How a Startup Should Interpret Survey Data

The beer goggles of survey data analysis
The beer goggles of survey data analysis

Conducting a survey is a phenomenal tool for startups to use for marketing, customer research, and product decisions.  The “survey” series began with this post, and we’re ready to nerd out and interpret your survey’s results.

Let’s start with a few lenses through which you should view your data.  Kinda like “beer goggles” for number crunching, without the hangover or walk of shame.

  • You’ve just conducted an online survey, so there will be a slight bias in your survey results towards the tech savvy.  I don’t believe that it is enough of a bias to get your boxers in a twist, or your proverbial panties in a wad.  Gross.  Nevertheless, keep this in mind.
  • Never believe anyone’s numbers but your own.  The survey expresses “declared intent”, but this can vary dramatically from “actual intent”.  In English, people often behave differently than what they say in a survey.  Don’t base your entire company’s strategy on survey results.  It is only another data point to consider.

Before you begin, annoint yourself “Chief Data Scientist“.  You’ll feel much smarter and more capable of extracting incredible insights from numbers, charts, and graphs.

With your new title, take the first question, and copy the question itself and the responses to a Google doc.  Do the answers show:

  • An upward or downward trend?
  • A preference towards one topic or another?
  • A picture of who the audience is, was, or is becoming
  • Groups of data, that if bundled together, make a different point?  See this example here
Q by Q analysis in a Google Doc.
Q by Q analysis in a Google Doc.

Write down your observations below each question.  Come up with different interpretations of the data.  List out ideas of what you might change based upon that answer.  Here’s an example of what this looks like.

Do this for every question.

If you’ve used SurveyMonkey’s Audience, go back through each question and see if adding in demographic info creates new insights.  Do age ranges create different preferences?  Gender? Income?  The example below shows that demographics in mobile are evenly spread for local news consumption.  This is surprising because the TV audience typically skews older.  Pretty cool, and a valuable data point.

Double your pleasure with demographic data
Double your pleasure with demographic data

Share the document with your team.  Politely ask them to read through your analysis in order to poke holes in it as well as add different observations.  The key point here is that you absolutely need more than one pair of eyes on your survey.  This will create better analysis and less bias.

Surveys are fun because you get to use your analytical and data crunching left brain with your thoughtful and creative right brain.  That’s up next – making your results look pretty to put in webinar and whitepaper.

 

 

Comments are closed.