What’s in a Name? (Part 2) — Knowing vs Doing
Background
In my last article, I started a sub-series questioning the great proliferation of roles in the practice of data analytics. For example, in the last 20 years, we have created roles such as Data Scientist, Decision Scientist, Data Engineer, AI Scientist, Data Visualiser, Analytics Translator, Data Connector, etc. I question whether this increasing ‘speciation’ of roles serve us well in deepening the practice of data analytics. There is also a proliferation of names to describe the collective function of data analytics within the organisation. The problem with this role proliferation is that it is often accompanied by a proliferation of practice sub-domains (e.g. decision science is NOT the same practice as data science); it is also accompanied by a proliferation of department names where the Business Intelligence function became the Data Science function or the Decision Management function (an interesting name!).
In this second article of the sub-series, I will write about the evolution of roles that seek to inhabit the intersection spaces of business, computation and analysis (see diagram below).
(I write a weekly article on bad thinking and bad practices in data analytics / data science which you can find here.)
Seat at the Table
We know the trends. The rise of AI, both traditional and generative. The rise of digital automation. The rise of data-driven thinking. Data Analytics functions have long aligned themselves to these changing winds — in part to stay relevant, and in part to raise their profiles. Data Analytics functions are keen to get a seat at the executive management table. However, I would argue that most data analytics functions DON’T deserve a seat at the adult table because they don’t contribute to the broader discourse on business strategy. I have seen many a chief data scientists and chief analytics officers who can’t articulate or attribute the economic and strategic value of their work. This is one of the reason these roles have been made to report into other executive functions.
My corporate career has been in retail banking with Citi. I was the first data analytics head on the country’s executive management team, and was similarly on the regional executive management team subsequently. So I completely understand the struggle for that ‘seat at the table’ — you earn it by giving inputs into the operating strategy of the organisation. I used data evidence to not only back my opinions, but to also challenge and shape the opinions of those at the table. In short, a seat at the table equates to decision making capabilities.
In Citi, we rebranded the data analytics function as Decision Management as part of that functional elevation and seat-at-the-table. Having left corporate life for more than a decade, and with a lot more data analytics experience under my belt, I have had cause to re-think whether Citi pursued this approach in the right manner.
Decision Management vs Knowledge Management
Why didn’t business intelligence evolve towards knowledge management; it would seem like a natural fit. As shared in my previous article, the current state of knowledge management is a misnomer and evolved from the older practice of library science — which is all about the cataloging and sharing of information. There is currently a significant surge in interest in the field spurred by information overload and the rise of AI. This includes interest in the closely adjacent field of knowledge graphs. But Knowledge Management in its current state has very little to do with improving decision making or driving economic value creation. Instead, Knowledge Management seeks to improve the efficiency and effectiveness of the upstream input process into decision making.
But what about bringing efficiency and effectiveness to the actual decision making? Citi was amongst the first to use the term Decision Management to brand its data analytics function across the globe, originating out of the US. It was a controversial label with equal number of internal critics and supporters. The global leaders of the function (in US) emerged from Credit Risk Management and they borrowed the idea of “supporting informed decisions” as a north star. Credit Risk Management had a seat at the table; because they alone claim control over the flow of information (both the generation and interpretation of it) coupled with decision authority (e.g. credit approval). They applied this same ‘choke-hold’ thinking to the data analytics function and voilà, a seat at the table! (Of course I’m being somewhat rude here; the function did make real and significant business contributions.)
Knowing vs Doing
But there were real tensions between how the US defined Decision Management and how we thought about the ‘practice’ within Asia Pacific (APAC), where the data analytics function was differently organised because of its genesis. APAC’s data analytics function grew from the coupling of Business Intelligence with Database Marketing. There was a strong focus on execution capabilities and not just information processing. Even sales incentive management (design + payout computation) sat within APAC’s data analytics function. Citi’s global Decision Management didn’t see the value of owning the execution and wanted us to shed it. It didn’t happen. There was a revolt and down to the last man, we stood our ground to insist on the importance of execution ownership. The same revolt happened in Citi Mexico where their data analytics matured out of CRM (executional) capabilities due to the nature of their branch-based business model.
And this is the key to understanding the proliferation of roles and even function labels. The past movements in data analytics favoured the intellectual pursuit of Knowing, i.e. the generation and interpretation of information. Careers were built on that. Business Intelligence, Data Science, Decision Science were all in the pursuit of Knowing. The Citi US Decision Management function spent a great deal of resources developing pro-forma P&Ls (i.e. calculating financial results using certain projections or presumptions). It was mechanical. Without owning the execution piece (campaigns were outsourced to a 3rd party contractor), it lacked the deep ability to continuously probe and validate the assumptions behind those calculations. Saying something would happen on paper didn’t always translate in reality.
The need for Analytics Translators and Data Connectors is the gulf that has been created because we have neglected the Doing side of the equation. The old believe that execution was simply “dirty work” and not very intelligent is incorrect. There is a lot of intelligence at the execution end. I’ve long been an advocate for executional intelligence — campaign analytics, sales incentive analytics, etc. and have written about it here and here. Instead of role proliferation and arguing about labels, the practice of data analytics needs to grow up to enfold its arms around both the Knowing and the Doing side. One informs the other in a continuous cycle of learning and Intelligence. Data Analytics needs to be the pursuit of Integrated Intelligence.
Conclusion
There is a real need to simplify rather than role-speciate the practice of data analytics. The space has gotten too noisy. We need to stop this behaviour of finding niches to exploit into new labels and new roles. Instead, expand the practice of data analytics; be more encompassing. Recognising that both Knowing and Doing must go hand in glove will encourage an end-to-end integrated ownership of responsibilities that will shape a new generation of data analytics practitioners.
In my next few articles, I will continue to unpack these thoughts and explore its implications to formal and continuous education in data analytics, as well as its implications the evolution of data analytics leadership.
Stay tuned!
