Introduction

Let's consider a scenario where your leadership challenges your technology department to demonstrate the value that it delivers for the business and mobilize teams to focus on a new set of priorities. This scenario might happen because of a change in the market conditions of your business, new executives changing the strategic direction, company mergers and acquisitions, or a large-scale project.

The task of finding new "measures of good IT" might be difficult. This is because the new measures require the updated definition of organizational success in business terms to be translated to corresponding technology metrics that your team can track and improve on. As a leader and agent of change, you have an important role to play. Appropriate metrics can help track demonstrable progress and sustain momentum. You can then maintain the sense of urgency that's a crucial ingredient for organizational transformations to succeed. For more information, see Leading Change: Why organizational transformations fail on the Harvard Business Review website.

Inspired by Amazon's "working backwards" tenet, this article proposes a mental model for technology leaders at any level. You can then use this mental model to identify and critically evaluate technology metrics that can translate to business metrics and goals. The article details scenarios, along with examples of metrics and justification for each choice. Technical Account Managers (TAMs) within the AWS Enterprise Support team help customers apply business metrics to effectively drive cost optimization and sustainability.

Understand the problem

It’s important for technology teams to understand what a top-level mandate for "higher focus on financial discipline" is. For example, customers ask TAMs questions such as the following: • “How can I check that the business value of the cloud is increasing while my AWS consumption is growing with my business?” • “How can I see how optimizations affect the amount of business transactions that I can perform?”

Your first course of action might be to launch cost optimization initiatives across IT resources, licenses, and infrastructure. However, next-level cost optimization might require expensive engineering efforts. This creates a complex trade-off, where the engineering hours spent on fine-tuning infrastructure efficiency might instead be invested to develop new revenue-generating features that improve system scalability or reduce technical debt. In some cases, excessive focus on cost optimization might reduce the technology organization's ability to drive business growth or respond to market opportunities. This can negatively affect financial results.

Businesses are constantly looking to increase productivity. The common practice is to measure engineering productivity through metrics such as lines of code or number of releases. However, these metrics pose challenges. While they’re quantifiable, they don’t reflect organizational effectiveness. High line counts might indicate verbose, poorly designed code rather than valuable features. Also, frequent releases might indicate fragmented development practices or insufficient testing rather than actual value delivery. Realistic productivity metrics must focus on business outcomes, such as how quickly you deliver customer-facing improvements and how effectively you allocate engineering resources to strategic initiatives. Another business outcome to measure metrics for is how well development processes balance speed with quality and maintainability.

Finally, organizations measure customer experience metrics based on operational work rather than actual value creation. Traditional metrics, such as ticket resolution times, number of incidents handled, or system uptime percentages, can miss the bigger picture and create a false sense of effectiveness. It’s important to validate whether the traditional metrics reflect genuine improvements. Ask yourself these questions: • Are we repeatedly fixing the same issues instead of preventing them? • Does our impressive uptime indicate that customers can use our services effectively, or do we have systems that are technically up but practically unusable?

The key is to focus on outcomes that directly affect customers rather than internal operational metrics. The internal metrics might be easy to measure, but might not correlate with actual user satisfaction.

Review common business metrics

Metrics aren't a foolproof means of communication. It's crucial to understand the strategic goal represented by metrics instead of focusing on the metric itself. The following are some common business metrics, along with the underlying motivation for their adoption:

• Financial metrics: These include measures such as revenue growth, profit margins, cost of goods sold (COGS), operating expenses, and return on investment (ROI). Financial metrics are vital to assess the overall financial health and sustainability of a business.

• Operational metrics: These metrics focus on the efficiency and effectiveness of business processes. Examples include inventory turnover, production downtime, order fulfillment time, and customer service response time. These metrics are essential to optimize resource allocation and improve operational performance.

• Customer metrics: These metrics measure the quality of customer experiences and loyalty. They can include customer satisfaction scores, net promoter score (NPS), customer retention rates, and churn rates. These metrics are required to align business strategies with customer needs.

• Employee metrics: These metrics measure workforce productivity and satisfaction. They can include employee engagement scores, turnover rates, training hours per employee, and average tenure. These metrics are critical to maintain a motivated and high-performing team.

• Strategic metrics: These metrics align with a company's long-term goals and strategic objectives, such as market share growth, brand equity, and innovation index. These metrics help in assessing the alignment of business activities with strategic vision and mission.

• Project metrics: These metrics are specific to project management, such as project completion rate, cost variance, schedule variance, and resource use. These metrics are vital to track the progress and success of projects and initiatives.

