The two main alternatives to the Split Test Accelerator are
- doing a series of simple split tests, and
- using a different accelerated solution.
Let's look first at simple split testing, and . . .
"Why You Don't Want To Look For Needles In A Haystack One Piece Of Straw At A Time" (OR "How To Find
All The Needles In Your Haystack In One Shot.")
If
you want more sales
from your
landing page, you will have test some changes to
your landing page.
For instance, here are some changes that might help your conversion rate:
- a different headline.
- a video.
- putting testimonials at the
top of the page instead of at the bottom.
- leading with a picture
of someone using your product.
- using one of those cool peel
away ads.
- using a drop
cap.
- more order
links.
- Breaking your sales page into multiple
parts.
- Having information run down
the side of the page in addition to the main text.
- framing the price in a
different way.
- Or . . .
Those are all interesting ideas. And some of them
might work. But here's the thing . . . even if you
generate a whole bunch of testing
ideas, you
probably won't know which ideas will make a positive
difference, and
which ones won't.
In fact, most
of your ideas will NOT
be good ideas. That is, they won't lead to any kind of
significant improvement. (This is true even for very good professional
marketers and copywriters).
Charles Holland,
author
of "Breakthrough Business
Results With MVT", argues that out of every
100
ideas people have for
improving business processes, only about 6
of them will be test-worthy
and turn out to make a significant positive difference.
(Here's the breakdown:
Of 100
ideas, you will probably test only about 25, because the
others are too expensive, time-consuming or impractical to test. Of the
25 you test, about 13 won't make any real difference, 6 will HURT, and 6
will lead to significant improvement.)
Now here's the thing. Those 6
ideas that make a positive difference might double or triple your conversion rate!
So you need to find them!
But, since you don't know which ideas are good, and
which aren't, that makes the task before you a lot like . . .
 .
. . looking for needles in a haystack.
The needles are the few
good ideas that will
increase your conversion rate 20%, 50%, 100% or more. The strands of
straw are all the other ideas that will either hurt your conversion
rate, won't be tested, or won't make much difference.
And when all the ideas
are stacked up together,
you can't tell which ideas are the 6 good ideas, and which are the 94
useless ones.
That's why split
testing each idea . . . one at a
time . . . is kind of like . . .
. . . looking for
needles in a haystack, one
needle-shaped thingy at a time.
But
using the
Split Test Accelerator software and methods is kind of like putting
that haystack of
ideas on a magnetized hay sifter
,
and . . .
 .
. . giving it a good shake.
And . . .
voilà!

You
find what you were
looking for often 10 times faster than
you would with ordinary split testing.
In this case, you were
looking for those few
changes to your sales page that will deliver a powerful boost to your
conversion rate.
Now which would you
rather do? Take a week (or a month) and
test just one straw-shaped thingy, hoping it's a needle, and then
discover that it's a worthless piece of straw? (This often happens 3,
4, or 5 times in a row with traditional A/B split testing.)
Or would you
rather shake the heck out of
the entire haystack, and quickly get good results almost
every single
time?
If you value your time, and want to make money faster, and with less discouragement . . .
You definitely want to use ACCELERATED split testing.
But which accelerated split testing tool should you use?
You might have heard that GOOGLE offers an accelerated split testing tool -- for FREE.
They do. And it's not a bad tool. If you don't meet the requirements for the Split Test Accelerator, it's not a bad fall back plan.
But Google's tool is not as powerful.
It doesn't provide as much ACCELERATION.
And before I explain what you get with the Split Test Accelerator package, I want to take a moment to compare the Split Test Accelerator to Google's Website Optimizer.
"Here's How The Split Test Accelerator Compares To Google's Website Optimizer."
The main difference between the two tools is that Google's Website Optimizer is a "full factorial" tool, and STA is a "fractional factorial" tool.
With a "full" factorial tool, visitors are shown every single combination of factors on your page. And it can take a long time to cycle through the combinations if you have very many things you're testing at the same time.
With STA, you can find the best combination by running only a "fraction" of the combinations.
Based on what STA learns from the users responses to a very carefully selected fraction of the combinations, it does an elegant mathematical "projecion" to the best combination overall. (Don't worry, you don't have to know any of this to use STA. I'm just stating the main difference for those who are interested.)
