Alison Cherry & Joanna Miles, Beaconfire

When it comes to nonprofit fundraising, everyone wants to know "the best" way to do things. When the economy gets tight, doing things properly becomes even more critical, and even small decisions can cause stress and disagreement.

Imagine you're redesigning your donation form. Deceptively simple questions like these may come up:

• How many steps should we have in our donation form?

• Should our form be one column? Or two?

• What dollar amounts should we use in our ask string?

• What's the best language to ask for donations? Give? Support? Donate?
  

The answer is usually the one we all fear (and hate): "It depends." It depends on your mission. It depends on your audience, and the design of your site, and the goal of your campaign. We can quote best practices, and case studies, and get a pretty good idea of what will probably work well. But, sometimes what you know won't work turns out to be just what you need.

In this situation, there's only one way to know for sure what's best: test it for yourself. Take the opinions out and let the data decide for you.

Multivariate testing (MVT) is a great tool for making these sorts of decisions. With MVT, you optimize a web page by creating variations on key areas -- like form layout, ask string, or submit button -- and test combinations of these variations together in real time. In the end, you not only learn what variations work best to get real users through your donation process, but how they perform in combination. Then, you can take hard data back to your decision-makers and say, "This is what works best for us."

Simple changes, big results

Often changes that seem very basic -- like the text in a headline or the color of a button -- can make a large impact on user behavior, and can be tested with relative ease. Such tests can address a range of concerns, from design to user experience, to help you move real users through your donation process. At the very worst, you will validate what you're doing that works well, and learn what types of changes make a difference in your users' behavior. At best, you will learn just what makes your users respond, gather some compelling data, and boost your online fundraising.

Fine-tuning

Having tested and re-tested your page's core elements and arriving at a strong result, there are nearly limitless ways to further fine-tune. (Sometimes, the best results you can get are better questions for your next test.) You can tweak your most successful variations, play with style choices, or optimize more complex elements. For example, an ask string on a donation page could be adjusted in many ways:

• How many values should you have?
• What should the highest value be?
• What should the lowest one be?
• How should you order them?
• Should they have descriptive labels?

Even if you tested your ask string early on, you could still run an entire test simply optimizing it for some of these additional factors. When a number of smaller changes like these work well together, they can have just as big an impact as more dramatic changes.

Validating the hard choices

If you're attempting to make major changes, or changes that are internally controversial, MVT can validate your decisions. Just test the new versions against the old to confirm whether the changes have a positive impact, or whether they make any difference at all. This approach provides stronger data than looking for changes in your analytics. By testing both versions at the same time, randomized to the same audience, you account for day-to-day and week-to-week variations in your data. This can be an invaluable strategy for moving decisions along in your organization: saying "I want to test this" can be an easier and less risky path than "I want to change this."

Testing on a budget

In multivariate testing, the benefits grow over time with repeated rounds of tests. It's not a magic wand, though. While you may read studies where a non-profit grew their donations by 30% or more with a single test, such results are far from guaranteed. Instead, it is the cumulative effects of smaller gains that make MVT work.

Suppose in your first test, you find two variations that lift your completed donations by 5% each. (This is purely hypothetical -- a successful change for your website could be larger or smaller.) Together, they might give you a 10% lift. If you run another round of tests and find a new variation that boosts that result by 5% more, now you're 15-16% ahead of where you started. Each subsequent round of testing will build on and strengthen your earlier results. Over time, increased donations should exceed the costs of running the tests.

At a time when so many organizations are cutting back, starting a new program can seem daunting. But the prospect of leaving fundraising dollars on the table, or letting potential donors turn away, should seem even more daunting.

With MVT, gathering extra data has a small marginal cost, and a direct positive impact on your online donations. It can help you make simple, low-risk, high-yield changes to make your donation process as effective as possible.

To learn more about how multivariate testing works, take a look at Beaconfire's new white paper, "Is Multivariate Testing Right for You? A Guide for Nonprofits".