Headline Formulas and the Science of CRO Copywriting

Copy Hackers Blog Post Science CopywritingCopywriting is NOT an art.

Sorry, let me rephrase: Effective copywriting is not an art.

It’s a science, first and foremost.

As John Caples teaches in Tested Advertising Methods (only $9.64!), copywriting should be approached the same way an engineer approaches the building of a bridge.

The same way a surgeon approaches triple bypass surgery.

The same way Heisenberg/Mr. White approaches his pure blue meth.

Once you know what to do, do it.

Do it exactly the same way every time.

Lather, rinse, repeat.

Follow the steps.

Stop wordsmithing.

Stop trying to impress yourself.

Just stop it already!

Do yourself and your site visitors a MASSIVE favor, and start “scientific copywriting”. That means taking a serious hiatus from being “creative”. That means looking at your website like a lab, not an easel.

If converting people with your copy is important to you.

Allow me to explain…

(Listen to this post on the go – with SoundGecko!)

Follow the Exact Same Process for Writing Every Single Time

Let’s get this straight: There is no muse on the top of the mountain.

You do not have to wait for your spirit to be moved by the words.

You just need to follow a proven process. The way a scientist follows the same process when setting up an experiment.

Here is the process I follow to write conversion-rate-optimized (CRO) copy:

Joanna Wiebe Copywriting Process Copy Hackers

When you follow this process EVERY time, you never need to worry about whether you “like” something or not. (See my post about this)

You throw subjectivity out the window.

Remember: This is not an art. It is a science. Be objective.

Now let’s talk about the most obvious and necessary way to be a “scientific copywriter”…

Use Headline Formulas

Why would  you write every headline from scratch when there are literally 100s of headline formulas to choose from?

And why, after finding a few headline formulas that work for your market and are easy enough for you to write, would you cease to use those few formulas?

Are you a masochist?

Do you have nothing but time on your hands?

Are you trying to set a record for number of headlines composed and trashed?

If you’re unfamiliar with headline formulas, go here, here and here to find a bunch – and check out some easy-to-complete ones below:

The Only Way to [Do Something Desirable] Without [Doing Something Undesirable]
The Only Way to Turn Off the Lights Without Clapping or Getting Out of Bed

[Do Something Hard] in [Period of Time] or [Promise]
Tune Your Piano in 15 Minutes or “Piano Tuner App” Is Free

[Do Something Desirable] Like [an Expert] Without [Something Expected & Undesirable]
Learn to Play Chess Like Bobby Fischer – Without Any of the Crazy!

Notice how the first and the third formulas are similar?

You could use both formulas as headlines on your site, and only the very observant visitor would notice that they sound similar – and he probably wouldn’t care! The first could be on your home page (since it’s essentially how to write a value prop), and the third could go on your “How It Works” or “Tour” page.

There would be no need to go searching for a new formula!

You do NOT have to reinvent the wheel every time.

So repeat after me: “I will not try to dream up new headlines when I can use proven headline formulas”. (Yep, that’s a tweetable!)

Next up……..

Be Direct and Specific

Scientists live in a world of facts, evidence, proof, precision and specifics.

When you’re copy hacking, you should, too.

  • Never ever ever ever ever EVER summarize when you can be SPECIFIC
  • Wherever possible, let the numbers tell the story – we’re talking “Join 46,700 Others” and “Save $3500 in 16 Weeks”
  • If you make a claim, immediately support it with some sort of proof or evidence
  • Use social proof, like testimonials, to make your points for you

FreshBooks uses very specific data in a direct, memorable way:

Instead of using some sort of summarized “Learn More” button, we at use this precise, specific button on our home page: uses social proof in a very bold, impossible to miss way:

PopSurvey Social Proof

Speaking of social proof, let’s get into a little Cialdini with the next point…

Apply What You Know About the Science of Persuasion & Influence

Cialdini Influence for Kindle EbookHere are the 6 persuasion principles Cialdini discusses in the must-read Influence:

  1. Reciprocity
  2. Commitment / Consistency
  3. Social Proof
  4. Authority
  5. Liking / Likability
  6. Scarcity

Read the book (Kindle version right here!) to learn the ins and outs, all of which are based on academic studies. Muchos recommendedos.

In addition to the proven ways to persuade that Cialdini discusses, a handful of scientific ways to persuade with your copy – and design – include:

  • Loss Aversion
    The benefit of moving from free to paid may be “Don’t lose the data you’ve entered” rather than “Save your data”.
  • Diagnosis Bias
    Use a testimonial from an authority as your home page headline to improve initial value judgments.
  • Contrast / Context Effect
    Make big prices look smaller by comparing them to even larger prices.
  • Pygmalion Effect
    During support calls, tell customers they’re going to have an easy time going forward, and they’re more likely to.
  • Conformity and Herd Behavior
    Make any decision other than using your product sound like it’s outside the norm.

