6 Tips for Setting the Price of Your Data Product

Building your data product is only the beginning. You’ve considered a wide variety of use cases, and settled on the one you’ll focus on. Maybe you’re going to help hospitals predict emergency room visits and optimize their staffing. Or you’re going to enable restaurants to reduce their food waste. Or maybe you just have some really unique data that you think might be of use to someone. Once you’ve determined your first “P” (the product), it’s time to think about the remaining three “Ps” of the famous four Ps of marketing: place, promotion and pricing.

Pricing has stumped product teams since the dawn of time (well, almost). As the data economy began to take off, “What’s my data worth?” became a question more and more people began to consider. Board members and company executives salivated at the thought of putting a line item for data on the balance sheet. The problem was no one really knew how to figure out what it was worth. Esoteric formulas, including measures of consistency or frequency, might sound good at conferences but it’s hard to actually put them into practice. The best way to determine data’s worth is to put it to use by creating a data product.

Some companies do implement chargeback if a product is used internally. However, the pricing question arises more often when taking that data product to market. The question should first be broken down into two parts: 

• How to charge for my data? 
• What to charge for my data?

To start with, the pricing model you decide to use answers the “how to charge” question. You must decide how you will structure the payment for your product. 

There are several pricing models to consider:

• Freemium: With a freemium model, there is no charge. Often it’s for a subset of the data, either a certain number of records or data from a limited period of time. A seller might use a freemium model to seed the market to allow students, developers or data scientists to give it a try. A freemium model can be used with any of the others models, like Snowflake’s Try-Before-You-Buy option. 
• Fixed fee: A pure, fixed-fee strategy offers access to the data product for a one time fee. Usually these models are for data that doesn’t change much, and nothing is being added to the data over time. With a fixed fee, you buy it; it’s yours. It’s like an all-you-can-eat buffet: Sometimes you use it a lot and get a lot of value. Other times not so much. But it’s yours.
• Subscription: When the data is more dynamic, you might want to use a subscription model offering access over a specific period of time. When the data changes more rapidly, it maintains its value and is often used close to the time of origin. In this case, subscription terms should be shorter periods of time, reflecting more continuous use. A longer time frame would be used if the data doesn’t change much and is used periodically over a longer time. When data is consumed more in initial months for planning purposes but is used to audit or adjust programs throughout the year, an annual subscription suits. For example, certain demographic data, which doesn’t change, might be used in marketing campaigns throughout the year. For sales transaction data, a shorter subscription term would be better. 
• Usage-based: A usage-based model charges customers only when they use the data product or service. Rather than an upfront fee or periodic payment, the consumer pays at the time of use. A usage-based approach means that companies can dial their use of data up or down depending on their needs, and not pay for any more than they use. This flexibility better reflects dynamic business environments. A usage-based approach is value-driven, directly linking the price paid with the value generated—or not. 

Usage-based pricing isn’t new. Most of the things we buy are usage-based. We pay our water and electricity bills on a monthly basis. Most of us, unless we are building reserves for Armageddon, buy food and other supplies on an ongoing basis. We’re all accustomed to the concept of usage- or consumption-based business models. And, most of us are horrified at the thought of wasting money on something that we don’t use. Same with companies. As a result, the adoption of usage-based pricing is accelerating, to the point that one expert sees an “inevitable shift to metered pricing for SaaS.” And, the rise of AI will catalyze that shift. 

The next step with usage-based pricing is to select the pricing metric. What element of usage will pricing be based on? Snowflake pricing, for example, is based on the use of storage, compute and cloud services. For listings on Snowflake Marketplace, data providers can opt for usage-based pricing by billable event or by query. With query-based pricing, consumers pay a fixed price for each query run that accesses paid data. With event billing, data providers can charge a price for specific types of usage of a data application. For example, consumers can be charged per row of data modified or used by the application, or per procedure call made by the application. Other events can be defined in application code. But that’s just the pricing metric. 

