May 2, 2024

What is a FinOps Framework and How Does it Help with Data Platform Optimization?

By Allyse Fristo

The cloud data platforms of today offer unprecedented adaptability and scalability, creating vast opportunities to leverage data for generating business value. However, their versatility can also pose significant challenges, often related to cost volatility, poor performance, and business need misalignment. 

The resulting combination of the cloud platform and associated technologies is a complex system that requires a proper architectural foundation and continuous management to succeed. 

The beauty of a properly implemented cloud platform is that it can facilitate significant business value while only incurring costs for what is actually used on the platform; the solution for striking this balance is FinOps. 

Financial Operations (FinOps) is an operating framework that promotes a culture of optimizing platform performance to ensure maximum ROI. It is the answer to getting everything you need from your data platform and nothing you don’t. 

This blog outlines the fundamental principles of FinOps, what it means for the modern data platform, and provides examples of common organizational circumstances that have been improved with this approach.

What is FinOps?

FinOps goes beyond simple tactics to improve performance and cost; it is a comprehensive approach to ensure platform outcomes are aligned with business needs and supports continuous management of the platform for optimal performance. 

This includes a core set of principles: 

  • Establish visibility into cost and performance

  • Institute controls and guardrails

  • Optimize for the desired business outcomes

These same concepts are the foundation of FinOps for data cloud platforms and are married with phData’s implementation logistics specifically developed for the cloud data industry. 

What are the 3 Phases of FinOps?

Track

The first phase of FinOps begins with establishing visibility and quantifying platform performance. These insights are critical for ensuring all operations running on the platform add significant value, resulting in a higher ROI. 

It’s essential to identify and manage the most significant resource demands that may drain resources and increase costs unnecessarily

In this phase, visibility is established with real-time cost and usage dashboards, comprehensive platform performance auditing, and automated monitoring and alerting. This lays the foundation for the subsequent phases.

Control

This phase aims to mitigate avoidable risks by implementing a set of standard parameters alongside custom controls (derived from Phase 1 insights) to establish platform guardrails, parameter best practices, integrated applications system contingency analysis, and custom regulators.

For example, Snowflake, one of the forefront cloud data platforms, is designed with parameters such as auto-suspend, resource monitors, statement timeout, and many others that facilitate cost and performance controls across the platform. This results in a robust risk mitigation strategy against wasteful cost overruns and gaps in platform performance and security, allowing platform teams to focus on valuable development work.

Optimize

An ever-evolving platform mandates continuous management to maintain optimal performance. In the final phase, a multivariate approach is leveraged to optimize workload efficiency in Snowflake, correctly sizing warehouses for cost-performance balance, employing architecture best practices for speed, and regularly analyzing queries to improve efficiency. 

Understanding Snowflake’s compute-based pricing is crucial for managing costs as your organization scales. This approach ensures a scalable, efficient, and cost-effective Snowflake environment.

How phData Approaches FinOps

As data services consultants, we have culminated our experience across 250+ enterprise clients to distill comprehensive data platform best practices. While there is some theoretical content for applying FinOps to the cloud data ecosystem, there is little substantiated in practice. The phData Toolkit is one example of our synthesized expertise; a variety of SaaS applications developed in-house to meet the needs of the ever-evolving data ecosystem. Ranging from comprehensive platform auditing to real-time monitoring and alerting, the Toolkit in combination with our team of experts leads the industry in FinOps for the modern data platform.

Our optimization strategies are organized into three functional categories:

  1. Implementation of Best Practices

    1. After first establishing visibility into platform functionality, this consists of warehouse configurations, account parameter best practices, and other modifications that serve as important guardrails for performance and cost, mitigating unnecessary risk.

  2. Update and/or Establish Platform Architecture

    1. Establishing a robust architectural foundation from scratch and/or updating an existing platform architecture is paramount to optimal performance and healthy platform growth.

  3. Continuous Management

    1. The platform ecosystem is an ever-evolving system that requires continuous oversight and management. This is accomplished through automated monitoring and alerting, monthly/quarterly best practices auditing, business outcome and ROI metric tracking, and ongoing optimizations.