For example, . . . . . . in STA's biggest test, you can test 20 factors at the same time, with over 4,000,000 combinations. You can't test that many combinations with Google's Website Optimizer, and if you did, it would take 4,000,000 visitors to get well-balanced data.
But STA can test all 4,000,000 plus combinations by testing just 36 combinations, carefully selected out of the 4 million. By randomly rotating these combinations through just a few thousand visitors, it can make very profitable recommendations.
The bottom line is that, with the exception of very small tests, STA will give you results many times faster than GWO.
There are some other differences in flexibility, and reporting, too, but rather than explain each one, I'll just lay it out in a table for you:
|
|
Google Website Optimizer
|
Split Test Accelerator
|
|
Accelerated (fractional factorial) testing
|
Limited
|
Extensive
|
|
Documentation for accelerated methods
|
A hint at the possibility
|
Extensive Documentation
|
|
Max combinations for reasonable fractional factorial test
|
About 1000
|
Over 4,000,000
|
|
Visitors Needed To Test 20 Factors
|
8,000-40,000, using 4 separate tests and careful monitoring.
|
2,000-10,000, using a single test. Just set and forget.
|
|
Can test LOCATION of elements
|
No
|
YES
|
|
Can test NON-CONTIGUOUS elements
|
No
|
YES
|
|
Can run a test across MULTIPLE PAGES
|
No
|
YES
|
|
Can EXCLUDE traffic by IP ADDRESS
|
No
|
YES
|
|
Can EXCLUDE traffic by DOMAIN
|
No
|
YES
|
|
Can EXCLUDE traffic by USER AGENT
|
No
|
YES
|
|
REPORTING
|
Up to 24 hour delay
|
REAL TIME
|
|
COOKIES
|
Session only
|
PERSISTENT (up to 1 year)
|
|
Control of Data
|
Google
|
YOU
|
|
Distinct Multiple Outcomes In One Test
|
No
|
YES
|
|
Main Effects Report
|
YES
|
YES
|
|
Report On Every Combination
|
YES
|
No
|
|
Standard Visitor Detail Report
|
No
|
YES
|
|
Standard Conversion Summary Report
|
No
|
YES
|
|
Standard Pairwise Interaction Report
|
No
|
YES
|
|
Standard Visitor Path Report
|
No
|
YES
|
|
Standard Domain Filtering
|
No
|
YES
|
|
Standard Keyword Filtering
|
No
|
YES
|
|
Standard Ad Code Filtering
|
No
|
YES
|
|
Standard Date Filtering
|
YES
|
YES
|
|
Can Test DYNAMIC content
|
YES
|
No
|
|
TERMS
|
FREE
|
Paid
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This information is to the best of my knowledge as of 12/2/2007
Now,
I want to explain what's in the package, and give you more details
about each component.
"Here's
What You Get!"
Here's
what you get with the Split Test Accelerator package:
 | The
Split Test
Accelerator 4.0. |
 | The
Manual:
"The STA 4.0 Setup Guide" |
 | The
eBook:
"How To Use
Accelerated Split Testing
To Get More Sales, Opt-ins, And Other Actions From Your Visitors"
|
 | Permanent
Access to the Split Test Accelerator Support And Discussion Forum
|
 | 6 Months Free Technical Support
|  | An Installation Option |  | An Unconditional 100% Money Back
Guarantee
|
"Here's
The Whole Package,
Broken Down, Piece-By-Piece"
Let's start with the software:
 Your
copy will be an instant
electronic download |
The Split Test Accelerator 4.0
|
This
is the workhorse that takes all your ideas and quickly finds the ones
that will help and the ones that won't. The result: More sales, optins,
or other desired outcomes.
Here's what the Split Test
Accelerator will do for you:
 | Gives you many multivariate
designs, including a test that allows you to test 20 different
factors (that can form over 4 Million combinations) on your
page with the
same amount of traffic required for a single A/B split
test. (Read that highlighted part
again. Most people don't really let this sink in.
That's why it's called the Split Test ACCELERATOR.
