Apr 23, 2003
SUMMARY: Do you think it is possible to run an email test in the morning, check your results, and roll out your full campaign that afternoon? Lots of people do. According to email research king Michael Wexler, they may be declaring the wrong winner.
In Part I of our exclusive interview with Wexler, you will learn why you should wait eight hours between test and roll-out. Also includes practical info on how much segmentation you really need to invest in (is one-to-one too much?) and which subject and from lines get more opens these days.
Email researcher Michael Wexler has a weird-but-perfect background for his job.
He was in grad school studying to become a psychology professor specializing in the field of personal relationships, when one day a friend phoned him up from a new company called Hotmail. "Hey, we need someone to help us figure out how to measure this stuff."
After much soul searching, Wexler bailed on the ivory tower to become a pioneer in the field of measuring email relationship marketing.
Currently he is with broadcast vendor e-Dialog where he helps clients including the NFL, Hasbro and Cheaptickets.com measure their results and figure out how to do better. He has got a great eagle-eye view of what works across many campaigns.
We grilled him for 90 minutes to get the following useful info for you. This week in Part I of our report you will learn:
a. Time of day, and day of week, test results
b. Best From and subject lines
c. Segmentation: How one-to-one do you have to get?
-> a. Time of day, and day of week, test results
Wexler is eager to debunk a big myth: Although it is true you may get a majority of your campaign results in within the first hour, he warns marketers against basing campaign roll-out decisions on them.
Unlike direct postal mail, there is no "doubling date" in email whereby you can double current results to get a clear estimate of exactly how the rest of response will pan out. Why? Because instead of being day by day, email is hour by hour.
Some demographic chunks of your list will only reply at certain hours. For example, some consumers only check personal email at night when they get home. Even in B2B, you can be affected by time zones.
Wexler says, "People have certain response patterns that are not dependent on the time you mail. We've seen test results flip flop as evening comes on, and as home users in California get to email. Your response curve looks a lot like your Web site usage curve."
Therefore, you cannot run a test in the morning and roll out in the afternoon with a large degree of certainty that you have picked the best creative. In fact, for B2C campaigns, Wexler recommends waiting at least eight hours before declaring a winner.
When it comes to day of week, Wexler says aside from sending business messages during the working week, he has seen less of a difference in results by day of week than one might expect. "A day of week tends to really not have as much impact as having a very relevant message. If they care enough about it, they will look for it and be interested enough to respond to it."
-> b. From, and subject lines
Wexler noted something we did not know (but hopefully you did): AOL versions seven and below only show the email address of the sender and not the 'from name' than you have carefully added in. He says, "If your email address is a long string of numbers from your vendor, then that's what they'll see."
You will get better response if your brand name starts the from email address, so talk to your vendor and your IT department to make that happen. "It does require a bit of tech configuration, but it's a really wise way to go," says Wexler.
If brand names are so powerful in "from" lines, when should you use a human being's name instead? Wexler advises that you do this only if the human is someone your recipients will have heard of. Familiarity breeds opens.
Subject lines are the same way. If you are mailing to the same list repeatedly (i.e. regular mailings to your house list), you will probably get a better long-term response if you keep an element of that subject line the same every time.
That does not mean use the same entire subject (too boring), but at least always re-use your brand name or something similar near the start so recipients can quickly identify that the email is something they want to pay attention to.
Aside from that Wexler repeated the advice we have heard from many other experts, "Misleading subject lines really irk people. Feel free to have a teaser, but try to be more straight forward. Don't try to hide what message is about."
Adding in recipient's names to the subject line used to work, but Wexler warns that it is probably too overused by unsavory mailers these days. Stay away from it, or at least regularly test name on versus name off.
-> c. Segmentation: How one-to-one do you have to get?
How do you figure out how much to invest in segmentation? Everyone knows segmenting your list to provide the right offer or content to the right person can help results, but it also costs more than it may be worth sometimes.
Wexler advises, "There's a high cost to entry for classic segmentation testing, it costs more than a simple test. However, this investment helps a lot. It stems lowered results, reporting of spam, unsubscriptions, and, it results in higher lifetime value per name."
Before investing in segmentation, be sure you have got measurement systems set up to track both email account lifetime length and value.
This is crucially important because unfortunately segmentation tests do not always show immediate results. "It takes a few emails to [recipients] for them to recognize that it's the right email for them. It may take six emails and maybe they'll only open three of them."
This is a hard lesson for marketers used to the idea of test results within an hour or day. Segmentation is about increasing results over a lifetime as your list slowly begins to trust that you will always send them useful, relevant messages.
Why not forget the whole thing and continue blasting your offer-of-the-week to everyone on your list?
"By sending out the same mass mailing, you're scaring away people who might be customers, and causing destruction to your brand because it doesn't meet the expectation of the user. You've worked hard so your store and brand resonates with consumers, and you destroy that when you're sending undifferentiated mailings."
How segmented do you really have to get? "Not one-to-one," says Wexler. "You don't need to segment to every single person. It just so happens that many people cluster together. A lot of people like Star Wars for example. You certainly have to treat them differently than fans of Harry Potter, but you don't have to treat them differently from each other."
"For resource allocation, pick the groups that appear to be high value, and then combine others. With high value groups you spend a bit more, with others you're throwing away a bit of power. You have to play with it to determine which ones may not make enough of a difference to segment for."
The base level is self-segmentation; allowing your subscribers to choose which lists of yours they want to be on. This works well for mailers who tend to be highly targeted anyway.
As your audience and breadth grows, you can use a combination of email and Web click tracking plus purchase history to slice out segments. This, of course, means your Web metrics and your buyer database will need to be integrated with your email database.
Wexler assures us that it is worth the work because high message relevancy is the best way to beat inbox-overload.
No matter how much other crud comes into your recipient's email boxes, if your message consistently matters to them, they will pay attention.
"We've seen clicks slightly decline over the past year, but it's not been that important," Wexler notes. "People are removing themselves or not clicking on lists that weren't that great for them anyway." So many of the people you are losing were never going to be high responders that were worth fighting for.
Wexler's touchstone truth, "The more we can make our email relevant, the more it can work."
In next week's issue of MarketingSherpa we will bring you Part II of this exclusive interview.