Implementation of Best Practices

Our teams establish a host of best practices and controls immediately. We often begin by running the phData Advisor Tool to provide a comprehensive report on platform performance and take action on the best practices the tool uncovers. This process is then repeated monthly or quarterly to ensure the best practices are maintained as the platform scales.

Establish Visibility

Utilizing the Elastic Platform Monitoring Tool, custom automated monitoring and alerting systems are set up to track data governance, security, performance, and cost parameters. These systems work alongside cost and usage reporting dashboards to offer real-time and historical platform data. This comprehensive approach ensures clear visibility into areas such as excessive under/over utilization of warehouses, unusual user activity, issues with materialized views, and any security or governance deviations.

Establish and Manage a Proper Information Architecture

Implementing stringent management of roles and access can dramatically reduce overhead and costs for support teams. This approach ensures proper warehouse design and sizing, minimizes compute costs, and (with automation) significantly curbs management overhead. Tagging enables showback and chargeback and allows for custom performance analysis. This foundation sets the precedent for all other platform operational capabilities and dictates the ability to scale platform value.

Continuous Management

The critical differentiator of FinOps is the continuous nature of optimizations. The cloud data platform is agile and adaptable, requiring ongoing management of performance, cost, and governance best practices. 

To strike the optimal balance of agility and risk aversion, the platform users need to be able to iterate quickly with reliable guardrails in place. Examples include re-evaluating warehouse sizing and workload placement, leveraging resource monitors to track and contain credit usage, and auditing for cost and performance best practices.

Operational Maturity Framework

FinOps is one component of our Operational Maturity Framework (OMF), a comprehensive model for evaluating cloud platform maturity. We conduct OMF evaluations on a quarterly basis to identify gaps, track progress, and define a roadmap for continued development. FinOps is related to many other platform maturity components and is fundamentally important for the high-level function of the platform.

Examples of FinOps Challenges phData has Solved

FinOps Implemented from Day 1 - Healthy Growth

Elastic Ops Platform Enablement

Client: A leader in the semiconductor Industry who started day one with platform best practices

  • Established a robust architecture foundation for secure and cost-effective onboarding of new use cases with operational scalability

  • Managed third-party applications efficiently, enabling the development team to focus on new use cases

  • Aligned strategic investments with business objectives, tracking outcomes, and measuring ROI on consumption growth

FinOps Implemented After Platform Was Established - Curtailing Cost Overruns

Elastic Ops Platform Efficiency

Client: A leader in semiconductor, automotive & aerospace, and electronics industries with out-of-control platform spend.

  • Lift and Shift migration resulted in widespread inefficiencies

  • Analyzed consumption and performance patterns to determine the most impactful changes

  • Implemented best practices for integrated application (dbt) processing and optimized warehouse utilization

FinOps Continuous Management

Elastic Ops Managed Services

Client: A leading national insurance provider struggling with cost and performance volatility.

  • Continuous, ongoing monitoring provided insights into several opportunities for optimizations over time

  • Optimized replication and storage processes, warehouse configuration, and account parameter best practices

How phData Can Help

Whether you are new to the cloud platform, are interested in optimizing your existing platform, or looking for ongoing operational support, phData’s Elastic Operations offerings can help make FinOps a success at your organization. Reach out to our Sales team to learn more about implementing FinOps.

FAQs

The most common inefficiencies that we experience typically boil down to design standards. A well-designed architecture means raw data gets enriched and standardized through multiple stages. As standards are defined this data should be persisted to a physical table instead of joining views together or creating join columns via subselected queries. This can cause unnecessary table scans and longer run times.

The most common problem our customers have is ingesting data into Snowflake efficiently. This can manifest itself in multiple forms, file sizes, file counts, file speed, file types; and they all lead to increased credit consumption. This problem can typically be solved via tweaks to the ingestion process but sometimes does require tooling that compliments the Snowflake data cloud and helps it meet broader enterprise needs.

You can learn more about FinOps from the FinOps Foundation: The FinOps Foundation is an independent association for and by FinOps practitioners in Cloud Financial Management and has excellent resources to help you understand the framework and practices of FinOps. You can learn more about FinOps as it applies to data cloud platforms by reaching out to the phData team for additional information.

Data Coach is our premium analytics training program with one-on-one coaching from renowned experts.

Accelerate and automate your data projects with the phData Toolkit