It really is magic, and the science behind the magic is
explained in the ebook that comes with STA.) |  | Includes
L36, L18, L12, and L9 designs. |  | Allows you to do simple
A/B/C/D/E/F split testing, too (for those "quick and dirty" jobs). |  | Comes
with full rights to install on any domain you own. |  | Serves
your ads randomly
(ensuring that your results aren't vulnerable to cyclical effects).
|  | Dynamically
assembles your ads "on the fly" (making it easy to set up a
test. You need set up only one page. And you don't have to worry about
which elements combine with which elements. The software does it for
you. And it gets it right every time.). |
 | Sets
a cookie on your visitor's
machine so they get the same page every time (If your visitor sees two
different pages, and then buys your product, you have no way to know
which ad should get credit for the sale, and that can mess your test
up. STA makes sure you don't have that problem.) |
 | Gives
you
live reports. (This lets you know at a glance which ads are
performing best, AND which new combination of factors is presently most
likely to be the winning ad, AND it gives you a pretty good estimate of
what your new conversion rate will be if you run the winning ad.) |
 | Gives
you the ability to switch
testing modes with "one click". |
 | Split
tests
your control against the winning ad with the push of a
single button (You don't have to monkey with anything to
switch to this important confirmation mode. Just push a button, and
you're there. Have second thoughts about switching modes? Just push
another button and you're back in Taguchi mode. No harm, no foul.) |
 | Gives you
complete control over which combinations to run in the confirmation
mode (That means you can pick any of the 4 million plus ads in the
biggest test. Use the suggested winner. Use the
best ad of the 36 that ran in the test. Or make up your own
combination if you think that might perform better.) |  | Runs
your winning ad "full time"
with just another push of a button. (This is the no-messing-around way
to get your new high profit page pulling orders immediately!)
|  | Allows
you
to string tests together along your whole funnel if you want.
(This
gives you amazing flexibility, whether you have one product or
many. You can even chop, slice, or dice a single sales page into many
parts to optimize each part on its own.). |
 | Leaves full
factorial testing (like what's done with Google's Website Optimizer) in
the dust. Some things you can do in a week with
STA, would take you many
months with a full factorial tool.
|  | Gives
you complete control over your data. Since you install STA on your own
server, no one sees your testing data but you. |  | Allows
you to test
different
locations on the page for various items.
Most other multivariate testing tools treat location as fixed
on
the page, so you can't easily test the location of an opt-in form or
image. They'll let you test various images in one location,
but
not various locations for a given image. STA will allow you
to do
both. |  | Allows
you to do advanced
analysis (if you want to -- you don't have to). |
 | Will
filter
your results by source (Find out which ad converts best
with Google traffic, which does best with Yahoo, which is best with ad
A, which is best with ad B, and so on. You can produce laser-optimized
landing pages for each source of traffic if you want.). |
 | Allows
you to filter results by
date range (this lets you pinpoint the exact day a new trend started,
which gives you even more power to convert your traffic). |
 | Tells
you
at any point during the test HOW CONFIDENT
you should be that one option is better than another (This is the
elusive "confidence measurement" that seems to be missing from all but
the best split testing applications. This makes it easier to determine
when it's time to stop testing and start collecting the higher profits
you've earned.) |
 | Tells
you which are the most
influential factors (some factors influence your conversion rate
greatly, and some are almost irrelevant, no matter what you try.
Wasting time on bum factors is the #1 source of discouragement A/B
split testers can face. STA arms you with the power to identify these
seductive little time and money wasters so you can safely avoid them in
future tests. ) |
 | Allows
you
to SEE the projected winning ad at any time during the test
(and you can see which is projected to be the best ad for Google,
Yahoo, MSN, ad A, or ad B, at any time during the test, too.). |
 | Allows
you to preview any of the
scientifically designed test pages with the click of a button (this
allows you to see what your visitors are seeing. It also allows you to
make sure they are seeing what you want them to see -- before you start
your test!).
|  | Allows
you
to set up your page with nothing but cutting and pasting. |
 | Saves
HOURS of your precious
time over setting up your tests manually with spreadsheets (keeps you
from making costly mistakes, too.) |
 | Allows you to "protect
your profits" by sending only a portion of your traffic to your
test. |  | Allows
you to get rid of "Junk Traffic Noise". (Spiders, "hits"
sites, etc. You can exclude visitors from the test by ip
address, domain, or user agent. Most of the main spiders are
already on the exclusion list, and it's easy to exclude other agents,
ips or domains with the click of a button.) |  | Enables
you to optimize multiple outcomes at the same time (How many users who
hit your landing page use your optin form? How many buy your
product? How many view your FAQ page? Now you can
find out, and optimize your page for whichever outcome is most
important.) |  | Gives
you complete visitor detail information. (This allows you to
find out even more about why some people buy and some people don't.
