Ideas for your work from MIT Sloan School of Management | Office of Communications
+ THREE INSIGHTS FOR THE WEEK August 15 – August 21, 2021
1. Data science teams can be a great source of value to the business, but they need proper guidance. MIT Sloan professor of the practice Rama Ramakrishnan, a former Salesforce executive, recommends leaders follow these steps to help data science teams realize their full potential:
- Point data science teams toward the right problem.
- Decide on a clear evaluation metric up front.
- Create a common-sense baseline. This will force the team to get the end-to-end data and evaluation pipeline working and uncover any issues, such as with data access, cleanliness, and timeliness.
- Manage data science projects more like research than like engineering. Understand there is a strong element of research in most data science work, which means a fair amount of time spent on dead ends with nothing to show for the effort.
- Check for “truth and consequences.” It’s important to subject results to intense scrutiny to make sure the benefits are real and there are no unintended negative consequences.
- Log everything, and retrain periodically. If every input and output is logged in as much detail as possible, investigating and fixing problems will be easier and faster.
2. Half a century ago, NASA built and maintained a physical twin of its Apollo 13 spacecraft so that it could troubleshoot problems without risk to the mission.
Now, companies are tapping the potential of “digital twins” — a dynamic model of a physical system that enables fast and creative experimentation at low cost and risk.
Writing in MIT Sloan Management Review, researchers Pushkar Apte and Costas Spanos say digital twins are at “an intriguing inflection point — transitioning from a specialized, tactical domain to becoming strategic tools with diverse applications.”
Multiple technologies have come of age that can now be combined in unique ways: sensors connected by the Internet of Things, 5G communications networks, and realistic 3D visualization enabled by advances in augmented reality and virtual reality.
Artificial intelligence and machine learning techniques can help model how a system functions and sometimes even predict how it might work in future scenarios. “Now we can do far more than just observe,” the authors write. “We can diagnose, control, and sometimes even provide a prognosis for diverse physical systems.”
Apte and Spanos see three high-level priorities where there is strong potential for harnessing digital twins: sustainability, smart innovation, and health and safety.
3. “Improved failure” might sound like an oxymoron. Not for Kara Penn, who explains how bold managers can exploit, design, and use failure as an asset in “Fail Better: Design Smart Mistakes and Succeed Sooner,” which she co-wrote with MIT Sloan senior lecturer Anjali Sastry.
As founder and principal consultant at Denver-based Mission Spark, Penn harnesses ideas from community development, management, and systems thinking to improve the social sector.
In a recent Q&A, Penn, MBA ’07, talked about risk-taking and the role of reflection in generating ideas:
- I co-authored “Fail Better” with Anjali Sastry because we were curious about why some efforts succeed and others fail. We developed a method and a set of tools that use the sandbox of your everyday project-based work as the driving force of new ideas and experimentation.
- One thing about the process described in “Fail Better” is creating space and time to embed the learning that comes out of the work you do every day. This is the most overlooked part. We run off to the next thing without reflecting in structured ways and without judgment on what we’ve just experienced.