Fixing the incomplete profile problem: How NET2GRID empowers utilities and users alike

Recognizing the inherent limits of profile-based disaggregation, NET2GRID has introduced a Profile Correction feature

For years, energy disaggregation has promised a new level of intelligence for utilities: identifying which appliances consume how much energy in a household and when. In practice, however, the accuracy of these insights has depended heavily on customer-provided data—inputs like number of residents, home size, appliance inventory, or lifestyle preferences. But in the real world, most users don’t—or won’t—fill out a profile in an app or portal. And those who do may forget to update it as life circumstances change.

 A new energy-efficient dishwasher, the recent purchase of a heat pump, or the installation of a home EV charger can drastically alter a household’s consumption pattern—rendering the original customer profile obsolete. For utilities and retailers relying on these profiles to tailor services or deliver personalized insights, this mismatch can dilute value and customer trust.

Moving beyond static profiles

Traditional energy apps and insights platforms often rely on a "profile-first" logic. They ask users to enter lifestyle and home details, which then inform the energy model used to disaggregate consumption. In theory, this provides a useful foundation. But in reality, profile completeness is a rarity. Utilities routinely report low response rates to these prompts, and when users do engage, the data quality can vary widely.

Another challenge arises from household evolution. Energy habits are not static—appliances are replaced, usage patterns shift, new residents move in. A fixed profile, even one that was accurate at the outset, quickly becomes outdated. And without active user engagement, these changes go unnoticed.

This is where NET2GRID has taken a meaningful step forward.

Automatic profile correction for low-resolution data

Recognizing the inherent limits of profile-based disaggregation, NET2GRID has introduced a Profile Correction feature that works even when the underlying data is low resolution (e.g., hourly smart meter data). The approach dynamically corrects household profiles based on actual electricity usage, without requiring user input. Through algorithmic detection, NET2GRID finds the presence of certain energy-intensive Distributed Energy Resources (DERs) like PVs, EVs and HVAC. Then, the utility could either receive the corrected information in a B2B report, or NET2GRID can automatically change the profile on behalf of the end user. 

This feature is mostly valuable for customers where the profile information is not available at all or those who have a personalized profile, but it’s outdated.

The automatic profile correction approach leads to significantly improved disaggregation accuracy, even for users who have never filled out a single detail in their app.

The benefits for utilities

This innovation is especially relevant for large-scale rollouts, where expecting personalized profiles for millions of customers is impractical. EDP, one of the biggest energy retailers in Europe, has been using the profile correction feature, improving the disaggregation quality by up to 40% for the user base where no profiles are available. Disaggregation categories like space heating, EV, PV and AlwaysOn were further drastically improved. But energy retailers and utilities operating across deregulated and regulated markets alike, the Profile Correction feature opens up new possibilities:

• Scalability without friction: Deploy disaggregation at scale without requiring profile setup campaigns.

• Improved personalization: Deliver appliance-level insights that actually reflect current household consumption, not outdated assumptions.

• Customer trust: Reduce user frustration by increasing the accuracy of appliance-level feedback, even for disengaged users.

  
  

Why this matters for the energy transition

As the energy ecosystem shifts toward electrification and decentralization, understanding consumption at a granular level becomes not just useful but necessary. 

NET2GRID’s approach acknowledges that while user data is helpful, it cannot be the sole input. By designing systems that adapt to real-world behavior and not just declared behavior they help close the gap between insight potential and insight delivery.

In a sector where customer trust, regulatory compliance, and operational efficiency must coexist, automatic profile correction offers a pragmatic, data-driven path forward.