How Do I Become Chief Analytics Officer?
Actively working towards that top job.
I would imagine that anyone in the data analytics / data science practice would relish the opportunity to be a Chief Analytics Officer (CAO) of an organisation. It’s an outstanding opportunity to really flex your intellectual “muscles”. I would argue that even those who loathe the responsibility of managing big teams aspire to be recognised for their abilities and contributions, and a CAO appointment can be a nice ego-stroke.
As youngsters, and even mid-careers, flock into the data analytics / data science practice, they are naturally wondering if there exist pre-defined pathways to the top of the heap, or is it somewhat haphazard and more to do with being at the right place at the right time, and having the right boss?
Granted that the world of data analytics / data science has evolved, with the practice now having a firm seat at the table in many organisations, the pathway to CAO and analytics leadership hasn’t changed all that much. As a former CAO and a practitioner in this discipline for 30 years, I believe I can provide a distinct perspective to this question. And so I dedicate my 54th weekly article to unpacking this topic.
(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)
Architect vs Engineer
Successful data analysts and data scientists must have an engineering mindset — you encounter a problem, and you figure out how you can overcome it in a logical and sustainable way. You diagnose the underlying factors of the problem, you test or simulate some possible solutions, and then you work towards its implementation. Many engineers will never grow to have an architectural mindset, which is the competency that is most required to be a CAO.
What sets a potential CAO apart from the flock is the ability to problem-find and problem-define, rather than problem-solve. Some would call this “big picture thinking”; it’s also related to design-thinking. Underlying this ability are a number of critical and essential competencies that need to be developed and honed over time. These are: data sensemaking (I’ve written about it here in article #2), and a knack for asking good questions (article #27). These are the kinds of competencies that AI cannot replicate today.
How should you go about acquiring this ability then? For starters, get involved in special projects and task forces; asked to be assigned if necessary. These projects usually have a strong strategic agenda to them, and it’s a great way to acquire strategic and expansive thinking skills. Secondly, as you work (as an “engineer”) on your solutioning, think about how your solution can be generalised, better monetised, and even developed into a competitive capability for your organisation. Doing this often will help to activate and strengthen your cognitive skills. Having a (good) mentor will also accelerate your learning rate.
Breadth and Depth
The path to the top is not vertically linear. You can’t code your way to the top job by becoming an ever better AI modeller. As a general rule of thumb, specialists don’t rise to become leaders. You need to have T-shaped skills to be a CAO — broad experiences across many data analytics / data science sub-domains and very deep skills in a few of them. If you started out as a data scientist developing ML/AI models, ask yourself:
- Do you have experience in visual analytics, dashboard design and KPI creation? Do you understand visual vocabulary?
- Do you have experience in campaign management, from simulation, pro forma modelling, A/B testing, monitoring, and evaluation?
- Do you have experience in pricing optimisation, or incentive modelling?
- Do you have experience in data warehouse implementation, and data modelling / data schema design?
How should you go about becoming more T-shaped? Job rotation within your organisation (rather than job hopping) is critical. Everyone wants to do the “sexy” work like advance modelling. But the non-sexy work need not just be “operational” in nature, and can be equally intellectually challenging if you allow yourself to be open-minded about it.
Decisions vs Data
To be a CAO, you need to learn to manage complexity instead of just complications. You need to operate as a “decision scientist” rather than a data scientist. It’s about creating impact and value from the data-oriented work. When you build a model, you are dealing with complications. The work isn’t easy, but it has zero value unless it gets properly implemented and utilised. Translating a data solution into a business solution requires you to think strategically about the decisioning ecosystems that exist across the organisation — how decisions are made and who makes them, and their downstream implications and consequences.
Affecting changes in decision-making is always messy and complex. Personal agendas often don’t align with organisational interests. And there are paradoxical truths in data-driven solutioning that can complicate matters — for example, implementing self-help analytics often leads to more work for the data analytics function and sometimes poorer decision-making. The ability to navigate these complexities is a necessary competency to becoming a CAO.
So how should you acquire this ability? You should get deeply involved in the implementation process of data solutions, in particular, the change management aspects. You should also try to insert yourself into various operating or steering committees to get a front-row seat on the politics of decision-making. Only then will you begin to appreciate the human dimension to decision-making that can either make or break a solution, regardless of how good or well-intentioned it may be.
Personality Matters
Last, but not least, personality always matters for any top job. The same is true for the CAO role. People must like working with and for you. You don’t need to be an extrovert (most data people are naturally introverts), but you need to (appear to) be approachable and congenial. Having the right personality allows you to “exert” influence in more constructive and productive ways.
This last ability is probably the most difficult for many data analysts / data scientists. The default belief is that you are what you are, and that personalities can’t change. Neuroscience would disagree with that view. But it takes constant and consistent effort to modify your personality. My go-to suggestion is to take every opportunity to train (i.e. teach others), talk (i.e. panel or guest speaking), and write (i.e. blogs, publication articles). Train-talk-write will build confidence, and help with your personality improvements.
Conclusion
The climb to the top can be arduous, and may not be for everyone. But it is not closed off to anyone. And if the CAO role is something you aspire to, my advice is to start early. Many of the needed skills and competencies require time to ferment and percolate; they require opportunities and investments. There will be some starts and stops along the way, but that is to be expected. Don’t give up.
