Do you have three tested direct mail packages waiting in the wings to use when your control starts to fatigue? If you don’t, you should. It’s never a question of if a control will die, it’s when. So what is most important to test now to get to that breakthrough package? Here are some ideas.
Smart direct mail marketers are constantly testing. It may be the offer, positioning, format or anything else — but what variable gets you the biggest bang for the testing dollar? And which test delivers the most favorable cost per acquisition?
Traditional mail testing can be very expensive, time consuming and yield limited insight if not executed correctly.
After personally overseeing and writing multitudes of direct mail packages, it’s still tough to choose just one variable to test. The reality is that several variables should be tested all at the same time to get to a new control faster. These are the types of tests I’ve found most successful in revealing key attributes for a new control:
- Your offer is highly influential in your response. If you’re testing price (most typical), you can test dollars off or percent off. I’ve found offering dollars off to be best, but every market is different.
- Are you including a bonus or free gift?
- Repositioning your product — or testing a new unique selling proposition — can reinvent your complete message and offer to produce sizeable increases in response rates.
- A new production format can refresh an existing control. Perhaps you’ve used a #10 outer envelope for a long time. A simple switch to a #9 or #11 envelope can make a difference.
- I like to include showstopper text and graphics on my envelopes, each worth about a half-second of time for the recipient to pause and study the OE. I’ve found elements such as faux bar codes, handstamps and seals yield favorable impact.
- Evaluating data overlays from models or profiles will return tremendous information and insights. But if you don’t spend the time to interpret it and imagine the possibilities, you can overlook great new ideas.
So with all these test possibilities and data, what variables should you test?
In my last column, I shared a new Bayesian Analytics methodology that I think will upend direct mail testing as we know it today. Bayesian Analytics isn’t new, though its current applications are new and spreading to many fields, including weather forecasting, insurance risk management and health care policy. Later this month I’m moderating an online session on this topic (learn more at my website).
A/B testing is effective, but usually builds a new control quite slowly (how many times have you tested, only to find the test performed under your control?). Multivariate testing enables you to isolate variables and achieve a new control more quickly, but it still takes several packages to confidently identify the winner. But the use of Bayesian Analytics in direct mail gathers substantially more testing insight and produces more cost savings, while taking less than half the time of traditional testing.
I believe in taking out the guess-work of testing where it’s possible. Otherwise it is easy to incorporate our own personal emotional appeals and biases, like when we say “I’d never respond to that!” We’re probably not our own market. We’re often wrong, even as informed as we are about our products and audiences.
My point is this: You must keep testing. Test outside your comfort zone. Let your prospective customers tell you what variables they’d respond to by using Bayesian Analytics methodology to deliver the emotional insights that big data can’t deliver.
If you don’t have at least three tested packages, or knowledge of what variables form the magic combination necessary to increase response rates, Bayesian Analytics will save you a lot of time and resources.
Download my new report, “Predicting Direct Mail Results Before You Mail” to learn more about Bayesian Analytics.