In partnership with

Unpacking Data Products in the AI Age: A Deep Dive for Product Managers

Hello Product Leaders!

I’ve been working in the data product space for as long as I can remember, although I’m not as experienced as Sanjeev Mohan (Find him on YT and Medium). I’ve had the honor of chatting with him a few times and gaining insights into his perspective on data products in the Age of AI.

Let me share some of his wisdom, combined with a product management angle.

But first, if you’re not familiar with the term, let me clarify. "Data product" has become increasingly prominent, especially with the rise of AI tools like ChatGPT.

In simple terms, it's a solution that turns raw data into insights.

Just like an advertiser, you trust Facebook for accurate targeting, a data product builds trust and provides a reliable experience by combining different pieces of data to create a differentiated product.

Today, we will focus on what to consider to build a successful data product.

Let’s dive in…

My Favorite Finds

👉️ With Brainnote, speak your thoughts and turn them into actionable ideas!
👉️ Kraftful analyzes user feedback from various sources, helping product teams gain actionable insights quickly and efficiently.
👉️ Save hours per week in PRD creation with our Free AI PDR (beta)
👉️ With Julius, Chat with your files and get expert-level insights in seconds
👉️ Spreadsite transforms spreadsheet data into interactive web dashboards and custom websites.
👉️ MinusX is a Chrome extension that automates data analysis in popular analytics tools, enabling natural language data exploration.

Recent Issues:

How Imperfect Plans Can Drive Success (Link)
6 Ingredients of AI Product Strategy (Link)
How to lead effectively when it's not your call (Link)

1- What’s a Data Solution…

A data product is designed to solve business problems by providing easy access to important insights for non-tech users.

For instance, Amazon might create a "customer purchase behavior" data product that combines purchase data from Amazon, Zappos, and Whole Foods, aiding decisions in marketing and supply chain management.

Similarly, Roche could develop a "clinical trial insights" data product, aggregating trial data and using advanced analytics to help researchers speed up drug development, ensuring consistent data use and reducing time to market.

My current company’s solution uses AI/ML to turn big datasets of clinical, scientific, and commercial data into useful information. This helps drug makers get their medicines to patients faster.

2- A Data product is always on the move

Creating a data product requires more than just compiling data; it demands a rigorous data quality management approach.

This involves overseeing a single piece of data throughout its lifecycle, ensuring it meets user needs, and maintaining its quality over time.

Super Hard trust me!

The product manager’s role is critical in ensuring that the data product remains functional and adaptable to the ever-changing business landscape.

For instance, consider a tech giant like Google, which may manage a “user search behavior” data product. As search trends evolve, this product needs to update its algorithms to reflect new behaviors and data sources.

Unlike pure platform play, continuous improvement is essential for preventing data contamination over time.

3- What’s a data product moth?

Not all data-related tools qualify as data products. A pure data product offers more than just a data set or a report; it provides an optimized user experience and addresses specific business needs.

Think about a company like IBM creating a tool to check how well their AI models are working.

This tool would do more than just show simple data; it would also have features to monitor, analyze, and improve the models over time. Another one is who built a data moat is CB Insights. You may have heard of them.

Just sharing a data set or a single report isn't enough to call it a data product.

Those killer data apps are all-in-one, offering regular updates with accurate, up-to-date information, and easy-to-understand values for its users.