The Guaranteed Way to Radically Improve Your Copywriting

image of dart hitting the bullseye

Good copywriters follow well-established best practices, studying sites like this one to learn more techniques that pull a reader in and motivate action.

Great copywriters delve into the psychology of persuasion, and make sure their copy is presented with effective design to make it even more engaging.

But the best copywriters don’t stop there. They know that those are just the factors that get you started.

The best copywriters continue to radically improve their copy over time. They know how to create copy that works even when it defies common sense. They find the “magic words” in headlines that will never be revealed in any swipe file.

How do they do that?

It’s not a secret technique or some special knowledge they learn from a guru at the top of the mountain. The truth is, great copywriters use split testing to radically improve their copy over time.

You may have heard of split testing, but unfortunately not a lot of us actually understand it properly or use it for conversion rate optimization. (Kind of the way everyone’s a social media guru whether or not they actually have a clue about how social media really works.)

So what the heck is split testing?

Split testing is not a buzzword and it’s certainly not a fad.

It’s very popular among large Internet companies (Amazon, Groupon, Google, etc.). In fact, almost all of them conduct thousands (or even tens of thousands) of split tests to increase sales and conversion rates on their websites and landing pages.

They have been doing split testing for years because it is a simple, methodical way to increase online conversion rate.

The concept of split testing is quite simple:

In a split test, you create one or more variations of a landing page and then randomly split traffic amongst them. Meanwhile you measure which variation led to maximum conversion for the result that you were looking for. Adopt the better performing landing page. Rinse and repeat.

For example, if you have a landing page that sells a parrot training eBook, how do you decide which headline to use? Should you use Teach your parrot nifty tricks in 24 hours (version A)? Or should you use Your parrot speaks 20 words by tomorrow or your money back (version B)?

You could ask on twitter, and get hundreds of very knowledgeable guesses. But that’s all they would be.

To find out, you would simply create two versions of landing page using one of those headlines for each. Send 50% of your traffic to version A and 50% to version B, and observe which version converted better. (Since there are usually two versions involved, split testing is also known as A/B testing.)

What should I split test on a landing page?

The easy answer to an easy question: everything! You never know what might be most influential to your sales and conversions so you should split test all aspects of your landing page copy.

Asking what to split test isn’t really the right question, then.

The right question is: What should I test first?

You have to prioritize the most influential aspects of a landing page, the elements that tend to produce the biggest changes.

I am the founder of a split testing product called Visual Website Optimizer with thousands of users, which gave me the privilege of observing thousands of split tests and their results from a front-row seat.

That experience showed me some of the key elements (in order of importance) that you should test on your landing page:

  • Headline (if you can test only one thing on your page, test your headline first)
  • Call to action (should you have Buy Now or Add to Cart … or offer a free trial instead?)
  • The product’s positioning
  • Highlighting different benefits and features
  • Length of copy (surprisingly, sometimes removing information on a page increases conversions)
  • Video vs. text (for some sites text works better, for some video prompts action)
  • Screenshots and other images (I can’t stress enough how important is it to split test different kinds of visuals you use on a landing page)
  • Freebies (visitors may respond better to one kind of bonus or free offer when compared to others)

Let’s play a guessing game

If you are still skeptical about benefits of split testing, let’s play a small guessing game.

Following are some real world split testing case studies. Read the examples and then try to predict which version worked best. The answers are below — but try to resist the urge to peek.

Case Study #1
For a PPC landing page meant for list building, which of the following call-to-action buttons increased signups by 28%?

  1. Version A: “Watch the video”
  2. Version B: “Get Instant Access”

Case Study #2
For a product meant for local business owners called CityCliq, which of the following headlines increased conversions by 90%?

  1. Version A: “Businesses grow faster online!”
  2. Version B: “Online advertising that works!”
  3. Version C: “Get found faster!”
  4. Version D: “Create a webpage for your business”

Case Study #3
In another PPC landing page, which call-to-action link increased conversions to next step in funnel by 14%?