Lance and I discuss quite a few of these on our ancient (but never outdated) 30 Days of Persuasion blog, and you can learn more by reading books like Sway, Habit and The Paradox of Choice or by taking HFI’s Persuasion, Emotion, Trust (PET) Course, which I took a few years ago and really liked.

Read the FREE Copyhackers persuasion ebook >>

Split-Test, Obviously

You’ve followed a sound process, used formulas, used specifics and applied decision-making principles to the copy you wrote.

Now you can simply roll out copy that is low risk / highly likely to perform well (because it is based on triangulated data)… or you can split-test everything possible.

This works best when you’re doing a redesign of a site rather than launching a new site. For example, I’m currently working with a client on revising their entire site in increments, with a new page tested every week. (Their traffic can bear it, lucky devils.)

When you split-test, be scientific about it – this is no time to start getting sloppy:

  • Have a sound hypothesis or research question for every treatment/recipe other than the control
  • Worry less about “winning” (or proving yourself right) and more about “learning” (which you can actually do something with)
  • Run to statistical significance, even if it feels like a royal pain
  • Don’t introduce new variables while running the test
  • Be aware of the impact of factors beyond your control, like seasonality and promotions, impacting the results
  • Use a proper testing tool, like the ultra-easy VWO (which now includes stuff like click-tracking, too)

IMPORTANT: When the test is complete, take the time to document what you’ve done and learned, as I’m about to discuss in this final point…

Head on over to for more about testing & optimization >>

Measure and Record Everything So You Can Learn

Let’s say you tested Headline Formula A against Headline Formula B.

You learned that Headline Formula B beat the control.

So what?

Do you just push the test to 100% of traffic and sit back, counting your cash?

Or do you take 20 minutes to WRITE DOWN what happened in the test so you can learn from it and, with any luck, repeat your success? You should note:

  • The hypothesis (pre-test) and the key learning (post-test)
  • The creative / recipes tested
  • Which test won
  • The people the test was exposed to (e.g., “Landing page for PPC Ad Group B”)
  • The start date and end date
  • The % significance reached
  • Any anomalies or unusual test occurrences, like having to stop the test for a day for some reason
  • Takeaways and recommendations (e.g., “Repeat test next on landing page for PPC Ad Group C”)

I know that seems like a lot of work, so I’ve gone ahead and made the worksheet I use available to you. You’re welcome, young copy hacker.

Once you’ve mastered the science of copywriting, then and only then you can start THINKING about the art of messaging to sell.

But I don’t know why you would.

After all, scientific copywriting works so well, is measurable and is repeatable, I can’t imagine why you’d want to get creative or artistic. Just take up a watercolor class at the local YMCA, and think like a scientist when you’re copywriting.


 Tweetables to Impress Your Followers (or Fill Your Buffer)

Tweet this: “Copywriting is not an art. It is a science. Be objective.”
Tweet this: Are you trying to set a record for number of headlines composed and trashed?
Tweet this: “I will not try to dream up new headlines when I can use proven headline formulas.”
Tweet this: Never ever ever ever ever EVER summarize when you can be SPECIFIC.
Tweet this: Persuade with loss aversion, diagnosis bias, context effect, pygmalion effect…

About the author

Joanna Wiebe

Joanna Wiebe - Copywriter and author of "Copyhackers"

  • Tim Brown

    Definitely kinda throws a barb at the end of this. Which got me to comment… kinda hurts to think of copywriting as not an art, If I’m honest.

  • Pedro Belfort


  • Tonne of value.. thanks!

  • Hey Joanna I loved reading this article, especially “The Only Way to [Do Something Desirable] Without [Doing Something Undesirable];. In my professional career, this formula really helped me a lot, though I have used it in a little variation. Thanks for such an informative post.

  • Jonathan Ginter

    Loved this article, but it has clearly been neglected. Some of the links no longer work. Some of the embedded graphics are gone, so it reads oddly when you reach a “…like this:” piece of text and there is nothing but blank space where a graphic would have normally been found. Just a bit of love and affection would keep this article relevant.

  • Aaron

    I know this is a really late comment, but I just want to reinforce that writing a homepage headline is different than writing copy for a brochure. The article dissenters should remember that a search for relevant information brought the prospect to the landing page. That’s why clarity (based on research and using the prospect’s language), rather than creativity or cleverness is required. No attention grabbing headlines required.

  • Timi

    Learning 1000 formulas doesn’t make you a copywriter any more than learning the theory of relativity makes me a physicist. Please stop encouraging people to be hacks in what should be a creative (even scientific) and skilled profession.

    Copywriting is (categorically) not an art? There is no art to copywriting?

