In the complex landscape of machine learning and business strategy (see ML Strategy Development: Content Analysis and Business Oriented Metrics), there is a powerful exercise that I often recommend to my clients and teams, one that inherently addresses many vital aspects of strategy development models. This exercise is simple but profound in its effects: measure in money.
Express measurements in monetary terms — such as revenue generated, cost savings or risk reduction — provides a clear and tangible understanding of the impact of machine learning (ML) models. This approach translates the results of abstract models into concrete economic terms, making it easier for stakeholders across the organization to understand the value and implications of these models. By quantifying results in monetary units, the often intangible benefits of ML become visible and tangible, encouraging a deeper appreciation of the model’s role and capabilities in driving business success.
Monetary measurements also facilitate comparisons between different ML projects and other business initiatives. In a landscape where resources are finite and prioritization is key, the ability to instantly compare the financial impact of various projects is invaluable. This comparison assists in strategic decision-making, ensuring that resources are allocated to projects with the highest ROI potential.
However, Converting traditional measurements into monetary units is a challenge. This process often involves assumptions that may be arbitrary, leading to questions with elusive answers. The resulting monetary measurements may initially appear unsatisfactory or inaccurate. However, the real value of this exercise lies not in the perfection of the final measurement but in the process itself. Trying to quantify the economic impact of ML models forces teams to dig deeper into the context of their work, revealing what they know and, more importantly, what they don’t know. This newfound awareness of knowledge gaps is more valuable than the false comfort of unexamined metrics.
Participating in this exercise transforms the understanding of team projects. It provides a framework for understanding the impact of their work in the real world, encouraging a deeper exploration of the business context and a more nuanced appreciation of the model’s role within it. Additionally, it instills a sense of how much confidence to place in decisions based on metrics. Before this exercise, there were uncertainties and assumptions. they were simply unacknowledged.
For those who argue that some aspects of ML’s impact are “not measurable,” Douglas W. Hubbard’s book How To Measure Anything is proof to the contrary. This work is not just a rebuttal to the notion of immensity, but also a rich source of methods and insights for navigating beyond such perceived limitations. It encourages a shift from a mindset of limitation to one of exploration and discovery, enabling teams to quantify and understand even the most elusive aspects of their work.
While this “monetary metric mapping” serves as a powerful tool for aligning machine learning metrics with the business context, it is important to remember that this monetary metric is not a complete representation of ethical reality. Financial metrics, while critical to business alignment, may not fully capture the broader social, ethical and environmental impacts of ML applications. This approach primarily quantifies economic impact, but should not overshadow the responsibility to consider and address the ethical dimensions and implications of ML projects. Therefore, while embracing the utility of monetary measures for strategic alignment, it is equally important to maintain a balanced perspective, ensuring that moral factors are not overshadowed by economic objectives.
In summary, measurement in monetary terms is more than a metric exercise. represents a fundamental shift in perspective that aligns ML projects more closely with business goals. This approach challenges teams to think critically, challenge assumptions, and dig deeper into both what they know and what they need to discover. While this method effectively quantifies value, it also reveals it, facilitating more informed decisions and a detailed understanding of the true impact of machine learning on the business domain. Importantly, this focus on financial metrics should be balanced with ethical considerations, ensuring that the wider social and ethical impacts of ML projects are not overshadowed by their financial results.