Thinklytics

Municipal Electric Utility · Energy & Utilities · Phoenix, AZ · 11 weeks

Smart meter data foundation built for 680K meters

A municipal electric utility installed 680,000 smart meters but only used the data for billing. The interval data sat unused. We developed an analytics system to process this data and launched the utility’s first demand response program. This directly cut capacity costs by $3.4 million in the first summer.

Challenge

The utility collected 15-minute data from 680,000 smart meters, but the data sat locked in a vendor’s system with no way to analyze it. Meanwhile, they spent $2.1 million each year on capacity reserves that a demand response program could reduce. The problem: they couldn’t pinpoint which customers to target or track how well the program worked.

Approach

We extracted interval data directly from the vendor system to create a solid smart meter data foundation. Then, we developed a load profile analytics layer to analyze customer energy use. Using load shape, peak contribution, and program eligibility, we identified the 42,000 customers with the highest demand response potential. Next, we built an analytics platform that tracks enrollment, dispatch, and curtailment performance in real time to monitor demand response program effectiveness.

Outcome

In its first year, the demand response program signed up 18,400 customers and cut capacity costs by $3.4 million. Building on smart meter data, we developed four more use cases: spotting non-technical losses, optimizing time-of-use rates, detecting outages, and targeting energy efficiency efforts. Early analysis of non-technical losses uncovered $680,000 in annual revenue leakage.

We dove into data everyone else ignored

So get this, the utility spent $180 million on smart meters and barely used them beyond sending out bills. Wild, huh? These meters collect a ton of data that can reveal how folks use power, catch non-technical losses, spot outages early, and even help with smarter pricing. We rolled up our sleeves, built some killer analytics tools to dig into that data, and shared the insights with multiple teams all at once. It was like unlocking a hidden goldmine and spreading the wealth across the whole company.

How We Found the Perfect Candidates for Demand Response Programs

We dove into load data from 680,000 customers to see who was driving peak demand and who might be able to shift their usage. Then, we gave each customer a score based on how much they affected demand and whether they qualified. From there, we zoomed in on 42,000 customers with the most flexibility. Focusing on this group helped us boost enrollment to 44 of every 100 in the first year.

Results

  • $3.4M Avoided capacity costs in first summer
  • 18,400 Demand response program enrollments in year one
  • $680K Annual revenue leakage identified through loss detection
  • 4 Additional analytics use cases enabled by the foundation

We spent $180 million on smart meters but were only using the data for billing. Thinklytics helped us create an analytics system that lets us do more with that data, like demand response, detecting non-technical losses, and a few other things. Just this past summer, we saved $3.4 million by avoiding capacity costs, which was better than we expected.

Director of Customer Programs, Municipal Electric Utility

In the client's terms

The question

Where does this number come from, and can the path be trusted?

The situation

Six hundred and eighty thousand smart meters produced 15-minute interval data that sat inside the vendor's system with no route to analysis. Meanwhile $2.1M a year went on capacity reserves that a demand response programme could have reduced, with no way to identify which customers to target.

What we found

The utility already owned the data that would have answered the question. It was not a measurement gap, it was an access gap, and the demand response programme had been treated as impossible when it was only unmeasured.

What it produced

The programme enrolled 18,400 customers in year one and avoided $3.4M in capacity cost over the first summer. The foundation enabled four further use cases, and early non-technical loss analysis found $680K a year in revenue leakage.

The roadmap

Interval data was extracted from the vendor system into a foundation the utility controls. A load profile layer on top analysed consumption by shape, peak contribution and eligibility, identifying the 42,000 customers with the highest demand response potential. A tracking platform then follows enrolment, dispatch and curtailment live.

The work behind this result

Related case studies

Frequently asked questions

What can smart meter interval data be used for beyond billing?

Load profiling, demand response targeting, non-technical loss detection, outage detection and rate design. This municipal utility had 680,000 meters producing 15-minute data that sat inside the vendor system with no route to analysis.

How do you target a demand response programme?

Score for flexibility rather than size. Analysing load across all 680,000 customers here identified the 42,000 with real peak contribution and shift capability, which produced first-year enrolment of 18,400 and avoided $3.4M in capacity cost over the first summer.

What is non-technical loss and how do you find it?

Consumption that is delivered but not billed. Early analysis on the meter data foundation here identified $680K a year in revenue leakage, one of four further use cases the foundation enabled after the demand response programme.