Evaluate the selection of metrics

When you choose technology metrics that align with your business outcomes, you must understand the common pitfalls and best practices. This helps you make sure that the metrics remain meaningful both initially and over time. The following are some key considerations:

• Avoid self-referential metrics. These metrics track team activities rather than value delivered. To test if a metric is self-referential, present the metric to someone outside your team with different business responsibilities. If their response is "so what?", then you might have a self-referential metric. For instance, in a cloud migration project, it's less meaningful to count migrated servers than track the performance and cost metrics for migrated applications. Even in a complete data center shutdown, server counts can be misleading because of varied complexity. For example, if you're migrating 200 servers, then the final 10 servers might require as much migration effort as the first 190. Another example is training initiatives. The number of sessions that you conduct or staff that you train reveals activity levels, but doesn't demonstrate business impact or actual skill improvement.

• Remove vanity metrics. These are metrics that boost confidence but provide no actionable insights. They can obscure reality and be particularly detrimental. For more information, see The Lean Startup on the Goodreads website. For example, total service signups might show constant growth while masking declining monthly acquisition rates or poor retention. Vanity metrics typically appear as absolute numbers rather than relative measures such as ratios or rates. These metrics focus on initial success metrics, such as signups for a service or first response times for case management. However, they don't show the complete picture of sustained value delivery. To identify vanity metrics, ask whether the measurement provides actionable insights and captures the full scope of value creation rather than only the initial steps.

• Recognize the limitations of proxy metrics. These metrics serve as indirect measures of complex scenarios. Even though these metrics are useful, don't confuse them with the actual goals that they represent. For example, when teams focus on improving the proxy metric rather than addressing the underlying business objective, they risk optimizing for the wrong outcome. If you adhere to the process and don't focus on the intended outcomes of the process, then you're at risk of business decline or Day 2 behavior. For more information, see Amazon's original 1997 letter to shareholders. Proxy metrics can be valuable tools to measure technology's business impact. Be aware of what they truly represent and their inherent limitations. Maintain focus on the actual business goal rather than allowing the proxy measurement to become the target itself.

• Use regular customer feedback to validate metric assumptions, especially for proxy metrics. Although technical metrics such as service availability might show that systems are functioning, customer input shows gaps between measurements and actual business impact. For example, a service might report high uptime when customers experience disruptive login issues. These issues might lead to service level agreement (SLA) disputes, even though the technical availability metrics are green. Initial reference points or thresholds that might not be relevant to business outcomes could influence decision-makers, a bias known as the anchoring effect. For more information, see Judgment under Uncertainty: Heuristics and Biases on the JSTOR website. This psychological tendency makes it important to maintain regular conversations with stakeholders, such as sales and customer support leadership. These conversations can help validate whether proxy metrics still accurately reflect intended outcomes or require adjustment. These conversations also help teams identify disconnects between measurements and actual business value. This, in turn, can result in metric refinement or implementation of complementary measures.

• Use the organization's mission and vision to resolve metric ambiguity. When business goals conflict, metrics can become challenging to interpret. For instance, an organization might simultaneously pursue aggressive cost reduction while investing in innovation to maintain market competitiveness. In such cases, an IT infrastructure cost metric might serve both objectives. However, the definition of success might vary depending on how you interpret the metric through each lens. When you face ambiguities, refer back to the organization's mission and vision for context. These foundational statements clarify the organization's ultimate objectives and plans to achieve them. The statements also help teams determine the primary purpose of their metrics and how to interpret results appropriately.

Map business metrics to cloud workloads

This article explores the following methods to map your business metrics to cloud workloads:

• Offset the cloud resource metric by a business metric. For example, the cost per sale metric. For this metric, we consider a cloud-obtained metric, such as cost in the case, over a business metric, such as a sale.

• Map business metrics to cloud resource metrics directionally. You aim for improvement in mapping when it's hard to directly map resources to business metrics. In these cases, you can make assumptions on cloud metrics that can directionally improve business metrics and monitor them. For example, if you have a business goal to increase website sales conversations, then map latency improvement with increased website conversations. Latency reduction might directionally improve the number of website conversations.

To map business metrics with cloud workloads, your organizations can take the following actions:

• Align cloud operations with business goals: Make sure that your cloud operations support strategic objectives, such as reducing operational costs, increasing agility, and improving customer experience.

• Enhance decision-making: Provide real-time visibility into workload performance, utilization, and costs to derive data-driven decisions.

• Optimize resource utilization: Identify underperforming workloads or resources. Then, optimize them to maximize ROI.

• Improve accountability and governance: Track metrics against predefined thresholds or goals to establish accountability and confirm compliance with governance policies.

