Data as a product is an approach to data that can create significant competitive advantage, according to McKinsey & Company.
Veeral Desai, senior expert data architect at McKinsey & Company told Digital Nation Australia that the concept of data as a product is unique to traditional approaches to data.
According to Desai, “[Data as a product] really applies a product thinking mindset to data and it packages it up in a consistent way that makes it useful for many customers to build different things, using them. So here, I would think of Lego blocks where you'd snap multiple pieces together, which are data products, to build something that's useful.”
User demand and user-centric needs are critical in building data products said Desai.
“You understand what [users] need, you understand what they want and you collaborate with the end consumer of a data product throughout the process to make sure it is what is required,” he said.
Being able to publish your data in a data catalogue is another characteristic of data products, so that users can easily identify them he said.
Desai also names strong ownership and governance around data products as key to their trustworthiness.
“It has to incorporate all of the knowledge that comes from the business, like a particular person in the organisation might have a particular view of how a particular attribute has to be defined. So having that ownership and governance around it is key,” he said.
The benefits to a business in considering data as a product are five-fold:
- Lowering risk of failure
- Better real-world contextualisation
- Higher quality
- Cost savings
The main impediments to the data product approach are cultural in nature Desai said. He names one challenge in the approach as the shift in ownership.
“The shift towards a more business-centric owner, who's got a P&L responsibility. This leads to some initial friction and it's due to the fact that there's the loss of power from that original central authority, but also, not the largest amount of confidence that they can deliver and trust in the governance process,” he said.
“This is often overcome by establishing something like a centre of excellence which is able to define these governance standards, bring IT along to support the technology side of the equation. And using the business as a key input into the way that data product is designed. So being a bit more cross-functional alleviates this."