Page 179 - Social Media Marketing

analytics related to or indicative of interactions originating on the Social Web?” The
connections between the two are found by linking the sources of Social Web conversa-
tions, influencers, and the incoming traffic to your website, support forum, or commu-
nity application.

To see how the data collected through a Social Web listening program is con-
nected to the data collected in a web analytics program, consider the application of
basic correlation. Figure€6.6 shows a visualization of correlation. The data represented
by Data Set 1 (solid squares) appears to be correlated—note how it falls along a com-
mon line—whereas there appears to be no relation in the data compared between Data
Set 2 (depicted by the dots), whose data is scattered all over the chart. Correlation is
the study of sets of observations, with the end goal being the deduction of a possible
connection between the sets of data. There is a big caveat here, and it’s an important
one: Correlation and causation are two different things.

157

Likelihood of ‘A’ happening Data Set #1 ■ ╇ W eb A nalytics

Data Set #2

Likelihood of ‘B’ happening

Figure€6.6╇â•A‰ Visual Chart of Correlation

Correlation is the more general of the two. Correlation implies that when-
ever A is observed, B is also observed. An example of correlation is the observation
that “People who bought dog food on the Petco website also bought dog leashes.”
Purchasing one does not in itself cause the purchase of the other. However, the two
tend to occur together more often than, say, people buying dog food and a bowling
ball, something that may well be observed on occasion in the checkout line at a physi-
cal department store like Walmart.

Causation, by comparison, is a direct cause-effect relationship between A and
B. When A happens, B results. Acquiring a dog, for example, fairly certainly causes the
purchase of both the dog food and the dog bowl. Cause and effect can be seen in things
like the increased checkout basket size that results when online shoppers are prompted
to consider additional items based on what they have currently placed in their shopping
carts: When people are told “others who bought dog bowls also bought dog leashes…,”
they tend to go look for a new leash.