Data Oracles
The best predictions come from deep analysis of today's data challenges and trends.
Most businesses are so agonisingly slow at change that the best predictions come from those providing deep analysis of today's data challenges and trends.
The term ‘oracle’ might bring to mind images of ancient seers uttering cryptic prophecies, but the data oracles of today share a different kind of wisdom. They are clear-sighted analysts, researchers, and writers who draw upon a deep understanding of the present to provide valuable insights for the future.
One of these data oracles who has deeply influenced my data work is Tom Davenport. A seasoned academic and writer, Davenport has written or edited 20 books and over 250 articles. Perhaps most famously, he co-authored the 2012 article "Data Scientist: The Sexiest Job of the 21st Century" with DJ Patil. This piece didn't merely predict a trend; it accelerated a trend by sparking a wave of interest in the field. I know people who pivoted their careers to move into data as a result of the article.
Davenport's book "Competing on Analytics" proved invaluable guidance when I was leading a major multi-year data transformation. Despite being nearly a decade old when I discovered it, its insights remained remarkably relevant. The book is a proven playbook for creating businesses that get value from and compete on data. I've yet to find another resource that rivals it.
One of the most valuable tools the book introduces is the DELTA model (roadmap captured below) for measuring an organisation's data maturity. This model has been extensively tested by the International Institute for Analytics (co-founded by Davenport), and they have shown a correlation between a company's revenue growth, average return on investment, and innovation. This correlation provides a compelling, board-level argument for why companies should invest in data transformation. I've personally presented these stats to multiple boards to underscore the necessity of top-down organisational change for scaling a data transformation.
Many organisations have not learned from his research and have ended up in what Davenport described as the “terminal stage” where, in the absence of exec support, data transformations stall.
Davenport's insights, much like those of a reliable oracle, aren't flashy prophecies about the next big thing in tech. Instead, they're well-researched observations and advice on current data trends and their implications. And, just as in ancient times, where it took time for the wisdom of the oracle to be understood and appreciated, it seems that Davenport's grounded understanding of the present takes businesses a generation to catch up.
If the past is a good predictor for the future, then Davenport's three books on AI and business likely contain some of the most accurate predictions for how AI will change business in the next ten years. I'd recommend starting with "All in on AI", or if you are looking for a shorter read, go to "AI and the future of work: what we know today".
Davenport isn't the only data writer I'm considering as an oracle, although his thirty-plus years of writing provide more proof points than newer data writers.
For me, three other data writers that are verging on oracle status are Barr Moses, Laura Madsen and Rachel Thomas. Like Davenport, they all write from deep expertise in their areas and research.
I was very nervous when I first had responsibility for data governance. It is a topic that is so abstract and removed from value. My gut questioned most of what I read. And I read a lot. I desperately wanted to understand this area. Eventually, I came across Laura Madsen's and Barr Moses' writing, and everything fell into place. They are both rethinking how data governance and writing with a clarity that will still ring true in 10 years.
I highly recommend reading Laura Madsen’s “Disrupting Data Governance” and following her posts on LinkedIn as she starts to question fundamentals such as Data Owners. Similarly, I’d recommend following Barr Moses on Medium and starting with her article on “What is Data Observability”.
More recently, I've been trying to move my thoughts on Data Ethics from interesting debates to solid, tangible topics. I recently came across Rachel Thomas’ work. She is most well known for co-founding the most popular deep learning course in the world. She has also written several of the most thorough and grounded pieces on data ethics that I know of. I'd recommend starting with AI and Power: The Ethical Challenges of Automation, Centralization and Scale.
In a world where data trends shift rapidly, these data oracles provide a steadying hand. They don't claim to predict the future, but their understanding of today gives us guidance that remains relevant for years, even decades, as businesses slowly catch up.
It reminds us that, while we can't predict the future with certainty, we can equip ourselves with a deep understanding of the present – and perhaps, that's all the guidance we need.