In the new book The Digital Mindset: What It Really Takes to Thrive in the Age of Data, Algorithms, and AI, by Paul Leonardi and Tsedal Neeley published by Harvard Business Review Press and out on May 10, the authors attempt to do something big. They attempt to skill up business leaders with enough digital couth that they can help lead their organizations’ transformation agendas. Honestly, the book is a nice book-end to Beyond Digital, which I reviewed late last year.
To make a point that business leaders can transform themselves, Leonardi and Neeley tell the story of Sara Menker. Menker during the summer of 2008 was a Wall Street trader. At the start of the great recession, she turned her attention from a collapsing real estate market to the impacts of where she was born.
Ethiopia had a history of catastrophic famine. And given her business experience, she saw the opportunity to create a digital startup that would use data to better structure business risk. To do this, she had to deal with the problems that plague most digital innovators including scattered data and the need to create predictive models.
Menker developed from her process what Leonardi and Neeley call a "digital mindset." With this, it was possible for her understand how a powerful digital platform, purpose built to help connect fragmented data sets, could help revolutionize agriculture and deal with famine in sub-Saharan Africa. (Page 4).
Clearly, many CMSWire readers are digital practitioners. However, the book aims to give your business leaders the vocabulary, knowledge and intuition to see the bigger picture that digital practitioners work with daily. The goal is for these leaders to do what Theodore Levitt suggested in Thinking About Management: ask questions and thereby, truly understand the impact from digital technology.
For business readers to be part of the digital practitioner equation, they do not need to become a master coder or even a data scientist. They need instead to “understand what computer programmers and data scientists do and to become proficient at understanding how machine learning works, how to use of A/B tests, how to interpret statistical models, and how to get a chatbot to do what they need it to do.” (Page 11).
The authors go onto suggest leaders don’t even need mastery of the topics presented but instead need to understand 30% of the topics. With this, they can add value add to the digital discussion.
The authors say digital is fundamentally about the interaction between data and technology. To succeed, business audiences need to develop a digital mindset. Leonardi and Neeley take an anthropologist’s perspective on mindset when they say it is a set of approaches to make sense out of the world. A digital mindset is a set of approaches used to make sense of data and technology and matters because digital forces are reshaping how we live and work.
Related Article: Can CIOs Changed Fixed Organizational Mindsets for Digital Transformation?
Leonardi and Neeley claim the increasing pervasiveness of digital technology means business leaders need to reconsider how they interact with machines and people. In contrast to articles that proclaim robotic process automation (RPA) and AI are taking away jobs, they believe digital workers are needed who can use machines to drive better business outcomes.
For this reason, non-technical workers need to understand that working with machines is not the same as working with people. This is even as we move from what Geoffrey Moore calls "systems of record" to "systems of engagement" thinking. It is important that business leaders understand as well that the AI we use today focuses on only one task.
Powering AI is data. For this reason, Leonardi and Neeley want business leaders to understand that data must be cleaned and converted to formats that algorithms can understand and learn. “Cleaning involves fixing or removing incorrect, incomplete, or duplicate data. Data aren’t always collected (or created) in consistent ways and combining multiple data sources often results in duplications, incongruencies, and mislabel.” (Page 29).
Next, Leonardi and Neeley look at the building blocks of AI are machine learning techniques. They explain that while humans use complex natural language and visual cues, computers work with numbers to generalize from examples and gain the ability to learn without being explicitly programmed. This is, they go onto explain, what has changed since we all moved from operations research to big data, data modeling. In other words, today there is vastly more data and the ability to process that data. In Big Data for Dummies the authors write, “Big data is the capability to manage a huge volume of disparate data, at the right speed, and with the right time frame to do real-time analysis and reaction."
The authors go onto explain AI, Machine Learning, and Deep Learning. With that accomplished, they provide a non-technical view into Natural Language Processing, Supervised and Unsupervised Learning and Reinforced Learning. Then they explain the supporting tech stack.
Related Article: What's on Your Artificial Intelligence Compliance, Privacy and Ethics Checklist?
https://www.youtube.com/embed/0MqB_t5ZX0E
With a description of how digital workers should work with computers, Leonardi and Neeley go onto explain how work is done in an increasing digital, virtual world. They recognize that work is increasingly being done in a hybrid or remote fashion. They explain that working in a virtual environment changes how we need to interact with each other.
In remote work, you lose the advantages of nonverbal and spontaneous communication. These have been critical to creating shared experiences, understanding and social bonding. To succeed at virtual work, they say it is essential that digital workers develop a digital work mindset. This means learning to stay in sync with remote collaborators.
Leonardi and Neeley believe achieving a digital mindset requires a level of comfort with analytics. They stress data is not a natural substance. They say it is a misnomer to say data are collected. It is more accurate to say that data are produced. To add perspective to digital computing, they suggest that the oldest form of data was manual input — it occurred when people started to write things down.
The problem was manual inputs varied from bookkeeper to bookkeeper. This is the case for digital systems — ergo the earlier discussion of data hygiene. With this said, many data are no longer produced manually. Here, data comes from barcodes and other means of input into transaction systems. Next generation technologies like RFID have transformed organizations.
