“Dark Data” & “Interactive Effects Theory” in Behavioral Science

 How engaged you are ?

In general there is a lot of discussion lately for the term “engagement”. What is actually the core difference with the term “loyalty” ? there are many articles defining the difference and some people point out that engagement is a metric of loyalty or engagement leads to loyalty… I kind of agree with most of them since they are related with revenues, product adoption, etc; however nobody puts the term human behavior in the centre of the whole macro-system: Human – Technology Interaction, HTI

To shape any user’s behavior using any modern digital service (app, online web, social, etc we live in the Digital Transformation era) initially some simple demographic and contextual data is required. However, if we want to approach engagement as it should be, we have to take the analysis one step further, and among the previous ones, we have to merge psychographics, sentiment, training and specific behavioral data that are not visible to the simple eye including their correlations

Basic queries to be answered are “Do users follow a particular pattern in the product/service ?”, “Why they are doing so ? ”, “What interaction they usually do ?”, “How results are determined by their initial actions and what exact actions are these ?” and “What is the impact of their behavior on simple daily actions they do”. In other words, accurate correlated behavioral data tells us not only what is happening, but also how and why it is happening.

Connecting the dots is the key, even with metrics that are not visible or measurable to the naked eye

Based on my research all these years, I refer here to two very important terms that we are lacking in almost all engagement/loyalty or customer experience reports and analytics procedures: Dark Data and Interactive Effects and of course the use of a complete Engagement Framework that will put all functions on a procedural map and orchestrate the whole system to follow the consumer journey, which is a multidimensional random variable

 

Dark Data

What I mean by the term Dark Data ?

Dark Data in Data and Behavioral Science are these KPIs/metrics or even raw data that are not easily measurable directly (via sensors or other devices) and they may exist only if we trigger the human system under specific conditions and for a small period of time

By using an app, I can analyse the clicks, pageviews and other directly measurable indices and derive patterns that will drive my UI, my app functionality or even understand my customer; however I cannot measure the “learning curve” or the “learning speed” of the specific user nor his sentimental reactions during his digital experience or what other friends are telling him that specific time about the digital service…not until I run a quiz, a test, a gamification approach, a survey or apply other innovations and analyze time responses, knowledge curves and other variables…this is Dark Data and they do play an essential role on human decision making that drives a part of engagement

Another example is to discriminate in a class the good students, in terms of grades and good behavior…I cannot measure that until I run a test and check if they will try to cheat…this is a wrong behavior that tells me a lot about his character but do not know until I trigger the system and do not relay only on the final grade

There are ways of measuring or creating on-the-fly additional data and metrics that tune the human behavioral model and reveal Dark Data…

so triggering the system, creating additional metrics and measuring dark data on a human-digital relation, drives better the engagement models

 

Interactive Effects Theory

A human behavior is comprised of actions and reactions. This can me modeled as a chain, or even better as a Markov Chain, where its state is only related to its previous state and in each state human can take only one decision. In modern behavioral science and engagement, we should analyze these Interactive Effects

For example, if I want to shift your energy behavior and make you save energy, then I could educate you and change your cooking habits. To do that, instead of sending you messages and tips, I could propose you a specific diet or I could offer food incentives that could shift your energy, by eating healthy food

This is an example of an interactive effect. To model that, you have to create a connected semantic model, identify the states (action, reaction, result), deploy a Markov chain with many parallel options and start playing with probabilities and then tune the model using the actual data you are getting from the digital service by following the human actions. We do that and it works !!

 

Engagement Framework

All the above points (and many more) should be on a unified map, should work together and should be correlated. Thus, we need a complete Engagement Framework, that will orchestrate the engagement analytics, the behavioral incentive models, the personalization, the dark data sensing, the interactive effects and drive actions and reactions on a timeline, in order to validate the engagement model and to achieve actual, measurable engagement metrics !

 

I will refer to all the above, I will explain how we use these secrets in our Utility Engagement Services with case results (some great numbers) from our DiG SaaS Engagement Platform (http://intelen.com/us/solutions/dig.html) in my upcoming talk about engagement and gamification at the European Utility Week 2015 in Vienna (http://www.european-utility-week.com/).

Join me there ! @ http://programme.european-utility-week.com/hub-sessions/smart-homes-end-user-engagement/session-47-understanding-consumer-journey/impact-game