21 {{ upvoteCount | shortNum }}
21 {{ upvoteCount | shortNum }}

How many visits to determine conclusive results for conversion testing?

by MrPete2000
18 replies
A quick question, and one that I honestly could never quite figure out.

How much traffic should be pushed to a landing page or offer before you can be *confident* that you have established a baseline CTR or conversion rate?

Same question, but now for keywords and traffic sources?

The key point being statistical confidence in our conclusions... how much traffic?

In my experience, things either convert from the very beginning (say 100-200 visits) or they do not... but I don't know if that's normal. Can you help me (and the community at large) out by sharing your knowledge and experiences?

Cheers,
Pete.
#conclusive #conversion #ctr #determine #landing page #results #traffic #visits
  • Profile picture of the author Randall Magwood
    I'd say about 500 hits. Low number of targeted hits can be deceptive. So start with 500 targeted hits then make a conclusion from there.
    Signature
    ==>> Free book reveals make money with affiliate marketing
    {{ DiscussionBoard.errors[7276351].message }}
  • Profile picture of the author GarrieWilson
    1,000 visitors PER item being tested and should run over a week. I've also read "X" actions. Actions being sales, opt-ins or whatever you are doing but I forget what "X" is.

    Garrie

    PS The copywriting section may have a better answer.
    Signature
    Screw You, NameCheap!
    $1 Off NameSilo Domain Coupons:
    SAVEABUCKDOMAINS & DOLLARDOMAINSAVINGS
    {{ DiscussionBoard.errors[7276361].message }}
  • Profile picture of the author MrPete2000
    Yes, I'm also interested in tracking signups and such... so generally 500 - 1000 hits spread over a week (or two possibly) should provide decent results.

    I'm assuming you guys are referring to PPC traffic since it's usually very targeted?

    Anyone have numbers for other types of traffic? Media runs and CPM type traffic where you would likely get several thousand impressions daily (but not clicks)?

    PZ
    Signature

    ...

    {{ DiscussionBoard.errors[7277723].message }}
  • Profile picture of the author WillR
    There is no set number you need to send.

    All you need to know is the more traffic you send to an offer, the more accurate your stats will be (obviously). That's the only answer I can give you. I could tell you 1,000 visitors are needed to give you a good idea, but if you were to send 5,000 visitors then your stats would be even more accurate and set in stone. Personally I would not draw any conclusions with less than at least 1,000 visitors... anything less is just too small a sample.
    Signature

    .


    .

    {{ DiscussionBoard.errors[7277741].message }}
  • Profile picture of the author Tyler Stearns
    I agree with Will. Each source of traffic will be different. One source won't be as targeted as the next. There are a lot more variables that contribute to CTR than just your landing page layout. The more traffic you can send, the better the test will be.
    {{ DiscussionBoard.errors[7278368].message }}
  • Profile picture of the author dlane1987
    2000 will give you a good idea of your convertions
    Signature
    Minecraft & SEO YouTube Channel
    {{ DiscussionBoard.errors[7279275].message }}
  • Profile picture of the author danwood
    It depends but in some cases, 100 hits can be enough to test
    Signature
    3123 Subscribers for FREE?!
    {{ DiscussionBoard.errors[7279282].message }}
  • Profile picture of the author TerranceCharles
    I usually go with about 100 for each test using a squeeze page to measure opt-in conversions
    Signature
    {{ DiscussionBoard.errors[7279668].message }}
    • Profile picture of the author Alexa Smith
      Banned
      Ooh, a statistical question.

      The number of hits you need to measure the conversion-rate with a high (90%/95%) chance of its being statistically significant (i.e. "valid") is in inverse proportion to the conversion-rate itself. If you had a conversion-rate of 50% (as people can, with a squeeze page), even something as low as 30 hits could be statistically significant. If the conversion-rate is as low as 1%, as it can be with a sales page and not-too-highly targeted traffic, then you'll need a really huge number, well into 4-figures.

      Don't forget that the conversion-rate may depend almost entirely on the traffic-source. (For example, as I know so well from my own business, many sales pages that don't convert SEO traffic at all can convert opted-in, warmed-up, subscriber traffic consistently well).
      {{ DiscussionBoard.errors[7279924].message }}
  • Profile picture of the author MrPete2000
    Thanks all, very very helpful indeed.

    I'm usually running tests with about 400 - 1000 visits (generally) but getting bad results. I know my landing pages aren't too great but I'm also questioning the traffic quality (targeting) but that's because it's mostly CPM type traffic.

    It might be time to test with PPC or similar and then run the higher volume campaigns on creative with "proven" performance.

    Cheers,
    Pete
    Signature

    ...

    {{ DiscussionBoard.errors[7281148].message }}
  • Profile picture of the author GarrieWilson
    You should also consider that "A" may have more opt-ins than "B" but make less money. A true test would track from the beginning (opt-in) to the end (sale).

    -g
    Signature
    Screw You, NameCheap!
    $1 Off NameSilo Domain Coupons:
    SAVEABUCKDOMAINS & DOLLARDOMAINSAVINGS
    {{ DiscussionBoard.errors[7281351].message }}
    • Profile picture of the author WillR
      Originally Posted by GarrieWilson View Post

      You should also consider that "A" may have more opt-ins than "B" but make less money. A true test would track from the beginning (opt-in) to the end (sale).

      -g
      Spot on. ROI (return on investment) is MUCH MUCH more important than optin rates.

      You will often find the higher your optin rate the lower quality those leads are. The more information you put on your squeeze page, the lower your optin rates tend to be BUT those leads are much better quality and are more likely to spend more money with you.

      So forget about optin rates and all of that jazz because it's not connected to your ROI. They are two separate things and at the end of the day, all that really matters is how much money you are making.

      Let that be your focus.
      Signature

      .


      .

      {{ DiscussionBoard.errors[7281629].message }}
  • Profile picture of the author Super Daddy Carter
    It depends on the traffic source and how targeted your visitors are when they see your offer.

    With targeted PPC traffic 400 clicks is enough to decide if an offer will convert or not.
    {{ DiscussionBoard.errors[7281906].message }}
  • Profile picture of the author MrPete2000
    I have a complete tracking system setup so I do track the total visits, clicks, opt-ins and sales.

    It's the learning curve that's killing me. I've been out of IM for almost 5 years and trying to jump in head first is making me realize that not only have I lost a few of my old tricks, but there is a lot of new things to pick up along the way. :-)

    Cheers,
    Pete
    Signature

    ...

    {{ DiscussionBoard.errors[7289616].message }}
    • Profile picture of the author Alexa Smith
      Banned
      Originally Posted by zivkovicp View Post

      It's the learning curve that's killing me.
      Yes, indeed.

      I'm as guilty of this as anyone here, but people posting in forums, giving advice, tend to overlook that aspect of it, which is pretty significant and tricky to negotiate when you're the one traversing it.

      But clearly, by definition, the answer to this question can only be a statistical function and inversely proportional to the conversion-rate, as explained in post #10 above. Anyone trying to pass off an "absolute figure" as an answer, without taking that into account, can only be "pulling it out of the air"!
      {{ DiscussionBoard.errors[7290085].message }}
  • Profile picture of the author Mike Grant
    1000 minimum per aspect being tested.
    Signature
    {{ DiscussionBoard.errors[7290088].message }}
    • Profile picture of the author Mark Andrews
      Banned
      An absolute minimum of 1,000 is correct yes.

      Smoking hot,


      Mark Andrews
      {{ DiscussionBoard.errors[7290210].message }}

Trending Topics