These metrics can provide valuable insights into how cloud resources can support your business goals, operational efficiency, and customer satisfaction.

Adopt a multi-level approach

You can use a multi-level approach to find a business metric and then find offset or directional proxies that link to that metric.

Metrics that are focused on key business goals align cloud performance with the overall business strategy. This alignment provides a high-level view that executives and stakeholders can easily understand. These metrics focus on objectives, such as revenue growth, cost reduction, or customer satisfaction, to facilitate strategic decision-making. However, they might lack granularity and fail to capture specific performance issues. They also might oversimplify the complexity of cloud workloads, which can potentially lead to misguided decisions if supplemental metrics aren’t included.

In contrast, metrics for specific business transactions provide detailed insights into workload performance for targeted optimization. These metrics are particularly useful for operational teams that monitor specific processes or transactions. They can help identify patterns or trends that affect business activities, such as trading volume or order processing time. However, these metrics might overwhelm stakeholders with excessive information. They might also narrowly focus on transactional performance, potentially overlooking broader strategic objectives. It’s a best practice to use a balanced approach that combines metrics for key business goals with detailed transaction metrics. That way, organizations can link business outcomes with resource performance and use the strengths of both perspectives.

Apply business metrics to cloud workloads

To apply your business metrics to cloud workloads, follow these steps:

1. Identify key business goals and map them to specific cloud workloads. For example, if your business goal is to increase customer satisfaction, then map metrics that are related to response time, service availability, and error rates to customer-facing workloads.

2. Define metrics that are specific to workloads: Develop specific metrics that are unique to each cloud workload. For example, a cloud-based ecommerce application might track metrics, such as page load times, transaction success rates, and average order value. These metrics help understand the performance of individual workloads and their contribution to business objectives. You can also decide whether you need to track a directional proxy or an offset metric.

3. Automate metric collection and reporting: Implement automation tools to collect, analyze, and report on metrics in real time. Automation reduces the time and effort required for data collection, minimizes errors, and makes sure that decision-makers have access to the latest information.

4. Create multi-level dashboards: Use dashboards that offer multiple levels of granularity, from high level business metrics to detailed workload-specific metrics. This approach helps executives to quickly see overall performance. It also allows operational teams to analyze specific workload metrics on a deeper level.

5. Leverage machine learning and AI: Incorporate machine learning and AI to identify patterns, trends, and anomalies in workload performance data. Predictive analytics can help forecast future trends, optimize resource allocation, and improve decision-making.

6. Integrate with business intelligence tools: Use business intelligence tools that integrate cloud workload metrics with other business data sources. This integration allows for a holistic view of business performance and supports cross-functional analysis.

The following table provides examples of cloud resource metrics and how to source them.

The following is an example metric graph. The X-axis shows the time period, and the Y-axis shows the number of vCPU hours grouped by linked accounts. We divided all Y-axis numbers by the number of business transactions that occurred in the same time period for that workload. This graph helps us see how resource usage changes because of optimizations that we make outside scaling of business workloads.

Conclusion

The application of business metrics to cloud workloads through dashboards is essential for modern businesses that aim to optimize cloud investments and drive strategic success. Therefore, align metrics with your broader business goals, and use tools such as CloudWatch or CURs for real-time monitoring. Also, regularly reassess metrics for relevance and use dashboards to effectively communicate insights to stakeholders. Use Amazon QuickSight to graph cost by unit metric. You can use the Cloud Intelligence Dashboards Framework to visualize and understand your cost and usage data.

By understanding the fundamentals of business metrics, mapping them to cloud workloads, and carefully identifying the right mix of metrics, organizations can effectively monitor, analyze, and improve their performance in a cloud-centric world. A balanced approach that includes both high-level business goals and specific transaction metrics provides the comprehensive visibility that's needed to achieve optimal outcomes.

AWS Support engineers and Technical Account Managers (TAMs) can help you with general guidance, best practices, troubleshooting, and operational support on AWS. To learn more about our plans and offerings, see AWS Support.

---

About the authors

Thomas Coombs

Thomas Coombs is a Principal TAM at AWS in Switzerland. In his role, Tom helps enterprise AWS customers operate effectively in the cloud. From a development background, he specializes in machine learning and sustainability.

Giacomo LOZITO

Giacomo LOZITO, MBA, is an Enterprise Support Manager for strategic customers in the Energy industry at AWS. Giacomo combines a passion for strategic thinking and leadership, a business focus on decisions, and a modern IT background to help AWS customers promote a culture of high performance and innovation across their organizations.