Posted on:
A client asked for some tips on creating a survey for their first small-scale research recruiting respondent. Worth sharing:
When you want to learn from others, whether it’s via an interview or survey or any other way, I find there’s always a tension: there’s what you REALLY want to know, and then there’s real or perceived constrains on what you can actually learn, ask, etc. For example:
- I want to know EVERYTHING. People have 24 hours in the day, and far less than 24 hours to give me.
- I want to get straight to the most valuable stuff. People feel vulnerable when sharing some of this “most valuable” stuff. You get the picture.
In general:
- You will rely on coverage to compensate for shortfalls elsewhere. In other words, if the ideal would be 10 1-hour interviews going into great depth, a more realistic thing might be 30 much shorter survey responses. The latter gives you some coverage that partially compensates for what you miss out by not having highly in-depth interviews. If you’re familiar with the concept of oversampling a signal to get a better digital/analog representation of it, there are some metaphorical similarities here.
- Don’t sacrifice what it is you want to learn, but do think about ways to make it easy and convenient for people to help you learn what you want to learn. Metaphorically, I think of the busy executive barking “let’s walk and talk” to their subordinate so that that walking time can be used to transfer information. That’s one reason I love Alpharun’s value prop. One idea to try out: the initial survey is literally just 1 or 2 questions, and at a later time (not too far into the future, but somewhat later) you ask those who completed the first survey if they would provide a bit more insight via a second survey. Of course you’ll get “dropoff” in that “funnel”, so you need to account for that, but it’s a way to make it easier for people to teach you what you want to learn.
- Remind people why their help here is valuable. If you can connect this back to how this will ultimately help them, even better. There’s a very real sense in which recruiting participants in research is not a research problem, it’s a marketing and sales problem. :)
- Have someone else look at your questions. Surprisingly often, they’ll spot phrasing that makes perfect sense to you and is not at all clear or meaningful to an outsider.
Tagged with:
More posts:
- Next: Play Structure, Missoula, MT
- Previous: Divergent, Not Convergent, RAG