My friend and colleague, Beth Kanter, asked me for advice about how A/B multivariate testing applies to the design and execution of calls-to-action in the non-profit sector, with a particular eye toward how we leverage the social graph. What follows is my response to her question, from which I would hope that others will draw some benefit.

Malcolm Gladwell reveals the key to success in a presentation that he made to the TED conference in 2004 entitled "Learning from Spaghetti Sauce".

He concludes that "by embracing the diversity of human beings you will find a sure way to true happiness".

We have learned that "One size fits all" is a failed strategy because no single message appeals significantly to a diverse mass audience. Why do we persist in sending out calls-to-action (such as an invitation) when the conversion rate is predictably too small to be considered successful?

You should apply Gladwell's lesson to how you determine your invitation strategy. I recommend that you do so through multivariate A/B testing.

Campaigns for causes are like branding/imaging campaigns for powers of association. Their pitch and presentation must be re-designed (in wording, composition, content, imagery, layout, audio, voice, video, typeface, user-interface, interactivity elements, recipient psychographics) to fit different target audience clusters in the social graph (based upon interests, activities, gender, income, age, political affiliation, ethnicity, media platform/network/channel preferences, etc.).

Given a sufficiently large testing sample, variations on the test message can be allocated for testing upon different recipient clusters. Doing so helps you to figure out the right match of message<->target cluster.

When we perfect a portfolio of clustered solutions, a better overall conversion and ROI is assured than when we pursue the single most "perfect" productive solution.