It also allows you to identify junk traffic quickly, purge
that data from your test, and
exclude that traffic in the future.) |  | Gives
you a "Conversion Summary" report to quickly spot which referral
domains, keywords, or ads are giving you the best results. |  | Allows
you to track the paths
visitors take through your site, and how long it took to take each step
in their path. (A nice report converts all this data into
conversion-relevant information.) |
Here's
what Robin Porter has to say
about a previous version of the Split Test Accelerator:
"Wish
I'd Discovered This Years
Ago!"
"Hi Jim, Well, I
don't normally write testimonials for people, but heck, I had
to
tell you how great Split Test Accelerator is! It's helped me turn what
I
thought was a "dead duck" into a "soaring eagle"! Ran the first round
of
tests, and sales started trickling in. Using STA, I was able to compile
the
best pulling sales letter as the "control" and then run a second round
of
testing with more improvements. Now
sales are "flooding" in! Conversion
rates are over 3% - and rising! Wish I'd
discovered this years ago.
Keep up
the good work!"
Robin Porter
|
 Uses
PHP and MySQL.
|
Now
that's the tool. It's a true workhorse,
and it will take all of your ideas and filter the wheat from the chaff,
the gold from the dross, the needles from the hay.
But
we're just getting started!
Here's that book I mentioned on page 2:
 Your
copy will be an instant
electronic download |
The E-Book:
"How To Use
Accelerated Split Testing
To Get More Sales, Opt-ins, And Other Actions From Your Visitors:
|
If
you want breakthrough results, this eBook is just as important
as the
software.
Here's what this eBook will
teach you:
 | Why
most ideas won't be
test-worthy (and how to decide which ones you should test and which
ones you shouldn't. You don't want to rely on your "judgment" for this.
You'll want to do THIS instead.) |
 | Why
most of
the ideas you decide to test won't help (and why that's not
a problem.) |
 | Why
you should picture yourself
panning for gold, instead of rolling a snowball down a hill (if you get
your metaphor right, you will run your tests with more focus, and you
will get better results). |
 | Why
you
might actually get by with less traffic
than you'd need for a typical A/B split test (this doesn't make sense,
but sometimes you get a free lunch). |
 | How
even good copywriters can
shoot themselves in the foot during the brainstorming process (and how
to make sure you don't make these mistakes). |
 | How
to
involve good copywriters in the process, if you have the
resources to do so (and how you might even get their help for free). |
 | How
to deal with "interaction
effects" (Most of the time you will ignore them. But sometimes you can
get slightly better results with an easy little trick). |
 | Which
factors have historically made a difference in other people's
split tests (Hey, you don't have to completely re-invent the wheel
every time. But you don't want to rely on "what worked in the past" too
much either.) |
 | How
to "jazz up" your options to
really supercharge your test, even when you think you're done designing
your test. (Fail to do this and you're leaving money on the table.) |
 | Why
variety
is more than just "the spice of life". Find out why variety
is essential if your tests are to have ANY life. |
 | How
to get your competitors to help
you brainstorm. |
 | How
to use
T.H.I.R.S.T. to create a headline section that makes your
visitors' eyes pop out (OK, that was a slight exaggeration.). |
 | How
to brainstorm like a pro
(scientific brainstorming is the difference between getting 10% or 20%
improvement, and
getting
truly breakthrough results. Sometimes a good brainstorming session will
lead to a test that produces a 100% or 200% improvement, or more).
