Conversion Rate Optimization and A/B testing

by Alaister

Posted: 10 years ago 14 replies

INTERNET MARKETING

There was a great discussion thread about Growth Hacking, what it means and how warriors can use it to grow their businesses. As a result of that thread the topic of A/B testing and conversion rate optimization came out of it.

I wanted to create this new thread and open up the discussion and see where you guys are at with A/B testing and CRO. Do you guys work on conversion rate optimization and perform constant A/B testing?

At Freelancer.com we are constantly analysing different pages and coming up with hypotheses/ideas on how to increase conversion rates. Some ideas include showing social proof like testimonials/case studies, having different CTA on buttons, using video on sales pages, emphasizing different elements in forms like our post project form and much more.

With each of these ideas we run A/B tests and compare data to see what impact it has on users, their behaviour and our conversion rates.

As I mentioned previously in the other thread we are running hundreds of tests daily on different pages and for different goals.

This is a topic that I love and am pretty passionate about. I'd love to hear about what you guys are doing and am happy to answer questions to help you guys use a/b testing and CRO to grow your businesses.

#a or b #conversion #optimization #rate

michaeloslier 10 years ago

I focus quite a bit on conversion rate optimization. I run an online store and I have things like live chat, testimonials, large product images to increase conversions.

I am in the process of working on my checkout process and making it simpler and easy to use. I think I am getting a lot of shopping cart abandonment and drop off there.
- Thanks
- 1 reply
{{ DiscussionBoard.errors[9187980].message }}
- savidge4 10 years ago
  
  This overall is not talked about as much as it should be. One of the primary issues in discussing this is setting kind of a minimum conversion rate to shoot for. I personally shoot for 3% ( again this is a minimum )
  
  Before it even gets brought up I will say if you don't think 3% is a valid number I can show you case study after case study, white papers, and in general retail reports that say yes 3% is a valid number.
  
  Once I get a product in the 3% area I then will go in and fine tune A/B test to bring those numbers up. With time I have been pretty good at bringing my numbers up in the 10%+ range on most things I promote.
  
  I sell RF modulators on ebay as an example. On a daily basis I sell 10% of the total # of RF modulators that sell. and 16% of those that actually click on my listing. It was a matter of testing the image, and the title to get the clicks, and then testing the images, the best shipping method, the description text, the layout of the text.. all kinds of stuff... test test test
  
  For me personally I have been testing the images on a slider that I have implemented on a store page that I have. I will say I have been more than skeptical with sliders in the past, any time I would put them in place, my store conversion would drop. Now they are more mainstream and so I am trying to work with them a bit more.
  
  Through testing I have found 2 and or 3 images max is optimal On a clothing site I am testing its 2 and on a sports site I have its 3. The duration of the images has a lot to do with converting. Longer is better, right in the 4 second area, seems to work the best with the testing I have done.
  
  Leaving the text the same but changing the image is statistically very effective. an example the words "All Sweat Shirts on Sale" leaving the words the same in 2 images and changing the image of a guy in a sweat shirt and a girl in a sweat shirt. - The girl gets clicked on more by the way.
  
  For me it comes down to this. Just because you are making sales does not mean that it could not be better. There does come a point that you will know that it is as good as it gets and you let it go.
  
  The most important variable in A/B testing is the number of times the change has been seen. If your site gets little traffic, it will take longer to test. First starting with a site I like to test things out with 100 visitors. So I set a bar, let 100 people see it. Make a change, let 100 people see it. again and again and again.
  
  Keep in mind that using 100... it may be days that I get 100 people across my "Test" and that's fine - not efficient, but fine ( you can throw very targeted PPC traffic at it to speed up the process )
  
  Once the site has more traffic, I go back in and test with 1000 views each cycle. If the site progress' further I may test again with 5000 views. Constantly working to get those double digit conversion rates.
  
  I actually have a optin for a halloween page I have that has a 80% conversion rate. kinda scarry ( excuse the pun ) It took over a year of testing to get it there. To be honest I have tried to replicate the page and offer as close as possible that I can in other niches, and it flat out does not work at all.
  
  Anyways that is kinda sorta my process. Yes, there is a lot more to it, but its a general idea!
  
  Hope that Helps!
  
  [ 2 ] Thanks
  
  1 reply
  
  Signature
  
  Success is an ACT not an idea
  
  {{ DiscussionBoard.errors[9191866].message }}
  
  AzzamS 10 years ago
  
  Kind of surprised that it is not talked about more often.
  I am allows doing split testing and CRO whether email or websites.
  Going to be using www.optimizely.com more aggressively
  
  Thanks
  
  Signature
  Download 101 Actions for a Complete Website SEO Technical Audit Sample FREE today and charge clients $$$ with it.
  SEO Case Study: 1.7M Visitors from 27,000 Keywords Click here to read the post .
  
  {{ DiscussionBoard.errors[9191898].message }}
grafx77 10 years ago

I am always split testing everything I do....from sign up forms, to images, videos, and headers. A/B testing can result in a quick bump in conversions as long as adequate traffic is there, however I like to use multivariate split testing for some of my more high traffic sales pages/landing pages.

This will only work with high traffic numbers however and don't recommend it to those who don't have the traffic to back it up as multivariate split testing needs to obtain a lot of data in order for it to show you significant increases. If you have the traffic to send, then your conversion rates will exponentially multiply for every variable you are testing (header, body, colors, positions, videos, CTA buttons, etc)
- Thanks
{{ DiscussionBoard.errors[9192030].message }}
TopLinks 10 years ago

Of course we do A/B testing. And A/B/C/+ testing.

