Dicing with Danger: Using Data Insights in Your Marketing

If you've suffered through my last two blog posts (#1 here, #2 here), you've read my reasoning that while marketing data potentially can lead your business into danger, they can also provide meaningful insight.

Now what?

With insight, you can start placing bets with data.

Hit it, Frankie:

While you will never out-cool Albert Francis Sinatra, you can probably make smarter bets.  Here's how.

• Start with your insights.  Review all of the insights you've generated from data.  Remember, these insights consist of guesses made on what you know.  They offer no guarantees of success, only increasing levels of certainty.
  
  Let's go back to insights we hypothesized in the last post.  Let's say you sell industrial coatings used in the electronics business and that you found large groups of electrical and chemical engineers in your database.  Based on discussions with subject matter experts, you learn that electrical engineers want to know detailed tech specs whereas chemical engineers will always request an MSDS, or Materials Safety Data Sheet.
  
• Turn your insights into testable propositions.  An insight means nothing until you expose it to reality.  If you have an insight, it usually means that you have an idea that breaks from the conventional wisdom.  Fortunately, many types of marketing offer the ability to test insight vs. conventional wisdom.
  
  In the case of our industrial coatings insight, let's test out our insight in email.  (Wow, an email about industrial coatings!  This must be how Robert Downey Jr lives!).  We could create a version of the email with a link to tech specs for the electrical engineer another with a link to the MSDS for the chemical engineers and a generic version with a link to a brochure.  In turn, we can split up the electrical and chemical engineer segments into control and test segments and...blah blah blah.  If you don't know how to do this yet, here's a good primer.
  
• Make sure you read the right results.  Direct response media, especially online direct response media, have lots of numbers that go up and down.  However, not all of them matter.
  
  With our breathtaking industrial coatings example, we'd sure want to know whether the engineers preferred the tech specs, MSDS or brochure, of course.  However, we'd also like to know whether they followed up.  This follow up might include a call-to-action on the downloaded material, response to a subsequent outbound call or even simply increased interest in follow-on emails.  What matters is that we calculate which insight--if any--provided meaningful results.

It doesn't take a Nathan Detroit to make good guesses with marketing data.  But it does reward practice.  Fortunately, you don't have to get it right all at once.  Every gamble you take gives you more feedback on how well your guesses work.  So now that you've learned how to play with danger, game on!