|  | How
many people you
should (ideally) involve in your brainstorming sessions (and, just as
important, which
people you should try to involve.) |
 | How
to brainstorm with fewer
people than you would like -- even all by yourself, if necessary. (And
how to brainstorm with people when
they're not in the same room, and aren't even brainstorming at the same
time.) |
 | How
to
"brainstorm without brainstorming" (sneaky ways to pick people's
brains without having to "organize" anything). |
 | How
to keep a brainstorming
session going until you have over 100 ideas (this trick keeps people
from getting bored, and makes it more likely that you will uncover
several profit boosting golden nuggets). |
 | How
to
avoid shutting down the brainstorming idea faucet (if you make one
common mistake, you won't get half as many ideas.) |
 | How
to keep large brainstorming
groups from becoming chaotic and petty (if you fail to do this, your
results might suffer). |
 | How
to
compensate brainstormers in a way that's fun, profitable for them, and
low risk for you. (If you want a way to be popular with your friends,
be the person who values their ideas and pays them too! And this isn't
just about "being nice". You will make more money yourself if you can
implement this idea.). |
 | How
to "prime the pump" so your
brainstorming sessions produce more ideas (What ONE question you should
ask when brainstorming. Don't overcomplicate things. If you ask more
than this one question, you will cause confusion and your session won't
go as well.) |
 | How
to get
by if you can't get anybody to help you. (don't go it alone
if you don't have to, but if you have to go it alone there are some
tricks to getting better ideas from yourself.) |
Here's
what Dr. Glenn Livingston had
to say about the document that led to this eBook. This was before
I had learned what I now know about the importance of brainstorming and
test
design:

". . . the clearest, easiest to
understand explanation I've seen . . ."
"Hi Jim - I just wanted to let
you know that I'm very impressed with the
way you've put together your Taguchi product*. [The manual is] probably
the clearest,
easiest to understand explanation I've seen about this type of
analysis."
Glenn Livingston, Ph.D.
* He refers
here to an earlier version of the Split Test Accelerator
|

|
Those two items are very powerful together. But it's no use to you if you can't get it
set up. That's why you also get . . .
 Your
copy will be an instant
electronic download |
The Split Test Accelerator 4.0
Setup Guide
|
This guide walks you through
the initial setup, step by step.
And then it shows you how to design
your tests, step by step.
It makes it all a simple,
"connect-the-dots" process.
|
and, you don't even have to worry about installation at all, because you also get . . .
"Installation
Option"
For just
$35 extra I'll install the software for you. You will be
given the option to purchase installation services after purchasing
STA. |
And
there's more . . .
Here's something that should put you at ease:
The
Split Test
Accelerator Support Forum
There is a support forum available for the Split Test
Accelerator
(there's even a prominent link to the forum on every page of the
application).
With already over
150 topics and over 750 posts, this forum will provide you with a very
high
level of support.
And, maybe even more importantly, you
will find a place where you can ask for input on your current
projects, without feeling like a mooch.
Far from a mooch, you
will be more like a hunter bringing back a nice big carcass for the
tribe to devour. And you will get input from
people who really understand what it takes to make your project better.
At the forum you can:
 | Get
Technical Support
(just post your questions, and I'll help you troubleshoot any problems.
I'll help you get set up, and I'll help you interpret your data, if you
need. Until the forum reaches critical mass, I will answer every
question personally.) |
 | Discuss
strategies for running your tests. We are an accelerated
split testing community and often share results with each other
(optional, of
course). You can also share and receive tips and tricks for getting
even more out of the process. |
 | Discuss
theory (If you're
interested, we can get into the methodology and mathematics a bit). |
 | Use
the
forum for brainstorming (if you want to
open your tests for general discussion, you can get some informal
brainstorming from others at the forum.) |
 | Discuss
anything else related to
split testing, marketing, or advertising. |
|
And the forum is NOT your only
means of support. You will also have direct access to my help desk for 6 months.
"Get
6 Months Free Technical Support"
If you
run into any
problems, even problems you created, you will be able receive assistance for 6 months |
|
Now,
what about the price?
Let me ask you . . .
"How Much Would You Pay
To Have The Power To Double Your Profits?"
(And see why $10,000
might
be a fair asking
price for the Split Test Accelerator.)
Here's one way to
answer the value question . . .
|