Most businesses A/B test their emails and landing pages without a second thought. Fewer think to do the same with direct mail, yet it works just as well, and in some ways gives you cleaner data, because there’s no spam filter or algorithm getting in the way of what you send.
If you’ve ever wondered whether a different headline, offer or design would get a better response from your customers, A/B testing your mailing is how you find out for certain, instead of guessing.
A/B testing means printing two (or more) versions of the same mailing, each with one meaningful difference, then sending them to similar segments of your audience and measuring which one performs better. The “one difference” part matters. If you change the offer, the design and the headline all at once, you won’t know which change actually moved the needle.
This is where most businesses get stuck. Print alone doesn’t tell you who scanned, called or ordered, so you need a way to trace the response back to the version someone received. A few reliable ways to do this:
The key is consistency: every recipient of Version A gets the Version A code, every recipient of Version B gets the Version B code, and you don’t mix the two.
A few things make the difference between a test you can trust and one that just creates noise:
The real value of A/B testing isn’t the one campaign it improves. It’s what you learn about your audience that you can apply to every campaign after it. Once you know your customers respond better to a specific type of offer, or that personalised mailings consistently outperform generic ones, you can build that insight into your standard approach, rather than starting from scratch every time.
Not necessarily. The mailing itself just needs two versions with a trackable element, such as a unique QR code or reference number, on each. The tracking side can be as simple as a free QR code generator paired with your existing website analytics, or a dedicated campaign tracking tool if you’re running tests regularly.
This depends on your typical response window, but most direct mail campaigns see the bulk of responses within two to three weeks of delivery. Give the test enough time to reach that natural response curve before drawing conclusions, otherwise you risk reading early or partial data as the final result.
Yes, and the two work well together. You could, for example, test a personalised version of a mailing against a generic version to measure exactly how much personalisation is worth to your response rate, rather than assuming it makes a difference.
Running a proper A/B test, from print production through to tracking and personalisation, is exactly the kind of thing our data management and personalisation team handles day to day. If you’d like help setting up a test for your next campaign, get in touch with our data specialists on 01273 464 884.
