split testing landing pages

by dj2590
14 replies
Hey everyone, What would be a good amount of clicks to start with for split testing 2-3 landing pages, i plan on using well known solo ad providers with good testimonials and good reputations.. I was thinking about buying a total of 1000 clicks between 3 different providers.. or should i just stick with one provider for the split testing? I will be using clickmagick for tracking.
#landing #pages #split
  • Profile picture of the author mjcole76
    I always split test with about 300-500 clicks when i split test, and that's good enough for me, so 1000 would be even better.
    Signature
    Private Message Me To Get The 2 Most Costly Mistakes Affiliates Make And How You Can Avoid Them Make Ten
    {{ DiscussionBoard.errors[10989373].message }}
  • Profile picture of the author dj2590
    Do you divide all those clicks between multiple solo providers? or usually just one?
    {{ DiscussionBoard.errors[10989378].message }}
    • Profile picture of the author vMartin
      Originally Posted by dj2590 View Post

      Do you divide all those clicks between multiple solo providers? or usually just one?
      I see no extra benefit of using just one provider, in fact you will be far better off also split-testing the solo providers you are using and not just the landing pages.
      Signature
      PROMOTE OUR NEW PLATINUM CONVERTING OFFER
      AND SKYROCKET YOUR CONVERSION RATES THROUGH THE ROOF 2021!


      CLICK HERE AND FIND OUT Get up to $120 per sale
      {{ DiscussionBoard.errors[10989663].message }}
  • Profile picture of the author Regional Warrior
    Originally Posted by dj2590 View Post

    Hey everyone, What would be a good amount of clicks to start with for split testing 2-3 landing pages, i plan on using well known solo ad providers with good testimonials and good reputations.. I was thinking about buying a total of 1000 clicks between 3 different providers.. or should i just stick with one provider for the split testing? I will be using clickmagick for tracking.
    The easy way is to forget the clicks and other BS and just set up the page(s) A/B/C and see which ones give you the most sign ups use that one

    Jason
    {{ DiscussionBoard.errors[10989769].message }}
    • Profile picture of the author dj2590
      yes, that makes sense.
      {{ DiscussionBoard.errors[10989926].message }}
  • Profile picture of the author rritz
    I usually split test with a minimum amount of 500 clicks to each landing page. You need statistically valid data.
    I also use different providers, although I have found that in most cases the results from 100 clicks are better than the results from say 1000 clicks.
    Can't say why.
    {{ DiscussionBoard.errors[10989804].message }}
  • Profile picture of the author mdallen
    The point of split testing is not to have too many variables. If you test multiple pages and multiple solo ad providers you will be unable to tell which variables affected the difference in outcomes. It is best to test two pages with one difference so that you can see which works better. refine one item at a time or you will just be guessing
    {{ DiscussionBoard.errors[10989806].message }}
    • Profile picture of the author vMartin
      Originally Posted by mdallen View Post

      The point of split testing is not to have too many variables. If you test multiple pages and multiple solo ad providers you will be unable to tell which variables affected the difference in outcomes. It is best to test two pages with one difference so that you can see which works better. refine one item at a time or you will just be guessing
      "The point of split testing is not to have too many variables."

      Lost for words... (Please anyone reading this NEVER EVER take this advice)

      You are basically telling OP to split-test as little as possible to find out each and every single variable that could potentially add some extra ROI versus split-testing several completely different approaches/angles that could actually make a HUGE impact on ROI.

      Headline testing should be done for optimizations purposes ONLY.

      What happens if OP tests let's say 4 different headlines with the same exact landing page and angle and the ROI difference is a matter of 10/5%.

      The chances his entire approach/angle/landing page are not profitable from the start is huge which makes split-testing just the headlines or other small aspects absolute useless.

      He should be testing different angles/landing pages/approaches that are VERY different from each other and see how that performs first. After he finds a winner he can work on optimizing his page even further but NEVER from the beginning.

      Obviously all this will depend on the budget but I would not recommend anyone to venture in paid traffic with a low budget anyway.

      Please again for anyone reading this AVOID mdallen's advice at all costs.
      Signature
      PROMOTE OUR NEW PLATINUM CONVERTING OFFER
      AND SKYROCKET YOUR CONVERSION RATES THROUGH THE ROOF 2021!