  1. Version A: “Read Sample of Book”
  2. Variation B: “Watch Video Preview”
  3. Variation C: “Watch my #1 Abs Exercise On Video”

Proof that you should be split testing your landing page now

Got your picks for each of the case studies above? Ok, now it’s time to reveal the answers:

  • Case study #1: The “Watch the video” button converted 28% more visitors
  • Case study #2: The “Create a webpage for your business” headline increased conversions by 90%
  • Case study #3: The “Watch video preview” call to action improved conversions by 14%

How many out of three did you get right? Unless you have already read Visual Website Optimizer case studies (or Googled them) it’d be pretty hard for you to have guessed all three of them correctly. In fact, if you guess randomly there is only 4% chance that you will get all three of them correct.

Logic won’t get you to the best-converting copy. Knowledge of copywriting techniques will give you good ideas to test, but it can’t take you all the way there. Even experience won’t get it for you.

But split testing will.

About the Author: Paras Chopra is the founder of Visual Website Optimizer, the world’s easiest A/B split and multivariate testing tool. Used by 5000+ companies worldwide, it allows marketers and designers to create A/B split tests and put them live on their websites in less than 10 minutes. We offer a free 30-day trial account.

Print Friendly

What do you want to learn?

Click to get a free course and resources about:

Reader Comments (74)

  1. says

    Paras :

    Thanks for the info on split testing. I now that Sonia once mentioned in a comment that copywriters test headings, offers, landing pages and sometimes body copy.

    I’m happy that Copyblogger is continuing the posting thread on copywriting.


  2. says

    Awesome read!

    I love the way copywriting is scientific and methodical. In other words, anyone who’s interested can learn it. I just hate to think talent is everything. :-)


    • says

      Talent is about 90% a function of work. :)

      The online world is wonderful because you can test quickly for very little money. Back when direct mail and print advertising were the major players, you had to invest a bunch of money (and time) before you knew your test results.

      But that,of course, means that the game will belong to those who use the tools.

  3. says

    I wonder if split testing on blog posts are viable. Sounds silly, I know, perhaps post two different version, two different times? Make sense?

    I has\d 2/3 by the way.

  4. says

    Great post – thank you, Paras. And that’s a great looking tool that you’ve developed. I’m going to take some time to look at it in detail – split testing is an area that our clients ask about frequently, and this could be just the thing for them.

  5. says

    I was wrong in all three. I forgot that readers are not logical. Or that they do not think like me. Testing will not make the best sellers, but neither do guessing. We better test therefore.

    Thank you for sharing your research findings.

    Jef Menguin

  6. says

    Oh man Paras. There’s actually something to this. I got all three of them wrong, and I was confident with two of them. Stink! Guess at the end of the day, you need to do lots of split testing.

    But how do you control what percentage of your audience goes to each landing page?

          • says

            Martyn, you set up the two versions you want to test – make two separate pages – then the software alternates each option.

            First visitor gets option A, second visitor gets option B, then third visitor gets A, fourth visitor gets B, etc. etc. Whichever version gets more clicks, orders, (or whatever you’re looking for) wins.

            Does that help?

          • says

            Yeah thanks Jodi, I think I’m beginning to get it figured out. :)

            After kicking around on the site some more, I read that the app will send repeat visitors to the same landing page, so they don’t know anything’s going on. Guess it’s scraping IP addresses and stashing them into a temporary database or something. Pretty cool.

            The more I learn about this, the more I like it. Definitely start using this when I get into landing pages *and* have more than two visitors per day. Cough.

      • says

        Which leaves me wondering, maybe some of what we think of as concrete rules for copywriting isn’t as static as I thought. You know, Useful, Unique, Urgent, and Ultra-specific. Paras’ tests show that maybe Ultra-specific isn’t necessarily a good thing. Maybe Unique isn’t either.

        But these are deep waters, and I’m in over my head. :-)

        • says

          The rules help guide you toward having better odds of success. But sometimes the copy that actually tests bests defies the rules. That’s what keeps things interesting. :)

  7. Melisssa Breker says

    Great post! I agree the only way to get a definitive answer on the right words is to test the waters. Often, people think they know the answers regarding the right copy … but testing shows the actual results. Thanks for an interesting read.

  8. says

    I have 2 main blog in 2 very different niches (weight loss and blogging) and the one thing that I have noticed is that something that works well for the weight loss niche, doesn’t always work for the blogging niche.

    Great article!

    Brian M Connole
    HCG Diet 411

  9. says

    Great information, I’ve been doing some split testing recently and I agree that’s definitely very important.

