New KPIs Are Smarter Than Ever
AI-enabled algorithms are boosting performance and value at leading-edge organizations, according to a new report
Q&A With Michael Schrage
“Legacy key performance indicators (KPIs) increasingly fail to deliver the information and insights leaders need to succeed” in their daily business efforts, according to a new report published by the MIT Sloan Management Review (SMR). KPIs fall short in tracking progress, aligning people and processes, prioritizing resources, and advancing accountability.
That’s not sustainable. Thanks to AI, there’s growing recognition that KPIs need to be smarter and more capable. Improved algorithms — empowered by AI — transform traditional performance metrics into intelligent, adaptive, and predictive tools. “Smart KPIs become sources — not merely measures — of strategic differentiation and value creation,” said Michael Schrage, a research fellow at the MIT Initiative on the Digital Economy (IDE) and a co-author of the report. Among the significant findings of the study, The Future of Strategic Measurement: Enhancing KPIs With AI, is that
“companies that revise KPIs with AI are three times more likely to see greater financial benefit than those that do not.”
Smarter KPIs lead to better outcomes.
Schrage and co-researchers François Candelon, Shervin Khodabandeh, and Michael Chu from the Boston Consulting Group, and David Kiron of SMR, surveyed more than 3,000 global managers and conducted interviews with 17 executives. They found that 60% of managers believe they need to improve their KPIs. However, only one-third (34%) are using AI to create new KPIs.
IDE Editorial Content Director, Paula Klein, asked Schrage to explain the state of smart KPIs and the study’s implications.
Q: You’ve done a lot of research on strategic measurement and the role of KPIs in recent years. What is the most important takeaway from this latest report?
A: I’d say there are actually three important takeaways that we identified and discussed. The first is AI — specifically, generative AI and machine learning technologies — makes KPIs even more important, more valuable, and more strategic to leaders who want to better understand and improve performance analytics and performance excellence.
This is not hype; if anything, we think the true potential and impact are underappreciated.
The second is that KPIs themselves are already becoming smarter and more intelligent. KPIs can learn. They learn to improve. They are becoming intelligent software agents that can learn from, learn with, and learn for their humans. We’re seeing GenAI-enabled KPIs making recommendations to managers and executives on the data and analytics these KPIs need to do their job even better. In other words, these KPIS are learning to learn.
Finally, KPIs are becoming so important and interdependent that executive committees and boards of directors will need smart KPIs to be sure they’re getting the best possible value, information, insight, advice, and performance from their previous KPIs. Organizations will use meta-KPIs to evaluate current KPIs.
Q: The case for better performance measurement seems strong, but is implementation a heavy lift? How much progress have organizations made on adopting smarter KPIs and how do they measure success?
A: Unsurprisingly, I’d describe the progress as mixed; some firms are more capable and committed than others. However, I’d challenge the premise that adopting smarter KPIs is all that heavy or intimidating. To the contrary, our research suggests that optimizing KPIs is less burdensome than many other data-driven digital transformation opportunities.
The biggest challenge we observed is that too many firms take a ‘set’em and forget’em’ approach to their key performance metrics.
That is, they engineer or reengineer their digital transformation and algorithmic investments around legacy measures like sales, cash flow, net promoter score (NPS), customer satisfaction, return on assets (ROA), and overall equipment effectiveness (OEE) instead of using their new wealth of data, cloud, compute, analytics, AI and machine learning to revisit and rethink the real value and purpose of what they’re measuring. It’s as if trucking and logistics companies believed agile navigation and on-time performance came from better Rand McNally maps, sextants, and magnetic compasses and not GPS, Google Maps and Waze! Sorry; that’s not going to happen. Our research suggests those days are fading fast.
In the wake of pandemic-driven digital transformation initiatives and greater recognition of data-driven value creation, enlightening your KPIs and digital dashboards should now be faster, easier and cheaper. Unfortunately, our research also shows
most organizations default to old ways: they use digital transformation to reduce the cost and improve efficiencies of existing processes, programs and projects instead of strategically upgrading their desired ends. And that means, they’re mismeasuring what matters.
Firms seriously investing in AI and GenAI are seriously investing in what I call KPAIs — metrics enabled and empowered by Key Performance Artificial Intelligence. That’s the success we saw at Walmart, Sanofi, Schneider Electric, and Pernod-Ricard.
Q: How do leaders infuse the concept throughout the organization? With so many transformational tools and competing imperatives how can smart KPIs bubble up to the top?
A: Automating workflows that save IT money but cost you employees and clients should be the harder sell; that’s a false economy. What we’re demonstrating is that if you’re not mapping out how your data flows and workflows measurably contribute to your key performance metrics, you’re solving the wrong problem! If you’re running a capital-intensive factory and you’re building digital twins to make sure you’re maximizing return on those hard assets, you’ve got the people, brains and technology to similarly simulate and “twin” the KPIs those capital-intensive assets are supposedly contributing to. Essentially, our research suggests KPAIs are the mechanisms for quantitatively exploring and exploiting what you want ‘excellence’ to measurably mean.
These issues are the essence of strategic, operational and competitive success. Our research reveals too many leaders have foolishly invested in intelligent digital technologies in service of not very intelligent metrics. I once remarked that, “Your strategy is your KPIs; your KPIs are your strategy.”’ With better data, compute and AI, I can now say that, “smarter strategies depend on smarter KPIs.”
Q: Let’s break it down incrementally and specifically: Where do they begin? What is required in terms of costs and resources?
A: These processes are what I teach in my MIT Executive Education classes and advisory work. You begin by asking and answering four simple questions:
• What KPI generates your most useful insight(s)?
• What KPI creates the greatest internal and/or cross-functional conflict?
• What KPI should your organization STOP USING RIGHT NOW?
• What new/novel strategic KPI should you embrace over the next 12–18 months?
I’ve learned that you need to begin where people are and make a strong case. We live in disruptive times: is there a new or novel metric — or KPI combination — that would help future-proof a key product, process or function? Start thinking of KPIs less as a way of keeping track and more as a way of organizing value.
Q: What about hurdles to AI adoption such as disruption to current organizational culture and leadership roles?
A: This question strikes at the heart of what we’re really saying: serious leaders and managers are going to have the same kind of relationships with their intelligent KPIs and dashboards as teenagers do with their phones and TikTo, Snapchats and Instagram. KPIs are their essential user interfaces. In the report, Sanofi offers a great example with its PLAI network and capability
Intelligently integrating intelligent KPIs into the enterprise does require a new cultural commitment to more effective partnerships between talented people and their technologies. How leaders learn from and with their smartest KPIs will determine their success.
