Friday, March 6, 2009

Testing is necessary, but be warned.

I love testing things.

But I also believe that it's easy to test things for the sake of testing. I've been guilty of this in the past.

Don't get me wrong: testing is great, and for direct marketers, essential. But time and time again I see test results that read something like this…

‘There was an observed difference in response, but the difference was not statistically significant’.

This can happen for a number of reasons but typically because either the sample sizes were too small or the test was just too tactical and not likely to produce a meaningful or different result. In other words it was just stupid.

I’d implore you to consider some long term (say, 12 month), strategic tests that will have a big impact on your program. The barrier to doing these is time, cost and direct marketers being incredibly impatient. We want to know everything yesterday, not in a year. Be patient my fellow fundraising friends, the longer term tests are likely to provide the most value for your program. And ultimately those brilliant causes you support.

Think about undertaking one/some of the following over a fixed period, not just one appeal:

• Testing the impact of different streams of donor care materials for different groups of donors and measuring the comparative value of those groups at the end of the test period.
• Testing the impact of treating different value donors differently. For example spending more on mid value donors to see whether they behave like high value donors. In other words, do they give you more?
• Testing the impact of varying ask strategies to try and find the optimum ask prompts for various donor group.

Don’t obsess with meaningless testing. They might seem like fun things to test but assessing whether blue font at the top of the response form out pulls red font ain’t going to change the world.

Oh and finally, for those of you interested in my little furry friend Ricky the raccoon. He’s doing well. But even more exciting was my spotting of three deer last night on my brief trip to the breathtaking Banff, Alberta. Didn’t see any grizzlies, but the deer sighting was memorable.

Jonathon

2 comments:

John Lepp said...

I LOVE testing. Sadly, a lot of my clients don't. What is even more distressing is, as you mentioned, a lot of people I deal with don't seem to know what a test is - or what makes a valid test. I've seen too much wasted money testing two white oe's with slightly different taglines. How we're the results? I'll ask. "Oh, they were basically the same..." Oh yah? Go figure... sigh. Sorry kids, that was a stupid test.

Jonathon Grapsas said...

Indeed John.

Should we tes? Yes of course. But I fear that sometimes we test for the sake of testing. Ive done it myself.

What I would like to see is more strategic, long term tests that have the potential to deliver significant learnings.