(US and Canada) Sharath Gokula, Director of Data Science & Analytics at Walmart E-commerce, speaks with Tristan Spaulding, Head of Product, Acceldata, about measuring productivity, centralization versus democratization of data, management’s challenge to rationalize data, today’s data professionals’ missing skill, and data leaders’ entrepreneurial spirit.

From a data analytics perspective, Gokula says, many frameworks have been created to measure productivity. However, measuring the core value of all operations, including marketing, finance, analytics, and data science, is difficult.

He hopes that as an industry and as data professionals, they should be able to measure themselves before measuring the world around them.

Gokula highlights the cloud transitions happening in organizational structures and management divisions. This shift has resulted in data centralization at Walmart and other organizations. The focus is on what should or not be centralized. Understanding the customer and generating ideas are fundamental to business operations.

Cloud providers cannot provide packages for everything. An organization needs to have a good sense of what they are doing from a data strategy standpoint. Centralization, to an extent, is necessary. It helps transfer information from one to the other in the hub-and-spoke model. 

Regarding management challenges due to the power and diversity of tools, Gokula says counterproductive shadow data departments are driving the need for technology centralization. He maintains that a solid data governance layer is required to ensure no shadow data departments are created.

Data professionals, he states, are people who get to see the business aspect of everything. They are curious, analytical, and understand business, but they do not recommend their data insights regarding what should be done. He adds that data people lack a voice in a room regarding business decision-making.

In conclusion, Gokula says the entrepreneurial spirit is showcased when a data leader can decide with 70% information rather than wait for the complete removal of ambiguity.