    I feel sorry for all the poor people who don’t know how the craft of copywriting works and only rely on your advice. This is a recipe for stale, simplistic and amateurish copy.

    Apples “Crazy Ones” campaign (look it up) was just following a formula, was it? It neither makes a promise nor does it even talk about the product, yet it was one of the most successful of all time. That’s because it was a piece of creative (not scientific) genius, and often, the most successful copywriting is. I could give more examples but I’m pressed for time.

    Of course there is a “science” to copywriting but that is far outweighed by the creative (and even artistic) elements. You claim people simply need to learn a formula (or a few) and stick to them, and that will work. As a professional (and successful) copywriter, I can tell all your readers that you’re plain wrong.

    That may work for a short while, in very limited situations, but it will not work long-term. You know why? Because copywriting is about connecting your product/service to a positive emotion in your target audience. But people change, societies change – as does the way they feel about things. Methods, formulas and situations that conjured up positive emotions just 10 years ago, don’t anymore. Advertising is evolutionary – and that’s why copywriting is an art. The art of copywriting is what allows you to evolve creatively with the times (not the science of it). Not to mention, the formula approach to copy generally only works (if it does) for direct response copywriting (e.g. email marketing). It does not work for brand awareness campaigns and the like. You need genuinely emotive and unique copy for that.

    And that’s the biggest flaw of your argument – if all you do is learn formulas, your copy will never be unique, never stand out and that means your words will not be read.

    You sound like a marketer who decided to “dabble” in copywriting. If that’s the case, don’t position yourself as a copywriting authority when what you really are is a marketing professional. If you substituted every instance of the word “copywriting” with “marketing”, then this article would make sense. In fact, it would be good advice (and that’s not me being snooty, I really think it would be accurate) but not for copywriting.

    • Lance Jones

      Thanks for commenting, Timi — we appreciate all comments… even those from passionate dissenters. 🙂

      To your final point, Joanna and I teach copywriting for marketers (really, is there any other kind?).

      Interestingly, we recently applied a value-proposition-as-headline formula to 11 startups’ home pages in a series of A/B tests. Know what happened? We generated positive lift in 9 of 11 cases, for an average conversion increase of 34%. It doesn’t get much more scientific than that, and we’ve even documented our entire process and results in our most recent ebook (we hope you’ll check it out!).

      We also don’t believe that using formulas and getting creative are mutually exclusive. In fact, we “got creative” in a couple of those A/B tests I mentioned, and it worked out nicely for those businesses.

      The problem is, based on our experience, startups tend to go immediately to writing “clever” copy when they think of “creative”, which tends not to work well on the web (and the web is where we focus our efforts). I wrote about it here:

      Apple? Always a great example of creativity, but they can afford to pay $1200/hr for those “creatives”. For more than a decade they’ve been one of the most recognizable brands in the world — headed by the most iconic business leader EVER. Getting creative when you’re Apple is fine, but not when you’re a startup with no brand awareness. For those small business leaders, it’s better to stick with what works for headlines — and that’s formulas.


      • Timi

        Hi Lance

        Your response strikes a much more delicate balance concerning the art and science of copywriting – and I agree with pretty much everything you said. I think the article could have struck that balance a bit better.

        It seems like in order to attract attention, the article made big bold statements that, actually, are not truly reflective of the full nature of copywriting and how it works. However, I take your point. For people who are just starting out, it’s essential to stick to the basics, learn the “craft” of copywriting and then move on to breaking the rules/ learning the “art”.

        As for Apple being a big brand, that’s also a fair point. However, the counterpoint is that not all their campaigns are successful. See the “Switch” campaign or further back (while they were still popular), the “Lisa” campaign. Those were not successes by any stretch of the imagination. My point is, creativity is an essential part of the best copy. It’s not some add-on that you only turn to if you’re a glutton for punishment or an idiot (as the article suggests). But when you’re just starting out (and this was not emphasized enough), definitely keep it simple.

        Because I’m passionate about copywriting, I may have overreacted in my first post (although not completely undeservedly so). I would like to say that I like the point about metrics and not just wading in the dark. It’s a bad habit in marketing that needs to go away. Researching your target audience is much better than going by what you like.

        Thanks for your response and for not being as abrasive as I was 🙂

  • Great informative post. There’s definitely a formula behind many creative pursuits. Recall the intro/arguments/conclusion structure school drills us to write essays. Even with something as freewheeling as human interactions, for guy’s pickup manuals, they have methods for escalating, etc. It’s all about internalizing the best practices through rinse and repeat until the rules are so deeply engrained, you can go outside the box and break them. That’s the sign you reached the top of your game.

  • Jacob

    Joanna, this was a great post. I’ve been signed up for awhile now and have been (admittedly) very bad about consistently reading your weekly bundle of awesomeness. This post has convinced me to schedule time every Tuesday to read your email and blog.