Too many people try an idea, get the results below their expectations, and then stop. Then, they might get results that meet their expectations, possibly even exceeds them, and they stop testing.

Testing can be a word for figuring out what works, but it can also be an exercise to see where trends are going and what works today as opposed to yesterday. If you do your testing and find these trends before they're publicized all over, you're far more likely to find a niche market and be able to profit.
- Thanks
{{ DiscussionBoard.errors[9192146].message }}
ecoverartist 10 years ago

I eat, sleep, breathe and live CRO. I write about it. I practice it. I read all the findings, case studies and white papers.

It is SO vital to a successful website and yet very few people know how to do it (or even know what it is, for that matter!)
- Thanks
Signature
Sherice Jacob - Conversion Expert & Copywriter
BIG BRAND Copywriting WSO! 6+ Years Running!
101 Blog Writing Prompts for Your Business
{{ DiscussionBoard.errors[9192350].message }}
NK 10 years ago

We do plenty of split testing too, and it still amazes me how often changes I thought would do better end up doing worse - such as making the discount label more pronounced with highlighted colors and bold text, but just regular plain text ended up working better.

Also did test runs with up to 4-5 different pages but found that limiting split testing to 2 and a maximum of 3 gives the best data to work with. Anything more just makes the end data harder to analyze since you end up comparing too many different variables to figure out what really works and what doesn't. Then again, I don't really like analyzing TOO many data at once
- Thanks
Signature

Affiliate Link Cloaking & Tracking Software - The most powerful tool available for affiliate marketers who are sick and tired of making way too little money!
{{ DiscussionBoard.errors[9192759].message }}

joaquin112

10 years ago

Wow hundreds of tests every day? I do spend between one and two hours optimizing my squeeze pages, but there's a limit to what one person can do. Has anyone had luck hiring someone to do CRO?

Thanks
1 reply

Signature

Publish your digital course at Accomplisher.com. We create the video sales letter, drive affiliate traffic and split the profits with you. If you want to start making money by teaching online, submit your application here.

{{ DiscussionBoard.errors[9194129].message }}

Alaister

10 years ago

Originally Posted by joaquin112

Wow hundreds of tests every day? I do spend between one and two hours optimizing my squeeze pages, but there's a limit to what one person can do. Has anyone had luck hiring someone to do CRO?

Yeh we have entire teams working on this, so it's not just one person.

You gotta hire someone that lives and breathes it like ecoverartist

I eat, sleep, breathe and live CRO.

Thanks
1 reply

{{ DiscussionBoard.errors[9194144].message }}

Alaister 10 years ago

It seems like we need to have a Conversion Rate Optimization and A/B testing forum here too!
- Thanks
{{ DiscussionBoard.errors[9194146].message }}

joaquin112

10 years ago

I agree, it is very important. By the way, how do you handle giving freelancers access to your website for AB testing? I mean, it must be very sensitive for a company of your size.

Thanks
1 reply

Signature

{{ DiscussionBoard.errors[9194151].message }}

Alaister

10 years ago

Originally Posted by joaquin112

I agree, it is very important. By the way, how do you handle giving freelancers access to your website for AB testing? I mean, it must be very sensitive for a company of your size.

What do you mean exactly?
We have internal teams working on our a/b tests and conversions

Thanks

{{ DiscussionBoard.errors[9194153].message }}

NK 10 years ago

That made me think, do people actually hire others to do their split testing for them? If so, how would you go about handling the access and data?

As far as I know, people usually do the testing on their own or have their own team to do it. But with so many things getting outsourced now, is split testing one of it?
- Thanks
- 1 reply
Signature

Affiliate Link Cloaking & Tracking Software - The most powerful tool available for affiliate marketers who are sick and tired of making way too little money!
{{ DiscussionBoard.errors[9194334].message }}
- SunilTanna 10 years ago
  
  What's the easiest way to grow your internet business?
  
  Get more traffic? (Not really... hard and/or expensive)
  
  Pay for more traffic? (Not really - diminishing returns & diminishing profits: price goes up per click as volume increases, plus keywords become less targeted)
  
  Answer: Turn more visitors into leads/buyers
  
  If you don't optimize your conversion, you are literally leaving money on the table.
  
  Imagine your page is only 80% as good as it could be: You are leaving 20% of sales (and more than 20% of profits, because of fixed overheads) unclaimed!
  
  In reality, most people's pages are no where near 80% as good as they could be: they are lucky if they even 50% as good as they could be! They are losing half their potential profits!
  
  If you don't care or don't care about optimizing, that's one of the main things that is holding you back and stopping you from succeeding.
  
  If you do know, and do care, you can start doing something about it: split testing.
  
  Here's how to do split-testing if you want to self-host, but of course there are hundreds of ways to do it. The point is that you must do something!
  
  PHP Script for Split Testing Sales-Letters,...
  
  Thanks
  
  Signature
  ClickBank Vendor?
  - Protect Your Thank You Pages & Downloads
  - Give Your Affiliates Multiple Landing Pages (Video Demo)
  - Killer Graphics for Your Site
  SPECIAL WSO PRICES FOR WARRIORS + GET THE "CLICKBANK DISCOUNT" TOO!
  
  {{ DiscussionBoard.errors[9194526].message }}

Conversion Rate Optimization and A/B testing

Trending Topics

Newbie

How did you turn it all around?

A healthy fascination for Hollywood...

best high quality crypto traffic.

Any thoughts on golf's latest phenom...Scottie Scheffler??