Bruce, Jason, and Rick make a solid point.
You know what the most important part of Glyphius is?
Yup, the writer.
People tend to think that Glyphius is a tool to replace good copy. It's not. It's a tool, just like a thesaurus, just like a dictionary, just like a word processor.
The idea is to understand the market, and understand the principles of GREAT copy. Then use Glyphius to validate -- not to replace.
Glyphius is not my favorite tool. It's a cool tool, though. Plus, we have to keep in mind that the stats behind Glyphius are based on winning ads -- such as online ads (like AdWords).
But the context of email (and open rates), as well as salesletters themselves, are different.
Testing is not gospel. People tend to forget the 2nd most important part. Aside from great copy (the copywriter), the 2nd most important (and probably #1, really) is the market.
Markets are different. In fact, I would say that copy is 2nd and context is first. The market, the delivery channel, the time frame, and of course, the copy around the "negative" keyword are ALL important.
When I test, I never say that it's gospel or should be taken as such. It's never statistically significant, unless there's a MASSIVE amount of traffic. The more traffic and the more results you get from testing, the slimmer the margin of error.
Next, test results are, again, contextual. I think the important thing about testing is knowing what to test. If a test result says "this works" or "this doesn't," you shouldn't cross-polinate those results to your situation, but now, at least you have something you can test in your business, too.
When I won a split-test challenge using Glyphius, Glyphius didn't win, it was my copy (my choice of words, my angle/hook, and the connection between my copy and the reader, i.e., the context) that won. I only used Glyphius as a tool "after the fact" to validate, tweak, and come up with variations.
Bottom line, let's understand that there's a difference between
correlation and
causality.
In fact, let me quote an answer I gave to one of my coaching students just recently about a blog post from James.
He asked:
My answer...
Quote:
James may be right as it relates to actual, bottom line, end-of-the-road results. Are headlines DIRECTLY conducive to more sales? The answer, of course, is no.
A headline is the most important part of the salescopy not because it is directly tied to sales. It is, of course, indirectly tied to the bottom line. It isn’t directly producing sales, but it is tied nonetheless.
It’s all about fundamental marketing: AIDA.
The primary objective and purpose of a headline is one thing and one thing only: to get people to start reading the next paragraph. And the next paragraph’s job is to get people to read the 2nd one, etc.
Simply because “nothingness” (no headline) wins in some cases doesn’t mean the headline tested doesn’t work or that it’s safe to say that it’s NOT the most important part of the salesletter. (That’s a correlation, not a cause. And I’ll come back to this in a moment.)
First of all, there are many other variables here that not taken into account:- For one, the first paragraph (in a no-headline letter) can act as a headline itself. Heck, the web page title (in the top browser bar) can be the headline. Who knows?
- The mindset of the reader may be “presold” before hitting the copy (such as coming from an affiliate promotion or keyword campaign).
- If the traffic came from a PPC campaign, the ad (keywords and ad copy) acts like the headline. People read it and want more information. (That’s what people go online for!) So if they hit a salesletter without a headline, they’re tempted to read it anyway.
- Headlines can sometimes scream “salesletter!” and when people see one, they may be pushed to scan or leave the copy.
- Finally but most importantly, they may not be targeted AT ALL. And if they are targeted, a headline may push them away. (Better said, a poor headline will.)
Again, the biggest bottleneck in any copy is almost always the headline. Because if people can read past it, they won’t read the rest. If you don’t get their attention (the “A” in AIDA), then the rest of the formula falls down the drain.)
But removing a headline in some cases may be like removing the bottleneck in the first place.
As Dan Kennedy once said:
“The truth about long copy is that, first of all, there’s abundant, legitimate, statistical research, split-testing research, to indicate that virtually without exception (...) that readership falls off dramatically at 300 words but does not again drop off until 3,000 words.”
Dan Kennedy - Tim Paulson Interview
Dan was talking about long copy in this case. In that, if people are targeted, they will read it. All of it. But if they're not, they won't even get pass the headline.
But the prevailing notion pervades headlines in and of themselves, too. In other words, if people are targeted, and the headline is RIGHT for them (they are targeted and the headline grabs their attention), they will read the rest. If not, they will leave the moment they read the headline.
If the headline is poor (and all other headlines are poor, too), then “nothingness” can win because you are in essence removing the bottleneck -- but not necessarily the cause.
If a really good headline was found, it might win over “nothingness”. And I admit that, in some cases, finding the perfect headline might be a challenge -- so “nothing” can be an obvious solution.
But it can also be the result of being too lazy to come up with better headlines, or not having enough traffic and/or time to test more headlines.
Bottom line?
There is a difference between “causality” and “correlation”.
In plain English, I mean that no headline winning over other headlines may be correlational -- in that, no other headline was good enough to produce a result, so nothing wins. “Nothingness” won because the other headlines it was pitted against were not the right headlines (or they were poor headlines.)
So there is a correlation. But it’s not the cause. In other words, “no headline” is NOT what caused a salesletter to outperform. It’s simply the lack of a bottleneck (a poor headline in the first place) that led to the copy outperforming with “nothingness.”
A good example of cause vs. correlation is, if you drink orange juice every day and you don’t have cancer, that means that drinking orange juice cures cancer. Right?
Obviously, that’s an assumptive leap, and it’s wrong. There may be a correlation there, but it’s not the cause.
And that’s the same thing, here.
No headline winning over other headlines doesn’t mean that the lack of a headline caused the copy to pull more. It may simply be that “no headline”, in relationship to all the other headlines it was tested against, won because the other headlines sucked, or that they weren’t right for that market.
So removing the headline removed the bottleneck, in this case.
But it’s also safe to say that, if you were to test more headlines, there would be one out there that may outperform “nothingness” as well.
My 3 cents.
|
Bottom line, interpreting some test results may be correlational. But it doesn't mean it's causal. Saying that Glyphius says "x" word is a bad WORD is making a leap of assumption that this is the case with your particular copy, your particular market, your particular offer, and your particular delivery channel.
And that's no different than orange juice curing cancer.
Just a thought.