    Signature
    {{ DiscussionBoard.errors[8874155].message }}
  • Profile picture of the author PPC-Coach
    I find most refuse to test ONE!

    They just want to direct link and give up quickly.

    Signature
    {{ DiscussionBoard.errors[8874307].message }}
  • Profile picture of the author vask
    At first I'll test just 2-3 extremely different landing pages, once I establish some stats, I'll start testing variations.

    Like for dating, I'll usually test the "Profile" LP, the "Rules" LP, and the standard dating LP. Then I'll go ahead and start testing variations (eg. copy, colors, layout, etc). I can easily test over 30 different variations and it doesn't really take that long to get the LPs set up. I'm pretty well versed in html/css/php though
    {{ DiscussionBoard.errors[8874430].message }}
    • Profile picture of the author crysper
      Originally Posted by vask View Post

      At first I'll test just 2-3 extremely different landing pages, once I establish some stats, I'll start testing variations.

      Like for dating, I'll usually test the "Profile" LP, the "Rules" LP, and the standard dating LP. Then I'll go ahead and start testing variations (eg. copy, colors, layout, etc). I can easily test over 30 different variations and it doesn't really take that long to get the LPs set up. I'm pretty well versed in html/css/php though
      Man, that's the BEST answer so far....and from this I assume you are pretty successful as well.

      That's the way to test LPs and even ads(talking by 2 years experience running CrazyCTR). Test totally different stuff....than tweak based on what works.
      {{ DiscussionBoard.errors[8874497].message }}
  • Profile picture of the author watsonovedades
    you need big money to test other then "whats working" i guess some affiliates are happy with finding a golden nugget who will churn and burn and off they go looking for another successful campaign, thats not necessary a bad thing,

    thats the fun of affiliate marketing, you dont wantt o stick with one thing, you can to use trends and other stuff to get cheaper clicks
    Signature
    Whatever your mind can conceive and BELIEVE you can achieve
    Follow me on Twitter - @DineroConPc
    I talk about Affiliate Marketing Methods
    {{ DiscussionBoard.errors[8874927].message }}
  • Profile picture of the author biggenius
    Because its costly both in terms of money and time. On the top we want stability instead of testing 100s landing pages. But if campaign is large enough, we can test multiple variations using automated tools.
    {{ DiscussionBoard.errors[8875085].message }}

Trending Topics