Top results
Reduced daily Snowflake compute costs by 55% through business-logic optimization, incremental processing patterns, and standardized transformation execution aligned to Snowflake’s consumption model.
Accelerated medallion-layer development across Bronze, Silver, and Gold by replacing hand-written Python and SQL with standardized node types and reusable templates. Work estimated at four weeks of development was completed in roughly one week. Building the Bronze layer by hand would have required a full month of dedicated effort.
Reallocated more than 50% of engineering time to high-impact projects, as standardized transformations, built-in lineage, and observability significantly reduced maintenance and firefighting overhead.
Improved data freshness from multi-hour refresh cycles to hourly updates, with 15-minute refreshes planned for select use cases as transformation efficiency and cost controls continue to improve.
Pain points
Pipeline reliability issues at production scale. Intermittent failures required constant monitoring and manual intervention, occasionally resulting in partially populated or empty downstream tables.
Engineering capacity consumed by maintenance. More than half of the data team’s time was spent addressing data leakage, stabilizing pipelines, and rebuilding stakeholder trust. The small data team spent the majority of its time keeping systems running rather than building new capabilities for a rapidly growing business.
Manual transformation development and lack of governance. Hand-written SQL and Python logic slowed delivery and increased the risk of errors as datasets, dependencies, and downstream consumers multiplied.
Lagging data freshness. Data freshness couldn’t keep pace with business needs, leading stakeholders to either pull data directly from source systems or make decisions based on stale information.
“The value add of Coalesce can be felt within weeks.”
The challenge
Stanley Martin had already adopted Snowflake, but the surrounding architecture built on legacy orchestration patterns and fragmented transformation logic became strained as the business scaled. As workloads grew, the architecture introduced latency, inconsistency, and rising operational risk.
Transformation logic was distributed across notebooks, scripts, and BI tools. Pipelines broke frequently and data quality checks often occurred downstream. Under growing data volumes and higher refresh demand, the Snowflake compute costs increased without corresponding gains in freshness or reliability.
To support aggressive growth targets and new AI initiatives, the team needed a Snowflake-centric transformation layer that delivered reliability, observability, and faster development at scale without becoming opaque or difficult to debug. That requirement led to a focused evaluation of modern data transformation platforms, including Coalesce.
“As a data engineer, it’s really important that I know exactly what’s happening under the hood.”
The evaluation
Stanley Martin evaluated transformation platforms with a clear set of architectural requirements. The team sought a Snowflake-native platform that reduced Python and SQL scripts while preserving code visibility, enforced medallion-style patterns, and addressed both maintenance overhead and compute efficiency.
After reviewing alternatives, Stanley Martin selected Coalesce for its ability to act as a transformation control plane rather than a black-box abstraction. Built with deep Snowflake integration, Coalesce allows engineers to retain full access to generated SQL while using standardized node types and templates to enforce consistent, cost-aware transformation patterns. The platform met Stanley Martin’s needs without introducing the added complexity of more generic, overbuilt solutions.
“Coalesce stood out because it was natively integrated with a lot of the operations within Snowflake.”
The migration
Stanley Martin partnered with Coalesce experts to implement a governed medallion architecture with standardized transformation patterns and institutionalized best practices such as incremental loading and data quality testing. The team began by rebuilding the Bronze layer for core systems like Microsoft Dynamics and Salesforce inside Coalesce’s visual canvas, then gradually redirected downstream Looker models and other consumers to the new tables.
Using Coalesce node types and templates, the team completed the Bronze layer in roughly a week that previously would have required about a month of manual SQL development. As additional workloads were migrated, transformations were further optimized, enabling the move from multi-hour syncs to hourly refreshes and setting the stage for tackling the largest, multi-billion-row cost analytics jobs.
“Compared to writing all the SQL myself, it probably saved me tenfold the amount of time.”
The impact
Within months, Stanley Martin transformed from a fragile, high-maintenance transformation environment to a governed, Snowflake-centric platform built for scale.
Business teams access reliable data to support operational and strategic decisions. BI developers and data engineers can move faster with consistent patterns and full SQL transparency. Workloads that previously required mult-hour refresh cycles now run hourly in a sustainable, maintainable way.
“Coalesce helps data teams work on projects that deliver the most impact for the business.”
Data engineers build faster with full transparency
Before Coalesce, transformation logic was spread across hand-written SQL, tasks, notebooks, and BI tools like Looker. Engineers had limited visibility into how pieces fit together, lineage was difficult to trace, and many data quality checks ran only after data reached downstream environments. When issues arose, troubleshooting required digging through multiple tools and scripts, which contributed to much of the team’s time being spent on maintenance instead of new development.
With Coalesce, engineers now work in a visual graph that still exposes every line of SQL, allowing them to inspect and adjust code without slowing development. Shared node types and templates enforce consistent patterns across bronze, silver, and gold layers and propagate changes across hundreds of tables. Copilot now reduces effort on some repetitive builds by about half. These gains show up in day-to-day signals as well, with issue-related emails from business users dropping from several per day to roughly one, typically requests for new analytics rather than bug reports.
Business teams access fresher data for better decisions
For business users, the most visible change is trust and alignment with the data team. Previously constrained to multi-hour refresh cycles, sales, operations, and finance teams often worked with data that lagged behind reality. By optimizing transformations in Coalesce and tightening warehouse management, Stanley Martin moved key workloads to hourly refreshes and is now planning 15-minute cycles for select use cases. This improved cadence supports twice-daily rebuilds of massive home cost analytics tables containing billions of rows.
IT leadership scales output without scaling headcount
For IT and data leadership, the impact is visible in both capacity and cost. A lean, six-person enterprise data analytics team now supports a company of more than 1,000 employees, sustaining industry-leading growth without a corresponding increase in engineering headcount. By consolidating transformation logic in Coalesce, optimizing SQL, and right-sizing warehouses, Snowflake compute costs fell by roughly half, even as new data products and analytics use cases came online.
Equally important, the team has reclaimed time to focus on higher-impact initiatives such as market studies, dynamic pricing, and cycle-time analysis across thousands of homes instead of repairing fragile pipelines. Coalesce Catalog is becoming the system of record for business definitions and KPIs, laying the foundation for future conversational AI experiences on Snowflake that will give employees faster, more trusted access to data.