    Regarding this post, could you please elaborate on your CRO copy process — specifically #1: Research? I realize what each of those items are, but I get the sense that there are likely some hidden secrets to doing each well.

    Thank you for inspiring and keeping your all-new site simple!

    • Joanna

      New name for my newsletter: “Weekly Bundle of Awesomeness”. 🙂

      Regarding my process, research is the biggest part of it. I come from (or maybe founded?) that copywriter school of thought where we believe that the messages DO NOT come from our own heads. Not at all. Messages come from existing customers, lost customers, prospects and general people of the world. Which means messages come from research.

      I do loads of research when I write copy. LOADS. That always includes, at minimum: 1) phone interviews with customers, 2), 3) competitor content audits, including blog comment audits, and 4) some type of survey, depending on the copywriting task. Additionally, unorthodox research, like mining Amazon book reviews and forum posts, can go a long way. But surveys and interviews usually reveal ~70% of the messages I’ll end up using. What I’m looking for is exact phrasing, interesting analogies, repeated adjectives, and insights into what people want, what they lack, what they loathe / worry about, what’s going on in their lives at the time that they decide to make a purchasing decision, etc. I can’t give away all the tricks of my trade, but does that help? (Sorry I’m responding late to this comment!)

  • Johan

    I think anything can be split-tested — you just gotta be a little creative *pa-dum-pssh* 😉

    All kidding aside, I think that still mostly holds true. For example, when I’m networking, I split-test my “elevator pitch” by telling 10 people one control line, and 3 groups of 10 other people I use variations (1 variation per group of 10).

    The “elevator pitch” that sparks MORE questions and MORE conversation, wins.

    I’ve found a few that “converts’ really, really well.

    I’m not an expert (nor do I play one on TV), but if I had a brochure, I’d make sure there was a call-to-action in it, come up with variations of different kinds (headlines, body copy, CTA, etc), hand ’em out hand-over-fist and see which brochure converts more actions. To measure it, I’d make use of some kind of “code word” or discount for people to use when contacting me (like, “say the word COFFEE when you call, and get a Starbucks gift card”, or whatever — I’m sure you catch my drift).

    Also, great design —one sugestion: here in the comment boxes, the font color is a taaad too light. Not by much, I can still see what I type, but some might need to squint. Or, and this is entirely possible, it’s me, not you.

    • Joanna

      Thanks, Johan. Definitely loads of ways to test! Getting creative goes a long way toward figuring out what will really work.

      I think that, when it comes to testing, we often get caught up in this fear of doing everything in such a way that some university ethics board would approve it as a freakin’ academic study! We worry about the science of it and reaching statistical significance, both of which ARE great and SHOULD be considered IF they can be considered. But let’s not forget about good ol’ fashioned trying something and then seeing what happens. You give some great examples of how you do that. And most people in the free world would think quickly of the way “The Four Hour Work Week” was named ——– it was tested, but not scientifically.

      Great points!

      (Sorry about the font colour —- we’re working on tweaks to the new site design.)

  • Another great post, Joanna. I really struggle with headlines and have a hard time choosing between 4 or 5 that I’ve brainstormed. Split testing is obviously ideal but it only works in some media (i.e. can’t really split test in a brochure), and even then you need some real numbers for it to be meaningful, although you mentioned “directional data” once before, and I still don’t know what that is!

    Loving the new site design by the way.

    • Joanna

      Wow, Damien – you’re a total keener! 🙂 I just posted this, and you’ve already commented. Nicely done. And thanks for the new UI kudos. 🙂

      There’s a sense for a lot of people, it seems, that split-testing is one of those things that you don’t really have to do because people haven’t been “testing” for ages. True, there are lots of forms of media where you can’t accurately split-test —- you mentioned brochures, and I’d add the App Store and email autoresponders (in most cases). This can be super-frustrating. I can see how it’d be off-putting… but it’s worth it when you start doing it. Worth considering?

      (BTW, there are other ways advertisers have tested for decades. But they’re not as precise as A/B testing online.)

      As for directional data, I just mean that you can start to see a pretty consistent trend — a direction that the data is likely to continue down. Yeah? Nah? Bah?

      • Yep, I’m your copystalker! Seriously though … I was just going through my recently reinstalled hard drive to find my Copyhackers bundle and remembered you had mentioned a site redesign was underway, so I dropped over to see how it looked and was rewarded with this post 🙂

        Okay now I get the directional data stuff – and yeah I guess there are other ways to test than collecting massive A/B samples , like …. just asking a punter. Some internet marketers (e.g. Pat Flynn and James Schramko) are using Facebook to test ideas with their followers. They post two versions of a design/logo/whatever and get a bunch of feedback on which they base their decision. Pretty cool way to crowdsource!

Copyhackers Tutorial Tuesdays training calendar