Leonardi and Neeley suggest that having digital mindset means you are data literate and comprehend the risks from the processes used to produce data. Even automated data is not free from error or consequence. Given this, “the first step in building analytical insights is to recognize when we are presented with data, we need to ask how those data were produced, who had access to them, and how well they represent the behavior or activities we hope to understand.” (Page 80)
At the same time, a digital mindset requires an understand of data’s fallibility. Data literacy is important after you drive for a data culture. Data doesn’t exist apart from the systems we use to classify them. For this reason, a digital mindset means thinking about how data are classified. This means learning to ask the right questions about data.
Related Article: Artificial Intelligence Is About Collaboration, Not Job Elimination
With an understanding of data in hand, the next step in developing a digital mindset is knowing the right questions to ask of data. This requires understanding how to use data to advance business goals. The first step in doing so is developing an intuition for the underlying patterns in data.
To be clear, the issue isn’t having data; it is interpreting data and drawing conclusions from data. To do this, statistics represents the best means for analyzing the underlying data patterns. More specifically, it allows digital workers to draw conclusions about a population from a sample dataset.
Hopefully, readers have some understanding and course work in statistics. But if not, the authors provide enough of a review that you can start to ask questions and use statistics for making sense of data. The goal is to act like a "data detective."
In contrast to one of the textbooks that I used in my coursework, the authors provide some interesting examples from Electronic Arts and McKinsey & Company. Leonardi’s and Neeley’s goal is help digital workers be conversant in the language of statistics so they understand the conclusions that can be drawn from a dataset.
Leonardi and Neeley suggest that people with a digital mindset need to understand their organizations will experience security issues and must be prepared to deal with the inevitable security crisis. They should be able ask two questions:
They make an important point — the castle analogy is not useful or accurate when thinking about cybersecurity. Michelle Dennedy, Jonathan Fox and Thomas Finneran in the Privacy Engineers Manifesto agree on this approach and argue for “data centric and person centric security.” Leonardi and Neeley also suggest digital ecosystems do not operate in a way that supports a castle analogy. Castle thinking doesn’t work with the number of entrances and how systems are constantly evolving.
For this reason, a digital mindset requires an understanding dynamism and increasingly decentralization of digital systems and ecosystems. To fix things, Leonardi and Neeley suggest leaders do three things:
Apart of embracing ecosystem interdependence, Leonardi and Neeley suggest that business leaders need to help IT leaders acquire budget to eliminate tech debt. Far too often a hack was the result of patch not being applied or new security approach not being fully implemented. It is critical that time and money are provided to ensure digital systems continue to work safely and appropriately.
In terms of designing for privacy, it is important that business leaders understand what Cambridge Analytica did with personal data. Today, privacy is a security concern and ISO 27001 and NIST 800-53 standards make it so. Digital workers need to know when data contain confidential and proprietary information. The authors suggest organizations following the best practices from Ann Cavoukian’s “Privacy by Design” including its seven principles.
Related Article: Privacy by Design (PbD): A Definitive Guide and Why It Matters
The speed and scale of change in the digital era makes experimentation a requirement. Leonardi and Neeley say rapid prototyping and data analysis improve internal work processes, products and services. Central to making this work is establishing a culture that embraces experimentation.
The world has changed, and when change is constant, it is difficult to use hunches or theories to chart the best course of action. For this reason, Leonardi and Neeley say, “the rapid development of digital products and services, the increasingly fickle habits of consumers, and the immense leaps in computer processing power and data storage capacity power made annually are among some of the reasons why today’s customers, companies, and markets are in a constant state of flux.” (Page 149).
CIO Pedro Martinez of Puig agrees when he said in a #CIOChat that, “Any great idea that requires buy-in from distinct stakeholders benefits from a proof of concept. A POC is worth a thousand slides for both convincing and gathering significant feedback at early stages of the project.”
In a digital-driven world, change isn’t something that happens periodically. It isn’t something for which you can scramble to respond. Instead, digital transformation occurs regularly and requires organizations to redesign underlying systems and processes to align with evolving data and digital technologies. I would add from Beyond Digital that it is about redefining market value propositions and business models.
This is like the thinking of Jeanne Ross, Cynthia Beath and Martin Mocker in Designed for Digital: “managers in digital organizations rely more on coaching that hierarchical decision making to optimize business outcomes.” (Page 100). So the transition impacts people, processes and culture.
For this reason, Leonardi and Neeley stress digital transformation isn’t a goal you achieve. It is a means to achieving changing business goals. To succeed, organizations must integrate data silos. MIT-CISRs research finds that 51% of organizations still have their data locked away in silos. As with supporting people and process, it is time to identify the right tool sets. During this process it may be necessary to upskill the team.
Without question, no business will be unaffected by digital transformation and every form of work will be changed. To succeed, organizations need technical resources, but it is just as important to transform the mindsets of business leaders.
This means building new skills and seeing the world in a new light. Part of this involves reshaping how collaboration takes place. Your digital mindset, say Leonardi and Neeley, represents your new superpower.