    I know you mentioned images but beyond just the images I think overall page design like header graphics, ecovers, text format and things like that are important to test as well – I got that tip from a friend of mine who used to work for Matt Bacak.

  10. says

    This article confirms for me the nagging question… do I say what I think my potential clients are “thinking” or do I simply state what in their minds will get their attention? Many times what we think will get their attention is not a all a good play on words but instead, something simply stated!

    This case study post has once again proved me in error. Thank you for sharing.

    Needless to say… I got all three wrong too.



  11. says

    Great tip Brian!

    And uh… just WHICH version of the title on this post got the best mileage? Was it the “Guaranteed…”? or the “…Radical”? or some other variety of carrot entirely? 😉

  12. says

    Reading the comments, I don’t feel so bad about getting all three wrong. Still, I’m baffled by the results. I leaned towards the catchier call to actions, like ‘Get instant access’, ‘Online advertising that works’ and ‘Watch my #1 Abs Exercise On Video’.

    I’m assuming the others converted better because they’re slightly more direct? Either way, thanks for sharing and I’ll check your product out. Split-testing isn’t something that I normally dab in and, as nagging as it is, I know I should.

  13. says

    Enjoyable post – thanks for this. It makes sense that its focus is on content, given that this is a blog about copywriting, but there are other elements to landing page optimization to consider, as well. These include design elements like color palette and images, as well as page architecture. I’ve designed A/B landing page tests where something as simple as a different image provided more impact than content. Not to say that testing your headlines, calls-to-action and content isn’t critical; if you get any of these wrong the page will fail, as Paras does a great job of showing us. It’s best to consider content as one [big] piece to the puzzle, but not the only piece.

  14. says

    I’m a sucker for clever titles (truly clever not hype!!! titles). I’m trying to curb my clever addiction– from now on I’ll be direct. (Well, maybe).

    Thanks for the insight.

  15. says

    Hey! There should be two darts in that picture above! 😉

    I got all three right without looking at the case studies. Thank you copyblogger!

  16. Lisa Champagne says

    I would’ve predicted that the benefits-based headlines (IMHO those are B,C, C) would be more appealing to customers. They seem to follow the common advice to highlight benefits to the user. In these cases, the ones that drew best seemed a bit more passive.

    Perhaps this approach is less ‘threatening,’ and therefore removes barriers to conversion? I get that this topic is ‘testing is king.’ I also like to think it through to see if there is some learning that can help me to write more effectively.

    Valuable info, thanks!

  17. Elliott F. says

    my main curiosity here is, maybe there is some other factor that we’re not seeing? is it maybe the time of day that the visitor came across the call to action? maybe a geographical location (usually linked to time as well)?

    also i’m curious *why* some of those calls converted better than the others.

    • says

      A true A/B split will divide the traffic equally, so you can get as true a test as possible, which smooths out the effects you’re talking about. Unless you’re getting two clicks a day, you should be able to get a pretty good random split.

      Every copywriter is curious about that, and sometimes the answer is “just because.” We can always come up with a reason, but we have to be careful about that — our “reason” can prompt us to skip testing next time because we think we know the underlying principle.

  18. says

    Great article, Paras! I have retweeted it to my list.

    Shameless self promotion:

    Dear CopyBlogger readers, if you want to quickly try split testing for yourself using the free Website Optimizer tool by Google, please read my article:
    How to Start Your First Split Test in 5 Minutes

    Yes, this literally takes 5 minutes.

    You should still try Paras’ program, as well. I hear a lot of companies use it with great success. It probably has its unique features compared to GWO, I haven’t checked it out yet.

  19. Nate says

    Very interesting read. I’ve been pushing my company to do more split testing, and we’re slowly heading in the right direction.

    I missed all 3 on the guessing game, but it’s difficult without more information. For example, what service(s) does CityCliq provide? Web design, marketing, SEO? That would affect which CTA worked best.

    Split testing is the perfect way to increase your conversion rate. Start with your best ideas, what you know about your customers and everything you’ve learned from Copyblogger. Then keep testing and improving each element.

    Great article, Paras. Thank you for sharing.

  20. says

    Split testing. We’ve been working on this for a long period of time and yet this is the first time i fully understood what it meant and why it should be done. Anyway, I tried to answer those tricky case studies and got 2 out of three. Looking forward for more of these.

