I’m not passionate about testing per se. I’m passionate about accurate thinking, and making my clients money.

With one exception the list I gave, #7, those results were either from work I did with my own list, or work I did with clients. The numbers we’re talking about are statistically significant. I would hardly be so disingenuous as to post the results of a 100-piece mailing, Jason.

In the case of emails, the list was about 5,000 strong; in the direct mail examples we’re looking at anything from tens of thousands to over a million posted pieces, and ROI measured in the hundreds of thousands of pounds.

So they’re not inconsequential numbers.

I’m a business owner, too. So are all my clients. And my point remains: you don’t know what’s important until you test it. And that means testing as many things as you can.

As an aside: one great thing to test is (practically) getting rid of the sales letter entirely. Nurture your list properly with regular and informative emails and a 360 word sales letter which is nothing more than a call to action can convert at 16.6% (and yes, to statistically significant numbers).

Warmly,

Jon

Pat,

You come to your own conclusions about what works by doing your own work.

Don’t cut corners by assuming anything. Always go the extra mile to verify or disprove the “hunches” you have.

You owe it to yourself to do that.

Hey Jon,

Thanks for your comments. It’s clear you’re passionate about testing.

Many people aren’t. Which is why they don’t do it. My goal isn’t to suggest how to be perfect, it’s to move people one step forward.

Of course setting up tests in GWO is easy peasy, so that clearly isn’t the obstacle holding people back.

Something else is.

Not sure if all the results you cite are yours or not. If they are, that’s great. Testing clearly made a difference for you.

Although citing percentage improvements doesn’t really tell me anything, since a 100% improvement could mean $400 of additional revenue just as easily as it could mean an additional $4 million of revenue.

At the end of the day, do what works for your business. But for the vast majority of people selling online, the idea of “testing everything” is simply overwhelming.

I’m not talking to copywriters here. I’m talking to business owners. They have a lot more on their plate than simply testing stuff all day.

At any rate, I value your comments!

You need to test each one, Pat.

You NEVER know what matters until you do. A good example is video and audio which “everyone knows” works better than plain copy.

Except when it doesn’t.

On MY website, it does; on my wife’s blog when we tested audio and video response bombed.

I repeat: you have to test, and you won’t know what’s important until you’ve tested as many things as you can.

Warmly,

Jon

I’m currently testing three variants with a fancy front end, portfolio, and portfolio with copy. Surprisingly the fancy front end is in last place even though it has all the information of the portfolio and copy – not at all what I expected.

My next step is to test variants of the copy, which this article addresses perfectly.

However I have a question. If I have many sales pages selling similar products – do I need to test each one or is it fair to assume they’d get similar results using the same sales copy formula?

First, the only way to know what matters is to test – you can assume you know what your visitors want, but the only way to know you’re *meeting* that criterion is to test everything, whether “common sense”, your “intuition” or the Tooth Fairy tell you otherwise or not.

Let me share with you some off-the-top-of-my-head results of tests, both online and offline, that have nothing to do with headlines, the offer or the copy:

1. A direct mailing in which white envelopes outpulled green envelopes by 100%.

2. A similar mailing where a live stamp as opposed to a frank increased response by 60%.
3. An email blast where simple html outpulled plain text by 35%.

4. An email blast where having [sender’s name] in the subject line outpulled the same subject line without the name by 35%, again (the email contents was identical). On a different list the results were reversed (almost to the decimal point).

5. An online sales page where the ONLY difference was in how we emphasised certain phrases: underline or bold. Underline outpulled bold by 2:1.

6. An offline ad in a national newspaper where a dashed line around the offer coupon and a larger headline outpulled the original by 30%+.

7. An offline ad where one was in sans-serif and the other in serif… and the serif one outpulled the sans-serif by 100%.

8. The long-copy squeeze page that broke all the “rules” and outpulled a traditional short-copy one with the form above the fold by 12.5%.

I could go on and on like this, but the message is clear: you don’t KNOW what’s going to be important and have the big effect until you actually test, test, and then test some more.

And as for suggesting it’s a “waste of time” is ludicrous. How long does it take to test, say, two different fonts on a web page using Google’s Website Optimizer?

Maybe 10 minutes to set it up?

And then it runs itself.

Sorry, Jason, but your advice is bad and readers would do well to ignore it.

Warmly,

Jon

I agree Thesis is a great move 🙂 Good luck!

David,

Thank you for the comment. Yes, in my experience, having the right offer (which includes a product someone actually wants 🙂 makes EVERYTHING so much easier.

Wyndham,

Testing the headline content, copy lead and offer hardly constitutes everything. But thank you for your comment.