In the book “The Adventure of Copper Beeches”, the great detective Sherlock Holmes declared, “Data! Data! Data! I can’t make bricks without clay.” In other words, pieces of data were the essential building blocks the famous sleuth needed to solve any mystery. But as we find ourselves emerging (yes, hopefully so) from a global pandemic, data is also the building block required for reinvention. As companies reinvent themselves in the new world, they use data to look more deeply into their own operations; they take a closer look at their existing customers and scan horizons for new customers. And, they explore ways that partners and customers can do the same. They not only monetize their data by using it internally but they begin to commercialize their data externally. The companies that see this opportunity tell us that they used to have a data strategy—now their strategy is data. 

This new paradigm requires thinking about data as an asset—yes, we’ve heard that before—but also as a product. That word gets thrown out a lot these days as well. In the data world, the data mesh paradigm includes “data as a product” as a foundational principle. The technical discussion of data products has somewhat rabbit-holed down an architectural path. While useful, the notion of “architectural quantum” might be off-putting for certain audiences. 

The basic notion of a product is fairly straightforward. The Oxford Dictionary defines a product as “an article or substance that is manufactured or refined for sale.” We should think of data as the building blocks of a manufactured or refined product for sale. And, by “sale” we mean exchanged for value, either internally or externally. Whether or not internal uses result in charge-back or other exchanges of value remains an open question. Some shared IT services models advocate for charge-back models, but often the mechanisms for doing so are complex and politically challenging. So, again, let’s put that aside and focus on the product itself.  

Get started taking your data to market

For many data leaders, the opportunity is obvious. A frequently cited McKinsey study estimates that data collaboration generates $3 trillion annually. And, data collaboration starts with data sharing—making data available for others to use and derive value from. In other words, it’s about monetizing data externally. However, most data leaders find even getting started to be daunting. Some of the question they ask include:

Which data? Often, the first instinct is to ask the data teams to scope out data products. Data teams have historically taken the lead, right? Well, OK, not necessarily this time. Or if they do, the first step is to talk to the business stakeholders to get an idea of which data sources or data “domains” might be of interest. Think “customer data” or “product data.” Customer data would come from multiple sources and include customer profiles, transaction data, contact center logs, and anything else that touches the customer. Product data might include production data, launch data, sales and returns, defects, and anything else touching the product. Data teams might know what data they have but not necessarily the extent of potential use cases.

Which use cases? The question then is, who would use this data and how? Some data teams spend way too much time trying to come up with those use cases themselves. That might work sometimes but it’s not committee work. The best place to start is by asking how data is already being used, both internally and externally, and then see if others could use it in the same way. For example, telecom operators use network traffic density in cities to determine where to place retail outlets. Lo and behold, retailers, developers, and urban planners can also use that data for site selection. Or imagine a related use: A jet engine manufacturer uses flight operations data to improve future products. The airlines can also use flight operations data to improve efficiency. Enterprise applications like ERP now come with benchmarking services that compare metrics like downtime costs or return on assets across customers. This is referred to as the adjacent possible, like “Do you want fries with that?”  

Which form of product or service? That brings us to the actual data product or service, and the different forms possible. It’s not always just about the data itself. Selling the data requires a developer or a data scientist to do something with it, such as building an application or analytic model, to deliver business value. However, if the data product or service is an application or an analytic model, delivering insights directly to customers within a business workflow, a decision or action can be taken immediately. For example, at PepsiCo the data team created an internal product, ROI Engine, to measure the impact of marketing campaigns and media placement. The app aggregates data from over 60 sources and delivers insights to users across the business from ecommerce to the different brands and regions. The insights enable marketers to determine which campaigns and ad placements were successful, and decide which to continue or refine.

Between the data itself and data apps (seen in the diagram above) are custom interfaces which facilitate discovery and access within a specific business context. For example, Atheon Analytics SKUtrak provides an interface to help fast-moving consumer goods suppliers and retailers make data-driven decisions with flow-of-goods analytics. As products and services deliver more derived insights directly into a business context (moving along the orange curve in the diagram), time-to-value accelerates. 

When offering the data itself as a product, direct data sharing provides a better option than copying and sending the data via a download or file transfer. Not only does copying and sending require more effort, the data is also out-of-date immediately as it is only a snapshot at a point in time. Access to the data is also more difficult if not impossible to revoke. 

What’s the value? This question has stumped product teams for all eternity (well, almost). What is my data worth, and what do I charge for it as a product? Some data providers have told me that it’s like target practice, narrowing in on the bullseye by testing prices and gauging demand at each price. Others extrapolate from a value that they derive with an internal use. Another approach is to work directly with a customer or partner to benchmark and measure incremental value accrued with the application of the data. For example, a marketing campaign achieves a certain conversion rate but with new data to identify specific targets, conversion rates increase. A share of the lift can be attributed to the data. In all examples, it’s about adopting an agile approach to testing new data offers, and determining the value they deliver. In the end, the market will determine the price.

How to go to market? For most companies, data commercialization is not their primary business. GE Aviation and Siemens Mobility offer data products and services, but they are still airplane engine and locomotive manufacturers. Successful commercialization often starts with the right go-to-market partners or channels. Many consulting firms and service providers help guide companies through the process. And, the emergence of data marketplaces makes data discoverability and access easier. The Snowflake Data Cloud facilitates data sharing (and selling), either directly with a customer or partner or via a data exchange set up among an ecosystem of partners. Instacart, for example, with over 500 million products in its catalog from over 40,000 stores across 5,500+ cities, shares purchasing trends with its retailer and CPG customers. 

With broader exposure and commercial features, Snowflake Data Marketplace provides a home to hundreds of data providers—and not just the pros. A growing number of Snowflake customers are exploring putting their own enterprise data into the Marketplace. In the vanguard, ADP, which processes about 25% of U.S. payrolls, has made its aggregated and anonymized geo-based U.S. workforce demographic and income data available on the Snowflake Data Marketplace. And others, such as  1-800-Flowers, are taking advantage of the Snowflake Data Marketplace to identify new data sources to enrich analysis and improve business performance. 

5 steps to start monetizing your data

In summary, to build and deliver data products and services, either to internal stakeholders or to external partners and customers, start with these five steps: 

  1. Establish data sources or domains as the raw material for data products or services.
  2. Identify potential uses by polling internal stakeholders, partners, and customers to determine existing uses and explore related, adjacent applications for the data.
  3. Determine the best form for the data product or service: raw, uncurated data; enriched data; customized interfaces; or context-specific applications.
  4. Adopt an agile approach to testing new offers and determining their value.
  5. Find the right go-to-market partners or channels, such as monetizing through the Snowflake Data Cloud.