  21. says

    I’ve heard of split testing before but I’ve never really understood the importance of it, how it would affect my business, or how to use it. Paras has done a great job at explaining all three and Brian added an extra comment that was extremely helpful explaining point three. Thanks for the helpful content.

  22. says

    Thanks for the insight to Split Testing.

    Just goes to show that it’s not what/how you (the product/business/content owner) think, or what excites you, but what your target audience think.

    I chose differently on all the case study options (being a logical-hung-up-on-choice-of-words kinda person) but it’s the majority that gets the vote and brings in the business!

    Obviously, in practice a number of parameters will apply e.g. you should have a decent source of data to start off with – quality and quantity.

  23. says

    What a great split test walk through Paras. Never realised the concept could be so profitable with a few tweeks and measurements. It makes us the viewer in effect and that power position is worth real leverage. Great, simply written article with lots of value. Cheers my friend.

  24. says

    At times it becomes tough for a website owner to know what title works best for the content on the landing page. To avoid the dilemma, split testing technique is what should be opted. Otherwise a website owner may end up not getting a good conversion rate.

  25. says

    Good post. It is always about split testing. That is the science of advertising, specificially direct response. Why do you think David Ogilvy implored copywriters to read Scientific Advertising seven times?

    Two points though.

    1) If you’re going to test two different headlines for a landing page to sell an ebook, surely the headlines decide how much traffic they get, I am assuming the rest of the copy is the same? you can’t devide traffic, thats the whole point of testing. If the copy is diffferent then I agree, it is a conversion test.

    2) To my shame, I got one answer right in the test! And I thought I was good at guessing headline results, shows what I know!

    Here are my answers 1) B 2) C 3) B

    Back to the drawing board for me then.,,

  26. says

    Tremendous piece! Split testing isn’t always a necessity, but it is the best way to maximize any page’s potential. Plus, once you’ve learned what does and doesn’t work, you’re better-prepared when you’re making pages for new products or services! Definitely bookmarking this article! It is nothing short of a step-by-step guide to split-testing for beginners… and as such, it will be a great reference!

    Thanks for sharing!

  27. says

    Great article. I knew nothing about split testing. Although I am a copywriter I also know nothing about the rules of copy either.

    Your examples all increase conversions because of common sense. It’s not random, as you suggest. It’s really very simple.

    I am not just stating this having seen the answers either.

    “watch the video” tells you what the next step is. “get instant access doesn’t”

    “create a webpage for your business” tells you what you are doing next. The other lines try and give you a benefit,

    “watch video” tells you what you will be doing next. “read sample of book” is too vague. How much book? Do I have to pay? Etc. The abs one is just silly.

    I totally see the value of split testing. Put several strong lines into the mix though.

    In my opinion, copywriting is not scientific in the slightest. There is no formula to good writing. I’m not saying I’m a good writer. I’m just saying I find formulas and rules depressing. Really depressing. Write how you speak. The second you deviate from your natural train of thought to think of rules or logic you are completely fucked.



    • says

      Correct, there are always set of best practices but they are not guaranteed to work in every case. It’s better to split test your ideas to really make sure you have the best copy.

  28. says

    Wow! I read your fantastic article from top to bottom and didn’t skip a word.

    Split testing is a new topic for me and a real eye-opener as I am a Web Writer with more and more demand from clients to optimize pages for SEO.

    I’m totally convinced.

    I thought I had all the answers to the “easy” test right, and “flunked” by only getting one correct answer out of three.

    You proved your point! And I understand what spit testing is.

    Kudos and thanks!

  29. says

    Great post about A/B split testing! Variations of sites is key seeing what converts best. Plus its pretty easy to split test sites no a days used to be pain in the neck years ago.

  30. Amber says

    This does work. Today we just quit spending money on several keywords because we found that while they were getting several clicks, we weren’t making money off of the leads from these keywords.

    Trial and error really does play a huge role in these campaigns. Our next step will be to improve parts of our landing pages and see if that increases our conversions.

  31. says

    Hey Paras,

    I am regular reader of your blog at Visual Website Optimizer website…glad to see you here.

    Any action on web, specially COPY is useless without TESTING.

    Very few people (even good copywriters) understand this, i am glad that you have taken time to write about it.


Comments are open for seven days. This article's comments are now closed.