It’s a well-worn maxim that companies can’t manage what they don’t measure. But using A.I. transforms both what companies can measure and what they should manage. Several corporations have learned to combine A.I. with Big Data to identify new performance indicators and refine existing ones, which is helping them develop fresh ideas about what drives their performance.
Surprisingly, A.I. often upends traditional assumptions about the drivers of performance, profitability, and growth. Using the technology therefore helps companies transform their performance and sustain that transformation, not just keep track of legacy metrics, making A.I. a critical factor in performance measurement. According to the BCG-MIT 2023 Global Executive A.I. Survey (the Survey), seven of 10 respondents agree that better key performance indicators (KPIs) are critical for success.
Our studies show that the A.I. leaders are using data and technology, such as supervised and unsupervised machine learning or deep learning to measure and manage KPIs in three ways: They create new KPIs with A.I.; they prioritize the KPIs that matter using A.I.; and they improve alignment across the organization with A.I.-designed shared KPIs.
Creating new KPIs with A.I.
Despite deploying an army of data scientists, Google was unable to boost the performance of a primary digital channel for years. Although Google collected vast amounts of data, they couldn’t quite identify the key parameters to measure, manage, and monitor in order to improve the performance of clients’ campaigns. Frustrated that its data, analytics, and talent weren’t moving the needle, Google eventually turned to A.I. It developed an algorithm, fed it all the data it could, and asked the A.I. to figure out how the channel could deliver better results for clients, which would boost its financial performance.
Using unsupervised machine learning techniques, the A.I. identified connections, correlations, and causations that Google’s engineers had missed; ranked the importance of new performance indicators; and identified the most critical ones. The engineers learned from the A.I. that some indicators that were thought to be critical were, in reality, unimportant. The A.I. also suggested that other metrics on which the platform hadn’t focused until then were among the top performance drivers, so Google started focusing on them. Six months after it implemented the A.I.’s recommendations, the campaigns saw a 30-point improvement in performance.
A.I.’s ability to identify novel performance indicators that matter is impressive. According to the Survey, 34% of companies already use A.I. to create new performance indicators—and as many as 90% of those say that their KPIs have improved as a result. Moreover, companies that use A.I. to create new KPIs are twice as likely to improve the efficiency with which they operate. A.I.-designed KPIs also make performance more predictable; these companies say they’re three times as effective at predicting future performance than those that don’t use A.I.
That isn’t true only for tech-savvy giants; A.I. can identify hidden problems in analog medium and small companies as well and find ways to track and tackle them. One of South America’s retailers, for instance, created a dataset made up of time series data on, among other things, demand, supply, sales, warehouses, trucks, et al. Then it used an unsupervised A.I. model to develop new KPIs that helped optimize its logistics across nine countries. In the first 90 days, the A.I.’s recommendations led to a major 14% decrease in the retailer’s logistics costs.
Prioritizing KPIs with A.I.
Like several other leaders, the Singapore-headquartered DBS Bank started using the technology to track conventional metrics such as the KPIs for function, which each was responsible for optimizing. However, the bank struggled to become better, and the early experiments proved to be unsuccessful (albeit ones from which DBS learned).
Three years ago, DBS decided to adopt the concept of consumer “journeys”—such as the journey to provide a consumer with a credit card or a housing loan—and created an A.I.-driven control tower to track all the journeys. Doing so allowed the bank to identify the most crucial factors that drove the outcomes it desired in terms of customer experience, profitability, employee experience, and risk level, and, crucially, to prioritize their importance. DBS made the data visible to its cross-functional teams, whose members now had a stake in optimizing results in all four categories. That gave the bank the confidence to make every decision with A.I.-driven insights. As a result, DBS’s profits before tax rose from around U.S. $5 billion in 2021 to over $6 billion in 2022, and Euromoney and Global Finance rated it one of the world’s best banks last year.
Industries have become more complex, and companies are larger, so prioritizing KPIs has become critical. An organization, its businesses, national operations, teams, and employees have to pursue multiple objectives, but the number of KPIs could become overwhelming for many. Using A.I. to develop a streamlined set of KPIs, sharply tuned to each employee’s performance goals, will help channel efforts towards organizational objectives that matter.
KPIs themselves need key performance indicators. Their efficacy demands periodic evaluation, just as companies regularly assess employee performance. Developing such key performance indicators for KPIs, which will help with KPI parsimony, helps companies better anticipate challenges, optimize resource allocation, and adapt faster to market dynamics. For instance, Schneider Electric has created a performance management office both to improve its performance and the metrics it uses.
Aligning with shared KPIs
A.I. is well suited to uncover the overlaps among KPIs and resolve the resulting trade-offs and inconsistencies. A.I.-generated shared KPIs can lead to improved organizational alignment.
In health care, for instance, reducing admissions is important for cost reduction, but also a key outcome indicator. In legacy provider organizations, CFOs manage costs and reimbursement flows, and chief medical officers (CMOs) emphasize the quality care of patients and their release from the hospital. A.I. can now enable the analysis of patient data, identify the root causes of readmissions, and suggest targeted interventions. Using this information, CFOs and CMOs can share a “patient readmission rate” KPI (the lower, the better) when identifying the root causes and predicting interventions to simultaneously improve outcomes and reduce costs. This shared KPI promotes alignment across the organization and would not be possible without A.I.-driven pattern recognition.
While in limited situations, some companies might prioritize a single North Star KPI (e.g., maximize sales per customer, or maximize repeat visits), most companies are nowadays part of business ecosystems, which makes the use of a single indicator impractical. Overlaps create conflict; each entity will bring different data sets, dataflows, and workflows to the table, and their priorities may often clash.
As the objectives of different ecosystem players, business units, and functions become increasingly intertwined in the pursuit of organizational goals, managing a shared set of KPIs will improve their alignment. Above all, it will enable executives to stop chasing siloed metrics and deepen their understanding of how the organization creates value. While data-sharing between teams is necessary to deliver results, developing a business case to prove the efficacy of the shared KPIs will rally teams behind them—and boost their performance.
…
Companies that use A.I. for performance management will be able to optimize their existing KPIs and design new ones as well. They must balance this by using the technology to prioritize the KPIs they count on. Doing so will allow CEOs to stop managing by looking back in the rearview mirror and to do so by looking ahead through the windshield. Thus, A.I. seems likely to open up a new era of performance measurement—and new frontiers of business performance.
Read other Fortune columns by François Candelon.
François Candelon is a managing director and senior partner at BCG, and the global director of the BCG Henderson Institute. You may contact him at Candelon.Francois@bcg.com.
Shervin Khodabandeh is a managing director and senior partner at BCG.
Michael Chu is a partner and associate director, data science, at BCG X.
Gaurav Jha is a consultant at BCG, and an ambassador at the BCG Henderson Institute.
Some of the companies featured in this column are past or current clients of BCG.
