High Performing Companies are best served by DATA - here's how to build a culture of data science

Questions, they say, will provoke your thoughts. Let me begin by asking a subjective one. What facets would you look for in a career with Data? Cutting edge Tech., Cross team collaborations and problem statements (Acquisitions, Retention, Payments, Product, Customer Experience, Fraud & Risk Operations, Financial Control, Quality Assurance), a growing industry. Well, all the above items are a reality at Junglee Games. And while we create Games, this quintessentially is an entertainment business which focuses heavily on User Personalization. It is hard to over emphasize the importance of analytics within the gambit of our work. But, with it comes certain challenges.

There is a need for consistent sense of data evangelism and what can be done in given time frames.

Each Data team member is effectively a sales rep for swift and reliable decision making, that covers all bases (Think OLYMPICS - Citius, Altius, Fortius). Variability in multiple business streams always remains. And not to undermine the sheer overlapping nature of business goals tagged to you. Sounds Daunting? Allow me to break it down, as I walk you through my experiences

Start with People, Process & the Team

Excellence is a process. Both in terms of personal and team growth, not merely jumping on the high tech train. While there is no denial on the tools and technologies needed to reach a destination, the idea is to look at the wider picture and have more folks on-boarded with you, vs just the data team. Your real team is composed of product owners, management, technologists and many others, because the work you do and the insights you obtain inevitably impact the entire business. You don’t exist in a vacuum, but rather are the conduit between what data you have (or, more often, don’t) and what the business needs to improve.

Source

While the above info-graphic could serve as a guideline liaison structure, where each item is pivotal for success, the need is to essentially connect the dots (did someone say "Are we on the same page?"). This is where critical processes come in. And they need to be done in consultation with teams, both internally and externally. Some typical items that we keep on Junglee's process list are Recurring Business Analysis, Data Requests and Change Management, Business Alerts Ownership and Escalation Matrix. Plus, there needs to be a robust Knowledge Management Repository (someone again said "Documentation?") which serves as a Single Stop Shop for Data Discovery. What it brings about is a Data driven Company Culture, Stakeholder involvement and collaboration.

Data Team Assembly and Evolution

The necessary and preferred Data Science skills can be seen here. And there are broad job titles that are around. Data Analyst, Business Analyst, Data Scientist, Machine Learning Engineer, Data Journalists, Data Architect, Data Engineer, Application / Data Visualization Engineer. There is a consistent need to hire in the Data Science world, not to fill up these roles, but to ensure that there is an alignment to company culture and no dearth of "implementers". The implementer is a loosely defined term, as it takes shape basis the issues. It could be a quick iteration to understand the data funnels, or putting together a visualization, or figuring out a simulation, or deploying a model at scale. This means adding on complementary skills at the beginning, which fosters collaboration and cohesion at a later point in time.

And there will always be time to focus on specializations and strive for the much needed evolution. The necessary steps could be

• Decentralized - Sporadic Analytics Resource Allocation & Utilization across functions
• Functional - Most Analytics Resources in one function, with little to no coordination across enterprise
• Consulting - Analytics folks working together, with "hire" philosophy applied by functions
• Centralized - Using Analytics in Strategic activities. Better funding and Resource Management. Creating federated teams from time to time will be additionally helpful

Living it up in the Matrix

Data Architectures, Feature Engineering, Algorithm Selection and Model Evaluation are all pretty and nice to have in your team's armory. Except that results are of pivotal importance which is what matters. While juggling priorities, negotiations, committing to deadlines, one of the overlooked capability is to communicate effectively, both internally and externally. Soon enough there will be know-how on who is good at what Data Science element, and then align structures to have people feed from each other. Have them broaden their horizons and, again, turn that into a repeatable process. You don’t need to step in to steer the discussions, but remember that personal growth comes from learning new things and passing them on, and that your team will be more effective (and happy) if processes are clear to everyone and if they value those processes as an integral part of their work. Plus, be candid in feedback and share it real time. This will build trust. And do not forget to offer support in a timely manner when an individual is stuck, but generally get out of the team's way (not for too long though)

A piece of advice to budding Data Science Managers

While it is important to understand the depth (life cycle) and breadth (tools and techniques) of Data Science projects, structured thinking and story-telling skills with an aptitude to ask questions and iterate on solutions is a must have! The essential leap that one always needs to lead a data science team (and thereby an organization) is to understand that it is not possessing the ability to answer the questions; it is about empowering the folks around you to be able to answer the questions.

It is not possessing the ability to answer the questions; it is about empowering the folks around you to be able to answer the questions.

Bottom line is always to not worry, it's Data Science, you are not building rockets just as yet, although great rocket ship launches are empowered by great data scientists. Regardless, when you are bogged down, just keep your head up and keep knocking down the barriers, till you get used to them. Trust that there will be a lot of time to do things, as changes evolve much more slowly than they actually seem to, because of the necessary human element involved in any organization, which in all fairness is good.

And lastly, Keep it simple. The barriers to Data Science acceptance in a lot of organizations is mostly due to the inherent digital culture, or the lack thereof. Efficient data processes challenge C-level executives to embrace horizontal decision-making. Front line managers with access to analytics have more operational freedom to make data-driven decisions, while top-level management oversees a strategy. This reduces management effort and eventually mitigates “gut-feeling-decision” risks. Basically, the cultural shift defines the end success of building a data-driven business. As McKinsey argues, setting a culture is probably the hardest part, while the rest is manageable.

Thankfully we’ve been able to build that culture at Junglee Games. Results are obvious with early profitability, close to 100% year on year growth in spite being a late entrant in the industry. A culture of Data Science has enabled Junglee Games to be one of the largest gaming businesses in India. It’s a challenging, high paced and exciting environment. And yes, we are hiring. And we welcome people smarter than us so we all get to learn and build something meaningful, together.