      CLICK HERE AND FIND OUT Get up to $120 per sale
      {{ DiscussionBoard.errors[10990021].message }}
      • Profile picture of the author dj2590
        i guess the more i read this the more it does make sense.
        {{ DiscussionBoard.errors[10990054].message }}
  • Profile picture of the author dj2590
    Anybody use ighor khiefets solo ads? i was thinking about trying one of his solo ads but they are SO Expensive lol
    {{ DiscussionBoard.errors[10989929].message }}
  • Profile picture of the author JohnMcCabe
    Originally Posted by mdallen View Post

    The point of split testing is not to have too many variables. If you test multiple pages and multiple solo ad providers you will be unable to tell which variables affected the difference in outcomes. It is best to test two pages with one difference so that you can see which works better. refine one item at a time or you will just be guessing
    Originally Posted by vMartin View Post

    "The point of split testing is not to have too many variables."

    Lost for words... (Please anyone reading this NEVER EVER take this advice)

    You are basically telling OP to split-test as little as possible to find out each and every single variable that could potentially add some extra ROI versus split-testing several completely different approaches/angles that could actually make a HUGE impact on ROI.

    Headline testing should be done for optimizations purposes ONLY.

    What happens if OP tests let's say 4 different headlines with the same exact landing page and angle and the ROI difference is a matter of 10/5%.

    The chances his entire approach/angle/landing page are not profitable from the start is huge which makes split-testing just the headlines or other small aspects absolute useless.

    He should be testing different angles/landing pages/approaches that are VERY different from each other and see how that performs first. After he finds a winner he can work on optimizing his page even further but NEVER from the beginning.

    Obviously all this will depend on the budget but I would not recommend anyone to venture in paid traffic with a low budget anyway.

    Please again for anyone reading this AVOID mdallen's advice at all costs.
    Actually, you are both right.

    If the OP tests, say, three landing pages and three different ad vendors at the same time, there are nine possible combinations that yield results.

    I would run this as three separate tests. This way, you can tell which vendor's clicks reacted to each landing page. You may (probably will) find that each vendor's list performs differently with each landing page.

    vMartin's advice on testing wildly different landing pages at this point is spot on.

    If you keep each vendor's stats separate, your one variable is the landing page.

    If you find that one page way outperforms the other two on each test, but they aren't the same page, look at what's different about the sources and try to figure out why they gave different results. You may want to repeat the tests to see if the result holds up.

    Which brings up another good point from vMartin, even though he didn't spell it out.

    Testing is an ongoing process. You start by testing the big stuff first, then refine from there.

    If you run a test on three widely different LPs and one looks like a runaway winner, repeat the test. If the result holds, you can retire the two losers and start refining the winner.

    Keep at it until you reach a point where any improvement costs more than it's worth. Why spend time and money trying to wring an extra 0.01% out of an offer when you can make a lot more progress by starting the process again on a new offer or product?
    {{ DiscussionBoard.errors[10990200].message }}
  • Profile picture of the author dj2590
    i've never really hard anybody mention this before, but can you split test an email follow up series too? lets say like 7 day series? I'm going to be working really hard on this... im going to keep testing everything until i can have a funnel that converts really well.
    {{ DiscussionBoard.errors[10990215].message }}
    • Profile picture of the author JohnMcCabe
      Originally Posted by dj2590 View Post

      i've never really hard anybody mention this before, but can you split test an email follow up series too? lets say like 7 day series? I'm going to be working really hard on this... im going to keep testing everything until i can have a funnel that converts really well.
      Yes, you can.

      The best way is to take your best shot at what you think will work, run some people through the sequence, and then look for any trouble spots.

      For example, let's say you see a sharp drop off on Day 3, meaning that most don't even see Day 4. They either unsubscribe after Day 3 or they simply don't open Day 4.

      You have two possibilities here. Either there's something wrong with Day 3's message, or there's a disconnect between Day 3 and Day 4

      To test, you'd need to run two tests.

      In the first one, you send subs to either of two nearly identical sequences. The first has a different message on Day 3 as the only difference. The other has a different subject line on Day 4. Does one change solve the problem?

      It might take multiple rounds of testing until you find the answer.

      Once you do, start over and look at results, looking for the next bottleneck.

      Repeat the 'think, test, examine' sequence until you optimize your funnel.

      One thing I forgot to mention earlier...

      Sometimes the result of your test will be inconclusive. There won't be enough difference to warrant declaring a winner. When that happens, try changing something else and repeat.
      {{ DiscussionBoard.errors[10990229].message }}
      • Profile picture of the author dj2590
        Thank you very much for all of your help john, I'm going to give that a try.
        {{ DiscussionBoard.errors[10990288].message }}

Trending Topics