Client
A top-six UK energy retailer
Sector
Energy & Utilities
Engagement
Strategy, delivery and embedded engineering - multi-quarter programme.
The challenge
What the client needed
The client had grown rapidly through acquisition and was running smart-metering data across four different storage systems, with inconsistent quality, no end-to-end lineage and a ten-day delay between meter reading and customer-facing insight. Regulatory reporting was manual, error-prone and increasingly unsustainable as the rollout of half-hourly settlement approached.
Our approach
How we worked
- Ran a six-week discovery to map every consumer of smart-metering data - internal teams, external regulators and customer-facing products - and quantify the impact of the existing latency and quality issues.
- Designed a target architecture based on a lakehouse pattern with open table formats, separating raw, conformed and product-ready zones with explicit data contracts between them.
- Stood up an embedded squad of senior data engineers, SREs and a product manager working alongside the client’s in-house team, using paired delivery from day one.
- Replaced manual regulatory extracts with an automated reporting service driven from the conformed zone, with full lineage back to source meter readings.
- Built a customer-facing API layer that powered three new products in the first 12 months, including a personal-usage app feature with measurable retention impact.
Outcomes
Measured results
All figures verified with the client. Specific identifiers withheld in line with our standard confidentiality terms.
- End-to-end latency from meter reading to customer-facing insight reduced from 10 days to under 30 minutes.
- Regulatory reporting cycle time reduced by 80% with zero significant data quality findings in the following two reporting periods.
- Three new customer-facing products launched on the platform within 12 months, with combined attributable margin improvement in the low single-digit millions.
- Total cost of ownership across the previous four data stores reduced by 42% within 18 months, after decommissioning.
- Internal data engineering team transitioned to full ownership of the platform within nine months.
“We have tried this twice before. The difference this time was the engineering depth on the ground and the discipline of the data contracts approach - both of which Halfteck brought from day one.”
Working on something similar?
If this engagement looks like the kind of problem you are facing, we would be glad to compare notes by email.