At Rev, Nick Elprin, Domino's CEO, continued to provide insights on managing data science based upon years of candid discussions with customers. He also delved into how data science leaders can utilize model management and help their companies become successful model-driven organizations. This blog post provides a distilled summary of the whitepaper, "Introducing Model Management". The whitepaper is a companion to his talk and is also available for download.
The Model Myth
Domino's latest whitepaper, "Introducing Model Management", covers how Model Management is a new organizational capability for companies that want to put models at the core of business processes. As models are the central output of data science, they have tremendous power to transform companies, industries, and society. Yet, despite the advantages of being model-driven, companies are stuck trying to get there. A recent MIT Sloan study found only 5% of companies were extensively utilizing models in their business. Why is this happening? Companies are treating models like software when they are, in fact, very different — this is the Model Myth. Even though models look like software and involve data, models have different input materials, different development processes, and different behavior.
Overcome the Model Myth with Model Management
To overcome the Model Myth, companies need to develop a new organizational capability called Model Management. Previously, model management referred to monitoring production models. Yet, it should encompass a much broader capability. Just as companies have built capabilities in sales, marketing, people management, finance, and so on, they need an equivalent capability in data science. Model Management is a new category of technologies and processes that work together to enable companies to reliably and securely develop, validate, deliver, and monitor models that create a competitive advantage.
Data science leaders and organizations that successfully build a Model Management capability will reap exponential rewards as more models drive better customer experiences and better margins. As models build on each other, more models also means more data and capacity for organizations to invest in new and better models. Those organizations will also better navigate common pitfalls that stymie the impact of models such as ethics and compliance risk. Ultimately, the haves and the have nots of this next era of computing will be determined by the quality of an organization’s Model Management.
The gguide, "Model Management", synthesizes Domino's learnings, distills the problem, and proposes a path forward to achieve the full potential of data science. Part One describes what a model is and discusses how models drive business value. Part Two focuses on the essence of the problem — that models are different from anything built to date and it is a myth that organizations can manage them like other assets. Part Three dives into the details of a proposed Model Management framework, including 5 pillars, which addresses the unique properties of models.