Writing Surveys to Collect Technical Information at Organizations

I have recently finished writing a number of surveys to acquire technical information from users. I actually love creating surveys because I learn a great deal about asking all the right questions for the information I need. They are also a collaborative effort sometimes involving quite a few individuals from subject matter experts to management.I thought I would write about it because I think this aspect of technical communication is rarely thought about. Here's what I have learned on creating surveys from my recent batch of them. I'm not getting technical like the type of instrument such using a Likert scale, but more in the area of basic ideas to keep in mind.

1. Focus on writing a quality survey. Talk to your subject matter experts for research
2. Do some usability testing - this step for me really counts. I find having a few people in your target audience try out the survey really helps.
3. Change it as necessary - feel free to change or customize a survey for best results, especially if users test and you discover readability issues.
4. Never assume users understand the questions. Make sure you rid yourself of a question if it isn't meeting the scope of the survey or standards you work by.
5. Keep an open-minded attitude on the survey and the data collected. This is definitely something I learned getting my doctorate. Also watch for outliers and think about them. You may or may not want to discard them but don't ignore them.
6. Focus on data and metrics that are actionable. Remember you need results to tell you something about what you are asking.
7. Finally, get more involved feedback from users, subject matter experts etc. That way you can continuously modify if the survey is ongoing.

Now my approach for my current surveys  is slightly different from research I have conducted for academic data collection I have performed in so far as I am involved with far more individuals while designing the survey than I was in grad school. This means there will be far more compromises on the questions as far as rigid research standards go but I do find better data collection usually takes place because as everyone works together better questions are formed to retrieve the data organizational management wants. The one thing that doesn't differ however, is the grave attention to detail. That is a must for any survey.