Learn more about how you can integrate Bidgely’s AI-powered customer profiling and segmentation capability into your operations to generate greater ROI from your AMI investment by downloading our Data Science Lab’s AI-Powered Customer Segmentation for Utilities white paper.

Transforming Customer Segmentation with the Power of AI 

Historically, one of the greatest challenges to successful demand side management (DSM) and other customer-centric programs has been that when it comes to making better energy decisions, there is no universal motivator or set of rules that applies to all customers. Traditional customer segmentation techniques helped incrementally, but with so much variation in customer personas and energy habits, even segment-tailored DSM programs and outreach sometimes miss the mark. Conventional tools such as mass surveys, focus groups and manual utility population data collection take time, are static, lack granularity, don’t update consistently and fail to account for behavioral and lifestyle aspects in their models. 

A Revolutionary Data-Driven Approach to Customer Segmentation 

Thankfully, AI-powered customer segmentation has opened up a new era of far more effective and impactful DSM program design and execution.

Bidgely analyzes raw energy consumption data obtained from household electrical meters at varied intervals to better inform rate plan design, in-depth customer profiling by power usage, and more.  The UtilityAI™ platform is uniquely capable of tapping into essential attributes that describe people’s behavior, lifestyle, income and other characteristics using sophisticated machine learning and statistical solutions to build a 360-degree profile of every customer. These hyper-personalized customer profiles enable the creation of  customer segments that group individuals based upon highly nuanced similarities and maximize the contrast between segments. 

With AI-powered customer segmentation, the substance of DSM initiatives is transformed from generalities into precise and profoundly relevant energy saving advice that more effectively influences behavior.  

For example, a customer who goes to the office daily should switch off lights and other appliances during working hours when they are not at home. However, for those customers who stay at home during the day, ongoing consumption patterns emerge, including those associated with daytime space heating, entertainment usage and lighting. AI-generated insights and data-driven energy use measurements associated with these daytime habits allow utilities to make a distinction between customers by lifestyle, and strategically apply different treatments to each group. 

The Science Behind Customer Profiling

Out of all available customer characteristics, Bidgely’s segmentation framework intelligently decides on the choice and hierarchy of customer attributes from data sources including:

  • Customer Raw Energy Data – Household members’ appliance usage, daily energy behaviors and choices. Variations in daily consumption provide fine details on a consumer’s usage preferences. For example, someone who is considered a night owl, with most of their consumption happening at night, will have a different potential to meet a utility requirement compared to an early bird who exhibits peak usage in the mornings.
  • Lifestyle Attributes – Details about activity at various hours, household occupancy, weekend and sleep habits and more that differentiate a household’s electric usage and needs. Bidgely employs a robust machine learning approach to extract the aspects of daily electric activity like cooking, entertainment, laundry and lighting on a weekday vs. weekend basis.
  • Behavioral Attributes – Greater detail about the occupancy and characteristics of those living in the household based on sampling-rate-level analysis. This granular level of analysis creates distinct and highly accurate customer segments.
  • Appliance Attributes – The disaggregation-based detection of major appliances like electric vehicles, HVAC systems and solar panels. Highly advanced computer vision, machine learning, statistical and signal processing techniques estimate the consumption of each appliance at the timestamp level. 

Grouping Customers into Meaningful Clusters

Customers are grouped into segments based on 1) the utility’s objectives; 2) customer profiles; and 3) data availability. Decisions about which inputs to prioritize represent one of the most crucial steps of the segmentation process.

Bidgely determines the ranking of inputs by combining domain knowledge with data-driven statistical inferences while also considering utility objectives, such as the desire to compare household monthly energy consumption.

Empowering Diverse Use Cases

  1. Customer Targeting  – Bidgely provides an exhaustive set of customer characteristics to facilitate customer-targeted actions, such as the time of the day the customer is most likely to be present at home to improve utility outreach, amplify conversion rates and lower acquisition costs.
  2. Appliance Propensity – With definitive information about customer’s lifestyle parameters like occupation, income level, family type, neighborhood affluence, consumption level and weekend behavior that influence energy product purchase decisions, utilities are able to implement advertising campaigns that achieve higher conversion rates and realize greater overall marketplace success.
  3. Rate Plan Design – Bidgely’s customer profiling solution provides details about a customer’s consumption level, peak usage months, and peak usage hours which assist in clustering, and developing relevant personalized rate plans for individual segments of customers. 
  4. Demand Response – Bidgely’s AI-driven segmentation framework segments customers based on their data-revealed inclinations to take the required action, allowing utilities to focus on the subset of customers most willing to take action in order to achieve the desired reduction in peak demand.
  5. Low and Medium Income Programs – Bidgely estimates customer income levels using a combination of derived features such as geography and dwelling size and incorporating credit score information and payment history to refine the algorithm.
  6. Similar Home Comparison – Household comparisons are most meaningful and accurate when supported with evidence that is logical to and easy for customers to interpret. Bidgely’s AI-powered customer-profiling is more accurate and therefore enables superior similar home comparisons. 

A Dynamic Segmentation Approach

AI-powered customer profiling captures essential aspects of a customer’s lifestyle over time, and reflects the variation in customer behavior or occupancy at different points during the year. Customer profiles reflect current household conditions and how they have changed from one month to the next, including the impact of unexpected environmental and societal events.  

Advancing Your Customer Segmentation with AI

Learn more about how you can integrate Bidgely’s AI-powered customer profiling and segmentation capability into your operations to generate greater ROI from your AMI investment by downloading our Data Science Lab’s AI-Powered Customer Segmentation for Utilities white paper.

Categories: Bidgely