Our SaaS engagement solution (DiG) has an integrated data-driven learning management/content module, over game mechanics, that sends to users personalized content in order to educate them and raise their awareness and knowledge; this has an immediate impact on their actions and their engagement status (this is called subliminal learning).
However is always difficult to verify what specific behavioral event or content has caused shifting in specific behaviors; this is a problem we call "Dark Data" and we are working in our R&D to actually measure that...
One good solution we follow is "extreme" personalization and service adaptation to the users' profiles by getting data-driven feedback from the user's digital profile and adapting more and more the content and the user experience
Timing & Frequency - Analytics on Personalization
The suitable timing and frequency of sent content is crucial as constitutes part of personalization. It is not only that sent tips and quizzes should have personalized content based on user’s needs and preferences.
Another aspect of providing personalization is related to the proper selection of time and frequency according to which content will be sent. This is important for two main interrelated reasons, whose balance indeed matters. The first is to minimize the level of interruption and annoyance the user might feel when notifications of new content arrive in inappropriate timing or too frequently. The second is to keep user engaged and satisfied. The content cannot be too rare while the observed interaction should account as the green light to the system to send new content, increasing so the level of convenience for the user.
Ad hoc analysis is required to determine what will be the optimal time and frequency of contacting each user to deliver either new content or communicating that a new challenge or an offer is available for him/her.
In many of our real customer deployed cases (see above), in our analytics division we are able to correlate real-time user actions, user behavioral feedback, user's digital trace in our systems (Web and mobile) and user's energy curves that give us great insights on the overall user profile and his ability to make actions that lead to results (ie. DR event, Savings, efficient usage of appliances, cross selling initiatives, etc). If you then combine demographics and psycho metrics (probabilistic decision trees) you may be able to predict specific behaviors and thus specific actions that lead to results...
Intelen's DiG SaaS is designed to deliver personalized content based on user’s profile, preferences and needs. To build rich user profiles, DiG collects and considers multiple data sources: demographics, analytics (behavioral data), psychographics/preferences (survey data), energy data, housing data, appliances related data, weather data etc. Then, data analysis techniques undertake to combine all that information in order to cluster users in one or more segments. Afterwards, the content that suits the most to each user, is delivered. Evaluating user’s feedback on the received content is crucial as allows system to update user’s profile and provide more and more personalized content.
At the very beginning, and in order to grab user’s attention and gain user’s satisfaction, DiG delivers content solely based on user’s preferences. A survey suitably designed, collects all data needed to understand user’s preferences and start with personalization.
Basic queries we can answer in the Utility engagement space are:
- “Do users follow a particular pattern in the app”,
- “Why they are doing so"
- “What interaction they usually do”
- “How savings are determined by their actions” and
- “What is the impact of their behavior on energy consumption”.
In other words, data tell us not only what is happening, but also how and why it is happening...
More real Utility engagement cases pretty soon...data-driven human behaviors over the digital dimension will shape things in many industries.