|
Testing, Testing... 1, 2, 3
If at first you don't succeed, "try, try again" and
"test, test again." One of email marketing's greatest
advantages is that it is easier (and much less expensive!) to
test than any other marketing vehicle. This means with just a
little effort, you can increase the effectiveness of your
campaigns and get the results you're looking for. Email
marketing author and expert, Kim MacPherson, explains how.
Testing, Testing... 1, 2, 3
by Kim MacPherson
How does a marketer get from Point A (actively marketing with
e-mail and getting fair-to-middling response rates) to Point B
(actively marketing with e-mail where campaigns are
continually optimized) effectively?
Over and above utilizing some of the many helpful tools from
the e-mail industry's higher-tech offerings, the answer, of
course, is to TEST. The procedure (in a nutshell) typically
plays out like this: Test one or more variables. Select the
winner/control. Roll out. Do it all over again and again,
testing new variables each time.
Not everyone knows how to set up an e-mail test properly. And
test campaigns are not the easiest things to set up, plus they
can make reading results a lot more complicated. If nothing
else, however, there are a few well-heeled rules that should
be followed and if you can apply lessons learned from each
test to your subsequent campaigns, chances are very good
indeed that you can dramatically improve your response rates
over time.
Testing can be as simple as breaking up a list into two parts
(the proverbial A/B split) and emailing each group a unique
"something."
For example, perhaps you want to improve your click-through
(and, hopefully, your conversion) rates. In that case, you may
opt to test offers. Or you may need to test the price points
of your product offerings so you may send out two versions of
your promotion - one with Price X, the other with Price Y.
There are a host of variables that you can test with each
outbound communication; and each new test can break new ground
as far as optimizing your results.
The easiest way to start is by creating a grid. Below is a
grid based on a very simple A/B split that is testing two
different offers.
| |
|
TEST CELL |
VARIABLE |
QUANTITY |
KEYCODE |
CHECKPOINT: MAILDATE, OFFER, COPY, ETC. IS
IDENTICAL? |
|
Cell A |
Offer 1: Free shipping |
5,000 |
1006V1 |
Yes |
|
Cell B |
Offer 2: $5.00 off |
5,000 |
1006V2 |
Yes |
|
| |
Most of the above is self-explanatory, but here's an
important point worth noting: As far as the "checkpoint"
goes (the last column in the grid above), if you're
testing more than one variable, it's critical to make sure
that you test each of them one at a time, meaning one
variable per test cell. This makes for more accurate
results.
Quick caveat: I caution you on testing too
aggressively with an outside opt-in list that you rent
when trying to acquire new customers and leads. Often,
you'll find that lessons learned in one test to a certain
list segment doesn't apply in a rollout scenario to that
very same segment.
For instance, long ago, my company created an e-mail test
for a technology company to test a lead generation offer
to promote its software products. The initial test,
e-mailed to decisionmaking members of a specific list -
let's just call it "Technology Today" (name changed to
protect the innocent). The list selects included job
title, industry, and business size. The test campaign,
e-mailed to 5,000 people of the 23,000 select universe,
did wonderfully well. For a business-to-business
acquisitions e-mail campaign, click-throughs and lead
sign-ups were strong - over 8% and 55% of click-throughs,
respectively. So we took the remaining 18,000 or so people
on that list and e-mailed the exact same campaign, exact
same offer a couple of weeks later. The results? While
still good enough for the client, they dropped by over
50%!
A house file e-mailing, however, can yield results that
you can take to the bank again and again - provided, that
is, that your test segments contain statistically valid
quantities (enough for 10% of your entire list to pull at
least 50 leads, click-throughs, or sales, depending on
what you are measuring).
Also keep in mind that if you're using an agency or a
"self-service" ASP for deployment, the company and/or the
tool will most likely plug in a unique tracking code of
its own for your "call to action" links for each cell.
However, if you're sending internally via homespun
methods, a date/version keycode such as the one above can
be helpful to use because, after the results are final,
it's easy to tell immediately which one of your cells is
the winner... without having to consult yet another
spreadsheet. Testing is just a far more cumbersome process
to go through without the help of a deployment solution.
When all is said and done with a test such as the one
above, you'll know more the next time you hit "Send."
|
Email
Marketing Video Tutorials Show you how to plan, create, and send
successful newsletters.
Read Email Marketing Tips
and Lessons written by Michelle Keegan, Constant
Contact's Email Marketing Diva.
Learn E-Marketing Lingo from the
glossary of Email
Marketing terms.
| by Michelle Keegan, Constant Contact's Email
Marketing Diva(TM) |
|
