Writing Global Surveys for Success, Not Failure

Interesting title, don’t you think?

Most companies that take on writing global surveys have failed at it.  Some of them did it once and the never again.  Others are still doing it wrong but since the results they get are what they want, they won’t change for fear of actually finding what their customers think of them.

This is a touchy topic since it is very easy to fall into stereotypes and make generalizations about entire countries or regions, and end up looking like a bigot.  Which, I am not.  Thus, I am not going to point to any specific country or region around the world and say “these people do this” or “those people don’t understand that”.  Instead I am going to point at the more general problems and trends, and we can later work together if you are interested in a specific country or region.

Four key things to remember when doing global surveys:

1. Some people are polite.  Some cultures are too polite to really tell you what you want to know.  The best you will get is a polite response saying things are fine, when they are not.  This means you will learn to write the questions in different ways for those groups, and you will use different scales.  I know of at least one case where a client of mine with a 90% satisfaction score had an attrition ration in excess of 26%!  How did we solve it?  We figured out what questions to ask and what were the best scales to find out what customers really meant.  Trial and error works great in these situations.

2. Some people are brutal.  On the other hand, some people are brutal.  You ask for feedback? no problem, you stink and your product is so bad it hurts me to use it… just like everybody else and all their products!  The key in this situation is to put things in proper perspective.  Sure, the brutal criticism may, just may, be well warranted – but chances are that everybody is getting similar feedback.  So, take the time to put it into perspective.  Try to benchmark your performance against your competitors or even your partners.  Work with a local service provider that can help you understand how your ratings rank in comparison, and work to improve your processes and products, not your scores so much.

3. Languages are all different, literal translations are bad.  Asking someone if they are satisfied with the bad behavior they received from the last interaction is not a good question to lead off a survey.  Alas, I have seen it.  When you construct questions in one language, literal translations sometimes change the meaning of the question in a different language.  The more diverse the languages and differing their grammars and syntax rules, the worse it is.  Do it right, craft new questions with similar meaning in different questions.  However, they don’t have to ranked or rated the same way, nor be the exact same question – why?  See point 4.

4. Reporting should be detailed, not generalized. Here is the biggest thing about doing global surveys, which I hope you gather from what i said before: each survey should target specifics for each region or country, and as such should be reported separately.  So, even if you use the same questions and scales, do not mix the results of two different locales in the reporting.  Let me say it again – do not mix results.  You probably already knew that, or probably figured out from the earlier points I made – so I won’t preach it again.  OK, just one more time – don’t mix results from different places.  Why? two words: skewed results.

Bottom Line: write your surveys for each region or country in the region or country, using local resources.  Don’t cross-tabulate as scores don’t mean the same in different parts of this world. And make sure you report for each region or country, not across all of them in a single, global score.

Need more? Got something to add?  That is what comments are for… feel free to let me know what you think.