Top results
Reduced ELT tool costs by 50%+: Eliminated licensing fees and decommissioned on-prem servers, delivering immediate savings while modernizing the stack.
Improved data processing speeds by 18-20%: Optimized transformations and cut nightly processing time significantly.
Achieved zero-downtime migration cutover: All code migrated fully with no disruption to downstream users.
Freed team from infrastructure management dependencies: No more patching, connectivity issues, or working around routine maintenance and coordination with infrastructure teams.
Pain points
Legacy on-prem tool throttled parallelism
Bottlenecks stretched nightly processing to 4-4.5 hours.
Infrastructure management consumed engineering time
Server maintenance and connectivity issues required coordination with infrastructure teams.
On-prem setup couldn’t scale with cloud strategy
The company’s cloud-first direction was blocked by its ELT tool, which required on-prem infrastructure.
Limited features held back modernization efforts
The previous system couldn’t keep pace with Snowflake’s evolving capabilities and modern data practices.
“Coalesce seemed like the right fit. It has the right amount of automation and the right features that we need.”
The challenge
By the time Texas Mutual had established Snowflake as its new data platform, using an on-prem ELT tool had become a limitation. Thread limits, infrastructure chores, and brittle upgrades left the team feeling constrained and unable to shrink its 4+hour nightly batch window or plan confidently for growth.
Texas Mutual wanted a simpler, Snowflake-native approach that improved performance without forcing the team into an over-engineered, code-only stack. The data engineering group began evaluating options that focused on ELT in Snowflake, supported a more visual way of working, and removed its dependency on on-prem infrastructure.
The evaluation
During the evaluation, Texas Mutual looked for a transformation platform focused on Snowflake that reduced custom tooling and extra schedulers while supporting both experienced engineers and colleagues who preferred a visual interface.
“When Snowflake evolves, Coalesce evolves. Coalesce is the right platform to allow us to leverage all of that.”
After comparing heavier enterprise tools and lighter cloud services, the team selected Coalesce as the option that best matched its Snowflake strategy, operational needs, and desire for the right balance between automation and control.
The migration
Texas Mutual approached the migration as a focused, like-for-like move of existing ELT logic into Coalesce, prioritizing all critical nightly workloads that previously took 4.5 hours to complete. The team worked closely with Coalesce to refine project structure and scheduling so the new environment ran in parallel with the legacy tool until the team was confident in every pipeline.
The team finally cut over all code to Coalesce with zero downtime for downstream users, then retired the old system and its servers. With parallelism increased from 7 to 20 threads and transformations tuned for Snowflake, nightly processing shrunk from about 4–4.5 hours to 3, resulting in an almost 20% improvement.
“Coalesce’s support has been excellent and fast in terms of delivering the features that we need.”
The impact
With Coalesce running on top of Snowflake, the data engineering team became free to focus on delivering higher-value data products and optimizations much faster.
Faster development, simplified operational support and infrastructure maintenance
Template-based mass code and pipeline generation helps the team develop pipelines much faster now. The intuitive visual interface also enables better coordination, faster impact analysis, and code management for the team. The team has also started leveraging AI based features in tool for auto documentation, impact analysis and pipeline generation.
Before the migration, a considerable amount of the data engineering team’s time went into keeping the on-prem environment healthy so the nightly runs could complete on schedule. The team consistently met service level agreements (SLAs), but did so through late-night checks and hot-fixes, absorbing impacts of server and OS upgrades, resolving tool upgrade issues, and constant vigilance over a complex stack sitting between source systems and Snowflake. Moving ELT processing into Coalesce simplified that picture. The team is no longer worried about operating system upgrades and patching, or connectivity issues with on-prem tool metadata database, built-in job scheduler issues, or tool-upgrade issues. Fewer moving parts meant fewer points of failure and less time spent chasing environment issues across teams.
Combined with decommissioned servers and retired licenses, Texas Mutual freed both budget and people to focus on data quality, new products, and governance instead of basic platform care and feeding.
Centralizing on a transformation platform that scales with Snowflake’s evolving capabilities
The previous transformation layer imposed hard limits on how far the team could push Snowflake, since thread caps and scaling constraints prevented it from safely increasing parallelism. With Coalesce sitting directly on top of Snowflake, processing now happens where the compute is designed to run, and jobs comfortably execute with a degree of parallelism of 20. The team organized more than 20 Coalesce projects to cover all its tables without running into previous bottlenecks or API issues. Because Coalesce tracks closely with Snowflake’s roadmap, Texas Mutual can adopt new platform features with confidence instead of worrying that the transformation layer will hold them back again.
Upskilling the data engineering team while delivering faster, more reliable data
“Collaboration and impact analysis are simpler. Much simpler. And in terms of our technical upskilling, Coalesce is a good fit. It helps us keep up with industry trends.”
For a six-person data engineering team inside a large data organization, tool choices directly affect how much learning and innovation are prioritized. The team now earns trust from stakeholders while also buying back time to deepen the skillsets of its members. With Coalesce, engineers can now focus on refining pipeline design, improving warehouse usage, and exploring Git-based workflows and Continuous Integration/Continuous Delivery or Deployment (CI/CD), rather than writing glue code or maintaining homegrown orchestration frameworks.
This extra bandwidth is especially important given Texas Mutual’s exploration of AI use cases, such as using Snowflake Cortex for search and agent-style experiences in addition to other Snowflake-native capabilities. In the near term, the team plans to capitalize on further enhancements to improve pipeline execution times and make the most of their Snowflake investment by optimizing warehouse costs and using Coalesce to experiment safely while keeping production data products reliable.