David Hamel, passionately holds high the banner of “the only metric that matters is if somebody buys something”. It’s a hyper-simplistic approach, but I think it serves to remind us that the real dependent variable we’re all trying to explain, in part through the Engagement Wars, is conversion.

Let’s do the Web Analytics thing and define conversion.

“TERM: Conversion
Type: Dimension (same term is also used for count)
Universe: Aggregate, Segmented, Individual

A visitor completing a target action.

This is a method of segmenting behavior as visitors interact with a web property. The
event represents a transition in the visitor state that may indicate 1) potential for
future behavior such as clicking on an advertisement, or registering for more
information, or starting a check out process; or 2) completion of a goal milestone
event such as completing a purchase on-line or requesting a quote. The latter (2) are
sometimes designated as target or goal conversions, and the former (1) are
sometimes referred to as step, support, mini, or micro conversions.

A conversion can be viewed and counted as any other event. In addition, conversions
provide a general framework for segmenting visits or visitors and attributing various
marketing activity and visitor actions to these segments. They provide the marketer
an additional tool for segmenting visitors other than demographics. In practice the
two approaches work together to understand the visitor and their on and off line
behavior with respect to various marketing activities. The best conversions indicate
that a visitor has successfully completed an objective of the site or business.

There might be two flavors of Conversion (possibly more) – for instance, if a user completes an action for the first time, this could be termed an “acquisition conversion”, but perhaps only if the said action forms some kind of lasting impression or relationship. If a user repeats the conversion, it can termed as a “retention conversion” or a “repeat conversion”, depending on the nature of goal.”

Source: WAA Standards Committee Definitions

Now for the “for instance”:

Some websites sole purpose to exist is to sell your eyeballs to other advertisers. For them, a key goal might be to keep you on the site for 5 pages. Since these sites are typically paid in terms of pageviews, this is a logical conversion figure. If a user accomplishes the task of 5 pageviews, the business can consider that the user has achieved conversion.

Another website might actually want you to buy something. In this case, buying something could be considered a conversion.

If we take Hamel’s argument that Conversion is the main dependent variable, isn’t a lot of the fight over a universal definition of ‘engagement’ a little bit futile. After all- we have very good methods for relating a series of website actions to conversion, and most of them are very good interval variables. Linear regression can (and in the cases of the people I work with) has, shed a lot of light on the mysteries of why some parts of the site are more responsible (seemingly) for higher conversion. Focus efforts there.

Should engagement be operationalized in a manner that is specific to a business or an organization?

Yes! It can serve as an important abstraction to readers and stakeholders that functionality and content is absolutely vital in getting people to convert.

Yet, in so doing, let’s make sure we’re really paying attention to the dependent variable that we’re ultimately trying to optimize.