Design Data Products with Porter’s Five Forces

Data monetization strategy is one of the most critical data strategies for companies. One of the most wrongly implemented data monetization strategies is creating data products.

An article on Demystifying Data Monetization from MIT Sloan Management Review presented two primary paths to data monetization: Internal and external. The first focuses on leveraging data to improve a company’s operations, productivity, products, and services. The latter path focuses on creating new revenue streams by making data available to customers and partners, i.e., commercialized data.

The commercialized data part should not only be just raw data. To maximize data monetization, these raw data should be computed and processed, and transformed into business insights that are more valuable to customers and partners. This is pretty much the concept of the data product.

In the last few years, most data teams must have been asked by the upper management to start producing data products for our customers and partners. Unfortunately, most management teams do not provide adequate direction or precise descriptions of what these data products should deliver for the customers and partners.

Would a robust search algorithm be a good data product? Would a GNN-based anomaly detection be a good data product?A robust search algorithm or a GNN-based anomaly detection is probably not an ideal data product in business cases.

They would be a great tool to produce a data product to meet the needs of customers and partners.

Generally, data products can be categorized under any of these three stages of data products: Perscription, Prediction, and Prescription.

• Description data product aims to answer the question “What happened?” (or What is happening?). Customers and partners can then use this information to diagnose the situation more accurately.
• Prediction data product aims to answer the “What will happen in the future, such as trends and events? “ This usually occurs after the description stage and utilizes raw historical data from the description data product.
• Prescription data product aims to answer the “What should we do next, given the trends and events predicted in the prediction stage?”

The given three stages of data products are a vast scope of creating a data product. It should give management teams a clearer picture of data products and their stages.

Data Products Strategy with Porter’s Five Forces

However, the larger question remains, what would be the customer value proposition to convince a customer to purchase our data products? Answering this question has been the bane of business for centuries.

Instead of doing the SWOT to identify the gap, Porter’s Five Forces Framework can be utilized to help identify customer value propositions in terms of its profitability. The goal of the data product should allow its customers to make a profit; Porter’s Five Forces provided an excellent framework for determining what would be a good data product for our customers and partners.

What would be a great data product to provide our customers and partners with bargaining power with the suppliers? Perhaps a Raw Material forecasting data product that can offer a precise snapshot of the current situation or a quick future trend could give the company can better negotiate power with the suppliers.

What about a data product that warns cutomers when rising competitors are on the horizontal? A simple logistic regression can be implemented in a Competitors Monitoring data product. The company can then react promptly before it becomes irrelevant.

How about a data product that helps a company better compete with rivalry? For example, NLP can be utilized in an Employee engagement data product such that the company could have better employee retention.

An excellent data product is not just a data science or a data engineering product. Its design should not start as a great data model with pipelines but should start as a great customer value proposition for customers and partners.