What are the most effective ways to experiment with your demand generation campaigns?

Carefully select the metric you're testing: If you are e.g. testing the headline of an email, you can select the opening rate. If it's the call-to-action you are wondering about, the opening rate will not make you any wiser, if it's the same for all recipients. If you want to test several parts of e.g. an email or newsletter, like headline, banner and call-to-action, start at the top, i.e. the headline. If the recipients are not attracted by it to open your email, they will never get to see the rest of it. When working with a small dataset, it's better to concentrate on the top aspect first. Otherwise, the number of recipients actually getting to the tested part could be too small to deliver robust results.

The most important thing I’ve found in experimentation is making sure your test segment has enough data or conversion volume that you will get a result fairly quickly. My maximum experiment length is 2 weeks. Any test that does not reach statistical significance by that time, I pause. Running tests with very small groups or pools of data can often lead to inaccurate findings. There are too many other factors in play when a test running for long time. Make sure your sample sizes are large enough and you’ll be much effective in your CRO program.

When testing emails and newsletters, using a capable marketing automation tool is very helpful. Choose a tool that lets you select the sample size and covers the randomization. If your total database is large enough, you can select a (still large enough for robust results) portion for the test, and use the remaining data to automatically send the "winning" email or newsletter after your test.

I find that folks get out of the habit of experimentation because they don’t want to “fail” or they have a few hypotheses that are proven false. Especially in B2B marketing where volume and data is so much smaller, remember that any test where you learn something is a successful experiment. If you prove your hypothesis, that’s a success. If you disapprove your hypothesis, that’s a success. The only “failures” are tests where you do nothing because you aren’t testing something that has a big enough impact or your data pool is too small.

One of the most effective ways to experiment with demand generation campaigns is offer. Marketers tend to focus test strategies on media channel and creative, but calls to action – what you’re asking the reader to do and why – often have much greater potential to impact engagement and results. This is especially true in channels like search (SEM), where offer needs to align with both the specific interest of the person searching but also where that individual is on the buyer journey. A buyer in the latter stages of research may be more likely to request a demo or free trial, whereas an early-stage prospect is more likely to want downloadable content (a white paper or analyst report) that doesn’t require engaging with sales.