The question of price—and by that we mean the price tag, the sticker price—comes up often in discussions of data commercialization. Frankly, there is no easy answer. There’s no silver bullet for setting a price. This list of questions might kick-start the internal discussion:

• What are customers willing to spend? 
• How have others priced similar data? 
• What value have we derived from the data internally?
• What additional value could be derived from it? 
• What price are we happy with as a business? 
• What did it cost to develop and deliver? (as a last resort)

Most of these questions are difficult to answer. Do you know how much you’d be willing to pay for a pair of jeans? Or a meal at a nice restaurant? We might think we’ve got an idea, but I’m sure we can all remember blowing our budgets on a unique purchase or experience in the past. 

In the face of the questions listed above, here are six key things to consider when trying to determine pricing for your data product:

1. Don’t ask them how much they’re willing to pay. 

In the same way people blow their budget on jeans, they’ll blow the budget for a unique data set if it’s valuable enough. Asking someone what they would pay is a futile exercise, particularly if the data product is truly unique. If your value proposition is true, that your data product will deliver differentiation, then they likely have no idea what it’s worth or what they are willing to pay. You can ask, but don’t expect an answer you can use. A recent discussion of innovation pricing concluded that “Assessing a customer’s willingness to pay is a critical discovery activity ... but asking customers what they might do in the future leads to unreliable feedback.”  

2. Consider the value you’ve derived internally. 

Some companies extrapolate from a value that they derive with an internal use. For example, a telecom operator uses its own network traffic data to select a new location for its next retail store. That crowd density data might also be of use to a restaurant or retail chain expanding with new outlets, or a real estate developer who needs to recommend sites to a customer. Value derived from an internal use case can inform the pricing of a similar use case for an external buyer.

3. Collaborate to estimate incremental value. 

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. 

4. Consider the costs of development and delivery. 

While cost-based pricing is not the best way to go, it’s certainly not to be ignored. In some cases, it will serve as a starting point or a minimum viable return, at least to break even in the short term. You don’t want to be justifying prices based on how much it costs you, but at the same time knowing the cost of goods sold (COGS) is important to calculating your ROI. 

5. Iterate, and don’t be afraid to increase prices. 

At some point you just need to put a stake in the ground, triangulating all of the inputs collected: know your costs, estimate the value they deliver, test new offers, and iterate to find an appropriate price. The key to finding the right price is to adopt an agile approach. Some data providers go so far as to say it’s like target practice, narrowing in on the bullseye by testing prices and gauging demand at each price. If the price is too high, no one buys; too low and the product flies off the shelf but value is left on the table. If the latter is the case, a price increase is likely warranted, as difficult as it might be to do. 

In a recent webinar on The Principles of Pricing, Andreas Panayiotou, Director of Pricing and Monetization at Notion Capital, shared the story of a portfolio company leader who claimed, “Every time I win a deal I double the price until people start complaining.” Of course, the increases will depend on the uniqueness of the products, and the perceived value. 

6. Build in increases with a tiered model. 

Another strategy for increasing prices is to add in new features, creating a tiered product offering. Tiered pricing facilitates entry and upgrade, and offers a runway for increasing share of wallet. An initial freemium model might seed the market, allowing customers to access selected data. The standard offer would provide access to the full data set, maybe with some basic analytics. Premium models add additional features, and an Elite offer comes with more custom capabilities. However, the tiered model doesn’t preclude the need to raise prices on one or all of the tiers. 

The bottom line is that pricing is a difficult exercise. As Andreas described it in our webinar, “Pricing is like going to the gym.” Basically it’s going to hurt, until you build the muscle with continued practice. Even when you get better, a new exercise might cause some pain. As they say, “No pain, no gain.”

In a previous blog post, we talked about the data product team. But when commercializing data, that’s just a starting point. With today’s burgeoning data economy, creating a data product is just the ticket to enter this market. After creating the product, if it’s monetized externally, that extended product team would need to account for the resources required to effectively take the data to market. Those resources would include a role or roles focused on the pricing of the product, the distribution or placement of the product, and its promotion. 

For more on the data product journey—from identifying audiences, use cases, and the form of the data product to pricing options and choosing which channels to market—please watch the on-demand session on Data Commercialization: Your Guide to Taking Data to Market from Snowflake Summit 2023. If you’d like to hear more about data products and monetization strategies, or want to share your own stories, don’t hesitate to reach out.