Good for Machine Learning and/or Good for Business?
By Emilio Lehoucq, Data Scientist, Northwestern University
There is a paradox in the history of machine learning (ML). To put it bluntly, what has overall been good for the rise of ML has been bad for most individual companies. However, there is a way to align both interests. Building on my doctoral research, I share here some important findings and practical recommendations for business leaders.
Economics Behind the Rise of Machine Learning
We take for granted that ML rose because it has allowed companies such as Google and Meta to process and extract profit from data. This is certainly part of the story. But there is another side to the story.
Since the mid-1990s, many corporations have repeatedly invested in ML, often discovering that it fails to significantly boost profitability. Even by the late 2010s, seven out of ten companies reported minimal or no impact from ML. Only about one out of ten corporations obtained significant financial benefits.[i]
The rise of ML has not only been about companies extracting profit from data, but also about adopting what I call an “open-ended relationship to profit.” Intentionally—and largely unintentionally—many corporations have invested in ML because they see the technology as intrinsically valuable regardless of profit or because they hope to be well-positioned for a longer-term, uncertain future.
Companies have adopted an open-ended relationship to profit in multiple ways, such as:
- Irrational Belief: Often companies invest in ML not because it solves a specific problem in a way that maximizes profit but because of ML’s promise, exploring its potential after the decision to invest.
- Copycatting: Many corporations adopt ML because it signals conformance to industry practices rather than because it improves performance.[ii]
- Siloing: In many cases, a lack of collaboration between data scientists and business teams leads companies to invest in ML solutions that fail to maximize profit.
What You Can Do as a Business Leader
Understanding these dynamics can enhance your decision-making. The issue is not that companies adopt an open-ended approach to profit, but rather that they often do so without intention and strategy. As a business leader, you can deliberately guide this process, ensuring that ML development is supported while making judicious investments.
These are some practices that you can consider adopting:
- Incentivize employees to contribute to open-source ML projects.
- Invite academics to play with your data and servers.
- Organize ML hackathons.
- Allow employees to devote some of their time to ML projects of their choosing.
- Engage in research and development around ML.
- Sponsor ML workshops.
- Fund ML conferences and training.
These are examples of practices that can be helpful. Some may be relevant for your company, particularly depending on the size and industry. Luckily, there is a range to choose from and many of them are not only good for investing in ML, but also to attract and retain talent.
Conclusion
The rise of ML has often not translated into profit for most companies. However, as a business leader, you can change this dynamic. With the current excitement around generative AI, now is the time to make strategic decisions that align the benefits of ML with business profitability. Do not let this opportunity go to waste.
[i] Ransbotham, Sam, Shervin Khodabandeh, Ronny Fehling, Burt LaFountain, and David Kiron. 2019. Winning with AI: Pioneers Combine Strategy, Organizational Behavior, and Technology. MIT Sloan Management Review and Boston Consulting Group. QuantumBlack. 2022. The State of AI in 2022--and a Half Decade in Review. QuantumBlack AI by McKinsey. Ransbotham, Sam, Shervin Khodabandeh, David Kiron, François Candelon, Michael Chu, and Burt LaFountain. 2020. Expanding AI’s Impact with Organizational Learning. MIT Sloan Management Review and Boston Consulting Group.
[ii] DiMaggio, Paul, and Walter W. Powell. 1983. “The Iron Cage Revisited: Collective Rationality and Institutional Isomorphism in Organizational Fields.” American Sociological Review 48(2):147–60. Meyer, John W., and Brian Rowan. 1977. “Institutionalized Organizations: Formal Structure as Myth and Ceremony.” American Journal of Sociology 83(2):340–63.
About the author
Emilio Lehoucq is a data scientist at Northwestern University. Before his current role, he earned his Ph.D. in sociology and MS in statistics at Northwestern, focusing on the sociology of data, particularly the economic dynamics behind the rise of machine learning in the United States since the mid-1990s.
By Emilio Lehoucq, Data Scientist, Northwestern University
There is a paradox in the history of machine learning (ML). To put it bluntly, what has overall been good for the rise of ML has been bad for most individual companies. However, there is a way to align both interests. Building on my doctoral research, I share here some important findings and practical recommendations for business leaders.
Economics Behind the Rise of Machine Learning
We take for granted that ML rose because it has allowed companies such as Google and Meta to process and extract profit from data. This is certainly part of the story. But there is another side to the story.
Since the mid-1990s, many corporations have repeatedly invested in ML, often discovering that it fails to significantly boost profitability. Even by the late 2010s, seven out of ten companies reported minimal or no impact from ML. Only about one out of ten corporations obtained significant financial benefits.[i]
The rise of ML has not only been about companies extracting profit from data, but also about adopting what I call an “open-ended relationship to profit.” Intentionally—and largely unintentionally—many corporations have invested in ML because they see the technology as intrinsically valuable regardless of profit or because they hope to be well-positioned for a longer-term, uncertain future.
Companies have adopted an open-ended relationship to profit in multiple ways, such as:
- Irrational Belief: Often companies invest in ML not because it solves a specific problem in a way that maximizes profit but because of ML’s promise, exploring its potential after the decision to invest.
- Copycatting: Many corporations adopt ML because it signals conformance to industry practices rather than because it improves performance.[ii]
- Siloing: In many cases, a lack of collaboration between data scientists and business teams leads companies to invest in ML solutions that fail to maximize profit.
What You Can Do as a Business Leader
Understanding these dynamics can enhance your decision-making. The issue is not that companies adopt an open-ended approach to profit, but rather that they often do so without intention and strategy. As a business leader, you can deliberately guide this process, ensuring that ML development is supported while making judicious investments.
These are some practices that you can consider adopting:
- Incentivize employees to contribute to open-source ML projects.
- Invite academics to play with your data and servers.
- Organize ML hackathons.
- Allow employees to devote some of their time to ML projects of their choosing.
- Engage in research and development around ML.
- Sponsor ML workshops.
- Fund ML conferences and training.
These are examples of practices that can be helpful. Some may be relevant for your company, particularly depending on the size and industry. Luckily, there is a range to choose from and many of them are not only good for investing in ML, but also to attract and retain talent.
Conclusion
The rise of ML has often not translated into profit for most companies. However, as a business leader, you can change this dynamic. With the current excitement around generative AI, now is the time to make strategic decisions that align the benefits of ML with business profitability. Do not let this opportunity go to waste.
[i] Ransbotham, Sam, Shervin Khodabandeh, Ronny Fehling, Burt LaFountain, and David Kiron. 2019. Winning with AI: Pioneers Combine Strategy, Organizational Behavior, and Technology. MIT Sloan Management Review and Boston Consulting Group. QuantumBlack. 2022. The State of AI in 2022--and a Half Decade in Review. QuantumBlack AI by McKinsey. Ransbotham, Sam, Shervin Khodabandeh, David Kiron, François Candelon, Michael Chu, and Burt LaFountain. 2020. Expanding AI’s Impact with Organizational Learning. MIT Sloan Management Review and Boston Consulting Group.
[ii] DiMaggio, Paul, and Walter W. Powell. 1983. “The Iron Cage Revisited: Collective Rationality and Institutional Isomorphism in Organizational Fields.” American Sociological Review 48(2):147–60. Meyer, John W., and Brian Rowan. 1977. “Institutionalized Organizations: Formal Structure as Myth and Ceremony.” American Journal of Sociology 83(2):340–63.