The Dependent Variable

Posted on by Christopher Berry

The dependent variable is the one that matters.

You can’t explain why without Y.

 

model2

Last week, I broadcasted 10 questions on Twitter and Facebook what they thought of dependent variables.

It’s what 18 smart managers, researchers, developers, strategists, planners, data scientists, executives, eComm marketers, direct marketers, and department heads thought. I thank them for their time. Thank you.

Take the n and the instrument for what it is – a dipstick check on just how aligned we are.

Here’s what was found.

H1 >=80% will state that Conversion is the Dependent Variable

Confirmed.

Conversion won every matchup it was in.

Number of conversions won head to head against impressions and engagement; engagement and unique visitors; impressions and unique impressions.

Conversion crushed with over 80% of respondents.

H2 >=80% will align their dependent variable preferences with the Impressions<Unique Visitors<Engagement<Conversion orthodox funnel

Confirmed.

The pattern of preference, in aggregate, in excess of 80% of respondents was:

  • Impressions<Unique Visitors<Engagement<Conversion

H3 <=80% will agree with the orthodox funnel when bucket statements are replaced with specific metrics

Rejected.

The specificity of metrics killed the unanimity of results.

Respondents were neatly split between whether Retweets or Unique Visitors was the dependent variable. (When up against Paid Impressions)

The matchup between impressions, URL Clicks, and Percentage of Target Audience Reached also resulted in a split. Nobody said that impressions were the DV. Opinion was evenly split between URL Clicks and Percentage of Target Audience Reached.

  • Paid Impressions<Retweets=Unique Visitors
  • Paid Impressions<Promoted Embed Clicks<URL Clicks
  • Promoted Hashtag Clicks = Percentage Target Audience Reached < Email Signups
  • Impressions = 0; URL Clicks = Percentage of Target Audience Reached

H4 <20% will state that engagement or impression metrics are Conversion metrics

Rejected.

  • 88% stated that Email Signups constituted a conversion.
  • 55% a URL Click
  • 50% ReTweets
  • 22% Promoted Embed Clicks
  • 16% indicated that Unique Visitors, Promoted Hashtag Clicks, and Percentage of Audience reached were all conversion variables.

Synthesis

In theory, we’re aligned.

In practicality, we’re not.

Specificity destroys unanimity.

Objectively, the traditional digital funnel is intuitive. All variants on that funnel (The Fish, as imagined in The Open Brand, the Forrester Consumer Journey) were efforts to reduce linearity and inevitability in the root causal model.

So, we might all conceptually agree that conversion is the most important DV, but, if there’s disagreement about what constitutes a conversion, there can be meaningless conflicts amongst people about which mental models make the most sense at operationalization.

This isn’t technically a problem so long as the DV is specified and defined. Once again, the rule of clarity and specificity of language within the model design will pay huge dividends.

It’s simply good to be aware that not everybody thinks the same way when a model needs to be defined.

Tell me what you thought about it on Twitter; or blog about it and let’s talk.