Q&A: How to Make Key Performance Indicators Key Again

KPIs will only be relevant and compelling if businesses revamp old metrics to match digital competencies of today — and tomorrow.

Businesses have always needed tools to track and measure performance, and therefore value, of employees, sales, and production. Key Performance Indicators (KPIs) — whether scorecards, dashboards or standard metrics — demonstrate how well, or effectively, a company is achieving a key objective. Recently, however, KPIs are falling short. Are they too outdated and arcane for today’s digital businesses? How can they be improved?

A new, global survey and research report published June 26 by the MIT Sloan Management Review (SMR) addresses these and other concerns about KPIs. The cross-industry survey of more than 3,200 senior executives at traditional and digital firms was coupled with 18 executive and thought leadership interviews. The results show that change is requisite. “Accelerating technological innovation, intensifying competitive pressure, and increasing customer expectations are forcing business executives to rethink how they use KPIs to lead and manage the enterprise.”

In fact, the study “finds business leaders worldwide struggling to strike a workable balance between tactical and strategic KPIs; operational and financial KPIs; and KPIs that effectively capture the moment while anticipating the future.”

Clearly, yesterday’s metrics need to make way for predictive analytics and machine-learning capabilities that will offer better insights to move the business forward.

MIT IDE fellow Michael Schrage, who co-authored the report with SMR’s David Kiron, highlighted some of the most important findings in an interview with IDE Editorial Content Manager, Paula Klein. What follows is a summary of the conversation.

Michael Schrage

IDE: Why the imperative to examine KPIs now? What business problem should they be solving that they are not? What’s at stake?

Michael Schrage: Why now? We live in the era of big data getting bigger, machine learning (ML) algorithms getting smarter, and innovative arrays of technologies that allow us to monitor and measure user experiences in ways unimaginable even a decade ago. Does anybody believe that legacy KPIs from 2001 or 2015 are adequate for tomorrow’s enterprise accountability or innovation? That’s absurd. Serious investors look at Amazon, Google, Facebook, Alibaba, Microsoft, Netflix, or Tencent, and immediately understand that yesterday’s approaches aren’t up-to-speed with how value gets created and measured these days. That was the original insight for this research and survey.

We found that the most sophisticated businesses — those that appreciate and understand digital transformation and accountability — take a more dynamic view of KPIs. They use the indicators to set transformation expectations and lead, not just manage the business. They see and treat KPIs as drivers of strategic change and opportunity. That’s a big deal.

On the other hand, however, some uncomfortable truths surfaced, as well. For the majority of organizations, KPIs are regarded as “key” in name only. KPI is a lazy name for just another metric.

For example, we were surprised that only a quarter of the executives (26%) say their functional KPIs are well-aligned with the organization’s strategic objectives. That means roughly three-quarters of respondents effectively acknowledge a disconnect between their functional and strategic metrics. They’re not aligned. They’ve got more data and better analytics than they’ve ever had before and yet their organizations don’t clearly align functional operations with strategic aspirations. That’s why this survey offers a clear warning to KPI underachievers: Perfunctory and commodity KPIs are going to get you perfunctory and commodity outcomes. That’s a recipe for stagnation and decline.

M. Schrage and D. Kiron, “Leading With Next-Generation Key Performance Indicators,” MIT Sloan Management Review, June 20108.

IDE: You say: “machine learning is poised to radically influence how executives use KPIs to monitor and spur growth.” Explain the potential connection between ML and KPIs.

Schrage: We see the beginnings of a “big flip” taking place. Traditionally, KPIs told management how well the business was doing. They were largely retrospective, coming from an accounting and/or financial reporting culture. KPIs were analytic outputs that informed executive decisions.

In a ML age, everything flips: Instead of KPIs being outputs for humans, they’re being used as inputs for machines. That is, organizations are using KPIs to train machine learning algorithms. Just as KPIs informed humans in the past, we can train machines to optimize KPIs in the future. We advise businesses to see KPIs as data sets for machine learning.

IDE: Describe how today’s KPIs differ from the dashboards and scorecards that many traditional firms use. How can analytics become a value creator? What’s an example?

Schrage: David Kiron and I are tremendous admirers of Bob Kaplan’s work and the Balanced Scorecard. Peter Drucker’s MBOs and Intel’s OKRs — now evangelized by John Doerr, the terrific Kleiner, Perkins venture capitalist — are excellent, too. However, they represent wonderful pre-Internet, pre-Big Data, pre-Machine Learning ways to think about accountability and business aspirations. Now, algorithmic innovation and data volume, as well as variety and velocity, disrupt these legacy KPI approaches.

Do you want to navigate the future with a compass or a GPS? How much faster do you drive if you improve the optics of your rear-view mirror? I think our survey makes clear that legacy KPIs have become less valuable for tomorrow’s accountabilities and aspirations.

Consider Netflix. Who would have thought that binge viewing would rightly become a KPI for business success? Netflix not only monitors and collects data on viewer behavior; it can also create and recommend programs based on that data. Netflix uses data and analytics to create more binge viewers and binge viewing. My bet is that Netflix has a “conversion-to-binge-viewer” KPI that helps its marketing team greatly.

While this survey mostly examined marketing and customer experiences, future research in KPI design and development seems ripe. I look forward to integrating behavioral economics and social psychology research into KPI and digital dashboard design. The opportunities for experimentation — both in the real-world and academia — are exciting, too. Most of all, this strikes me as important study to pursue. I hope we can even come up with KPIs for KPIs one day!