User experience on a website, a mobile app, or an e-commerce domain is one of those stories. But if you’re trying to determine from page hits alone whether your website is a hit, you’re going to be hitting your head on a brick wall. There is often no clear line-of-sight between browsing behavior and what the browsing individual may be thinking or feeling at any point in time.
Behavioral data is rarely a transparent window on browsing experience. Web analytics applications might have access to every possible type of experience-relevant browsing data from client and server applications and still get it wrong. That’s because point of view is everything when analyzing the browsing experience, and Web browsing data may not testify in any way to those perspectives.
It’s a bit like the experience of a family on vacation. To the parents, it may feel much like work, involving such steps as making hotel and travel reservations, packing suitcases, studying the places they plan to visit, having newspapers and mail suspended, getting there, keeping the kids from fighting in the back seat, and so on. But to the kids, it may all be pure fun and leisure, or (gasp!) a perfunctory grind of “are we having fun yet?” From the point of view of a third party documenting their vacation-going behavior, it may be difficult to distinguish a family that’s truly enjoying themselves from one that’s merely “going through the motions.”
In the broader sense, journey modeling is an essential tool when you’re using Web analytics data to assess whether users are having fun yet. Your data scientists should be identifying the journey variables that drive some customers to merely browse one site aimlessly while others click from site to site with specific goals and intentions. Data on specific Web users, sessions, and “hits,” as discussed in this recent article, is mere fodder in this exploratory process.
The cited article implies a journey model of Web browsing behaviors without explicitly referring to it. “Hit” data–pageviews, screenviews, events, transactions, social interactions, etc.–can say nothing meaningful outside the context of a modeled journey that associates these behaviors with specific sentiments and intentions. In addition, sessions (i.e., collections of hits from the same users) only derive their meaning from the contextual variables–such as user time on site or user time to view stream–that are most relevant to a standard journey as modeled.
And even the concept of “user” itself–a specific (possibly anonymized) individual with whom specific Web sessions and hits are associated–must be defined within the context of the journey. That’s because the “user” of interest may simply be someone who uses a standard browser for typical interactions with Web pages. Or they may be someone who employs browsers, smartphones, set-top boxes, and other clients to view Web pages, conduct mobile commerce, and/or consume streaming media across a wide range of sites. In either scenario, the “user” must be defined crisply so that the relevant Web analytics data can be gathered and correlated. This would enable analysts to identify the specific hits and sessions associated with specific journeys.
All of this contextualization is merely upfront spadework for your data scientists. What they’re trying to do is whip your messy Web analytics data into the proper shape. Doing so can reveal insights about what users might be experiencing as they click on this object, download that, and stream the other.
Without sound journey modeling, your increasingly diverse repositories of clickstream, sentiment, and other Web behavioral data will be more of a storage burden than a big-data business asset.