Skip to content
Pathline FoundryPathline Foundry
Projects and results

Load forecasting for a natural gas marketer

Client
A natural gas marketer
My role
I co-founded smartMFG Solutions, served as its CEO, and built FLOCK.
Timeline and attribution
smartMFG Solutions (2023 to 2025)
Scope
Load forecasting | Nomination advisement | Software development | Performance tracking
Evidence label: Measured Result
33%
lower penalty cost per customer, against the three years before
$170,000
lower penalty cost over the 2024 to 2025 year than the baseline rate, normalized for customer count
30 days
from delivery to daily use in the nomination workflow

How the work ran

  1. Records

    Daily usage and nomination records for every customer

  2. Forecast

    Customer-level forecasts, one for each customer

  3. Recommend

    Seven-day recommendations, built into the nomination workflow

  4. Decide

    The scheduling team is free to adjust any recommendation

  5. Track

    30-day trailing performance, tracked inside the tool

The situation

Facility and context
A natural gas marketer nominating daily gas volumes for hundreds of transport-service customers.
Problem or decision
When a customer's actual use differed from the nominated volume by more than 5%, the marketer paid a penalty on the excess. Earlier forecasting tools, including one from a third-party provider, had not solved it.
Baseline and data used
Daily usage and nomination records for every customer. Over the three years before the work, 2021 to 2024, penalized deviation averaged 25.6 dekatherms per customer per day, about $2.05 per customer per day.
Evidence label: Illustrative Example

How a day became a penalty, and what closer forecasts change

With earlier forecasts

Day 3: outside the 5% band, penalized on the excessDay 5: outside the 5% band, penalized on the excessDay 9: outside the 5% band, penalized on the excessDay 12: outside the 5% band, penalized on the excessDay 13: outside the 5% band, penalized on the excessDay 17: outside the 5% band, penalized on the excessDay 20: outside the 5% band, penalized on the excessDay 23: outside the 5% band, penalized on the excessDay 27: outside the 5% band, penalized on the excessDay 28: outside the 5% band, penalized on the excess

With customer-level forecasts

Day 13: outside the 5% band, penalized on the excessDay 20: outside the 5% band, penalized on the excess
  • Nominated volume
  • 5% band
  • Actual daily use
  • Penalized excess
Made-up data, not the marketer's, on the same scale in both panels. A day's use more than 5% away from its nomination was penalized on the excess. Closer forecasts keep more days inside the band, which is what brings the penalty cost down. The panels show the shape of that change, not its size.

Approach

Finding
Customer-level forecasts could cut penalized deviation roughly in half against the three years before.
Recommended action
FLOCK, a forecasting and advisement tool giving seven-day recommendations for each customer, built into the daily nomination workflow, with the scheduling team free to adjust any recommendation.
Investment and incentive
Not applicable: a software engagement, with no incentive program involved.

Results

Expected or verified value
Over the 2024 to 2025 year, realized penalty costs fell about $170,000 against the 2021 to 2024 baseline, normalized for customer count: a 33% reduction per customer. Following every recommendation without adjustment would have cut penalties by about 49%.

Penalty cost per customer

Baseline
33% lower
49% lower

2021 to 2024 baseline

2024 to 2025, actual

2024 to 2025, every recommendation followed

Modeled
Indexed to the 2021 to 2024 baseline, so the baseline is 100. Lower is better. The marketer was free to adjust each recommendation; the last column is what following them exactly would have produced. From smartMFG Solutions' FLOCK white paper, April 2025.

Impact

Implementation status
In daily use within 30 days, with 30-day trailing performance tracked inside the tool.
What happened next
The same discipline, a forecast measured against a stated baseline, is what Pathline brings to facility forecasting and monitoring.

Client described only as a natural gas marketer. Figures from smartMFG Solutions' FLOCK white paper, April 2025.

Performance services

Other projects and results

Building performance compliance for Washington school districts

West Ridge Energy (2025 to present)

Evidence label: Identified Savings

Which buildings the standard covers, how far each covered building sits from its energy target, and the measures that can close the gap. Of the ten Tier 1 buildings, four were above target, by 6.1% to 34.5%.

Read the case

Energy benchmark for an Idaho printing plant

Pathline Foundry (2026)

Evidence label: Identified Savings

The morning start-up is likely the biggest lever: the highest 15 minutes of the month sets the demand charge, and demand charges are estimated at 38% to 52% of the bill. The plant also sells more per square foot than 76% of its peers, so a comparison on floor area alone would understate it.

Read the case

Energy assessments for regional manufacturers

DOE Industrial Assessment Center, University of Utah (2019 to 2025)

Evidence label: Identified Savings

Each assessment identified a set of measures worth 7.2% to 26.5% of the plant's annual energy cost, 13.3% on average.

Read the case

Working on a similar facility?

Start with the facility, requirement, or decision in front of you, and I'll tell you what deserves a closer look.

Discuss a similar facility