Consumer centric analytics
Consumer centric analytics.
A lot of money is about to be spent convincing you that a 360 degree leveraged of the consumer can be constructed using scrapped data sources. The clickstream paradigm isn’t consumer centric analytics. I’ve said it before. In this post, we’ll look at a problem-solution-opportunity set.
The Problem Set
There’s a lot of counting going on. The counting of views through the iPad versus views through Facebook versus views through the work computer versus views through the home laptop. There are pockets of some pretty good usability analysis, some very good optimization, and, we’re finally getting some real statistical rigor into digital analytics in a few places. It’s great to see. Better information ought to be causing better decisions in several organizations.
Counting clickstream events isn’t consumer centric analytics.
The clickstream is a transactional record of what peripherals associated with a session did over time. For a very long time, people used three primary peripherals to interact with digital mediums.
- The earliest mainstream one is the game controller
- The mouse
- The keyboard
Recently, they’ve been joined by:
- the motion sensor
- and the touchpad
Digital analysts used to think in terms of the browser session. Now we have app sessions, as defined period of time associated with a distinct browser. What did those peripherals interact with the medium?
All of that information forms a transactional record. Different types of instrumentation return data at different fidelity. Adobe Flash used to be a black hole for a few years, because so much web analytics infrastructure was based around the pageload transaction based on sessions.
Many of us were on two devices in the period between 1997 and 2012. We had a work laptop and a home laptop. Web analysts would some variant of time-part analysis, or time-series analysis, and there was a period where differentiated usability was in vogue. Then traditional market research kicked it and suddenly we understood that we were dealing with the same person.
Survey research had the benefit gathering a large number of features (m) of personally identifiable people (n). Survey research had the bias of asking people what they remember thinking and feeling.
Clickstream analysis had a very large number of features (m) about a large number of sessions (n).
The market research generated a whole series of managerial heuristics, traditionally updated once every four years, and typically timed along the massive website redesign cycle. There would be a few very good business cases for digital to be built out of the analytics.
Device and peripheral analytics has certainly caused (m) to expand. The number of people continues to grow at a population growth rates. The number of sessions have likely increased as well.
The Solution Sets
The utility of data, instrumentation, definitions and models is only as useful as the decisions they’re seeking to drive.
What are the managerial implications of having a 360 degree view of a single personally identifiable person (a distinct n) across all their sessions across multiple devices (an expanded nxm)?
The first, and probably the most important for many, are the political business cases that can come out of that data. If you’re the head of traditional digital, and you see a small team developing around mobile, you want to bust down that silo. For the headcount. For the synergy. Evidence are great bullets for the civil wars that people play.
Second, think of the insights. (Go ahead, just think about them!).
However, I question if a complete record, a (n x (n x m)), would be as easy to mine. It isn’t just eyballing the data.
***
I’m Christopher Berry.
My answer is Authintic.