1. This website uses cookies to improve service and provide a tailored user experience. By using this site, you agree to this use. See our Cookie Policy.
    Dismiss Notice

[POLL] Do you split test CTA (buy buttons) on your Amazon affiliate websites?

Discussion in 'Affiliate Programs' started by jamie3000, Jun 8, 2019.

?

Do you split test your Amazon affiliate buy now (CTA) buttons?

  1. Yes - I do split testing on my CTA / buy buttons

    3 vote(s)
    60.0%
  2. No - I don't do any split testing

    2 vote(s)
    40.0%
  1. jamie3000

    jamie3000 Jr. VIP Jr. VIP

    Joined:
    Jun 30, 2014
    Messages:
    3,879
    Likes Received:
    2,160
    Location:
    uk
    Thought it might be interesting to see how many amazon affiliate website owners split test their Buy buttons and other call to actions.

    If you do. If would be interesting if you could share what you do exactly :)
     
  2. Ftr

    Ftr Junior Member

    Joined:
    Feb 23, 2017
    Messages:
    147
    Likes Received:
    165
    Occupation:
    Passive is Massive
    Location:
    Alpha Nomad
    I remember first reading some ideas on best ways to A/B test this on Hubspot, if you search 'Hubspot CTA split test' it gives you a little intro and some basic starting ideas on overall CTA A/B and gives you a tool to measure it & stuff, I just went from there and applied some of my findings onto AA.

    Mostly placement location and colour changes.
     
    • Thanks Thanks x 1
  3. jamie3000

    jamie3000 Jr. VIP Jr. VIP

    Joined:
    Jun 30, 2014
    Messages:
    3,879
    Likes Received:
    2,160
    Location:
    uk
    Did you see much of an increase in your click through to Amazon rates?
     
  4. Ftr

    Ftr Junior Member

    Joined:
    Feb 23, 2017
    Messages:
    147
    Likes Received:
    165
    Occupation:
    Passive is Massive
    Location:
    Alpha Nomad
    You know what I noticed differences, but it was different for each site like there was no exact universal placement that worked best for every site I had. It seemed to depend on the niche and how that site was designed. Mostly I had a pretty good placement from the start but sometimes I found a placement that did increase Amazon ctr a tiny bit or dropped it a fair bit, and it was worth testing out though.

    I was too scared to fuck around too much with any actual amazon buttons though because they were a bit iffy about that and I didn't want to complicate any guidelines.


    But for other CTA's like getting them to click on a review page I noticed over time bigger increases from when I started it was like did a little change found a tiny increase, so kept it, did another little thing that did nothing, did another little thing added a tiny increase, kept going and it all eventually added up, trouble is that it's kinda ever changing.
    I liked to play around with testing colours mostly and 1 or 2 words sometimes even changing something like the word GREAT into AWESOME can make a difference, I don't really understand why, again different for each site.


    My process is have enough current data traffic conversions and have your site all done everything good, then pick one tiny little thing, change it and monitor the data for a week. Keep doing that till you find an increase then you change another little thing somewhere else on the page and try find another increase, rinse repeat till you got a 100% ctr lol.
    I once changed an image from a guy and girl in the photo to just a photo of a beach and it gave a massive increase, I like seasonal changes for cta like christmas themed stuff.

    I already knew things like blue is good for building trust, red good for taking action quicker, have an entire notepad notes on those nuances. Only recently am I getting round to really testing it all out as it takes a long time you gotta give it a week or so for each 1 little change, there's some cool split testing software out there, only change 1 tiny thing at a time though to get the correct data.


    So i'm really into split testing going forward. Also don't just do things you think will work, try out a couple of things you think might not work (Id generally start split testing these on lower income sites, till happy with the knowledge)
     
    • Thanks Thanks x 1
    Last edited: Jun 8, 2019
  5. saku007

    saku007 Newbie

    Joined:
    Jun 5, 2019
    Messages:
    16
    Likes Received:
    2
    I did some split testing when I began a few years ago - not testing actively anymore. There max difference that I saw was about 10%, so I just picked the winner and ran with it since then.
     
    • Thanks Thanks x 1
  6. fasttrak

    fasttrak Jr. VIP Jr. VIP

    Joined:
    Sep 12, 2016
    Messages:
    2,631
    Likes Received:
    2,490
    Gender:
    Male
    Location:
    ↓ LOOK DOWN HERE ↓
    Home Page:
    Also interested to hear others experiences. Good question @jamie3000, thanks for asking.
     
  7. nichexposure

    nichexposure Jr. VIP Jr. VIP

    Joined:
    Jul 4, 2013
    Messages:
    2,227
    Likes Received:
    785
    Gender:
    Male
    Occupation:
    I just pretend to know SEO
    Location:
    2nd Page of Google
    Home Page:
    Not actually with the CTA but changing the whole landing page structures. I am testing multiple landing pages right now. Few works well, few not.
     
    • Thanks Thanks x 1
  8. jamie3000

    jamie3000 Jr. VIP Jr. VIP

    Joined:
    Jun 30, 2014
    Messages:
    3,879
    Likes Received:
    2,160
    Location:
    uk
    Wow. Thanks for such a detailed response. It's good to see people taking it seriously. Most people just want to pump out content and build links and forget to make the most of the traffic they have.

    How much traffic do you have on most sites to test changes with?
     
  9. Ftr

    Ftr Junior Member

    Joined:
    Feb 23, 2017
    Messages:
    147
    Likes Received:
    165
    Occupation:
    Passive is Massive
    Location:
    Alpha Nomad
    It's cool I didn't even think it was very detailed lol.

    I saw your discussion about the selenium undetectable jar and thought the same awhile back, regarding taking it seriously. Well I don't think many people have their eyes on the real prize or even make money outside of a job job.


    1. I focus more on like unique visits to a page over any period of time, personally I would start at min 5K VISITS and at least 200 CONVERSION, but that could have been achieved after 2 and a half months or something.
    2. I use to start at 10K VISITS and 100 CONVERSIONS back when first testing stuff though.
      But you gotta let your test run for a long enough amount of time in full week+ increments.

    First 4 days I'm just looking to not see any big drop, if the page is lower traffic I'm letting it run longer to give it more weight/confidence in my findings.


    But I will change this it ifs higher traffic but low converting or if its lower traffic & high converting, then for higher traffic low con id change something big and sooner.
    For higher converting id not touch it for a longer while and then only make the more little changes like colour of button or something.

    Changing big things like the headline or a layout are better to work with cause you can't test every variable on a page out. I don't sit around spending much time on this though I'm in then I'm done then I'm out, I'll look at results mid week and end of week. A few minutes spent on this, working wise.