2.5 quintillion bytes of data are created every day (source: IBM). In addition to legacy modes of data generation, new behaviours like social media posts, digital pictures and videos, activity tracker usage, and IoT sensor measurements have accelerated the pace of data generation. The astounding amount of data generated is hard to comprehend, let alone understand how to put to use.
Only a small fraction of the overall data generated is considered usable. However, once identified, usable data has tremendous value to an organization. The most obvious value is for internal use to enhance business intelligence and decision making around product development, risk management, customer targeting, and so forth.
The less obvious, but more quantifiable valuable proposition, is the benefit an organization can realize by participating in the data economy. Not all data should be stored in the company vaults. Instead, organizations can productize their usable data for commercial sale by external parties. These data products have the potential to provide new revenue streams while improving the common good.
1 – Recognize Which Data Are Valuable
The first and most introspective step of the journey is determining if data has any value to others. Given the amount of data organizations collect, this is no easy task.
As we have previously discussed, raw format data is often times the most valuable, even though it might have false positives and inconsistencies. However, this isn’t always the case. In other circumstances, datasets are more valuable when segmented into smaller and more focused data products, which can then appeal to different user groups. These focused datasets can take the form of analysed and enriched data such as insights, indices, patterns and associations.
Generally, prospective data vendors can take two approaches when deciding what data are valuable: 1) talk to potential buyers and external experts to define the scope of data or 2) test the market directly and offer what the vendor thinks is the best product.
The first approach would yield the most clarity and would help with pricing. Common buyers of data include management and technology consultants, multinational corporations, academics, and data analytics firms. Speaking with these buyer groups would help clarify what appetite exists for certain data. In addition, external experts such as channel partners, could provide insight into reference prices and an objective understanding of the market opportunity.
The second approach is a bit risky and could have drastic consequences. Without a firm understanding of the data market economy, a pricing hunch is nothing more than a high stakes guess. You might get lucky and get it right, or you could miss the board completely and have little insight into where you went wrong.
2 Structure a Data Product
Once earmarked for commercialization, a data product must be structured according to international laws & regulations and the information security policies of the organization itself. There are three main steps in structuring the data product.
First, all personal information must be removed. Selling identities is illegal just about everywhere in the world. Any data products for sale must be clear of personally identifiable information (PII). Vendors can still aggregate their user data into audience segments, sentiment indices and generic profiles, such as with digital advertising; however, all individual identity data must be removed.
Second, vendors should be diligent to ensure that the data products are not exposing company/organizational secrets that could erode competitive edge over their rivals. Simply put, data that could be detrimental to an organizations positioning or endanger public safety shouldnt be sold.
Third, vendors must still decide on the overall characteristics of their commercial data. The main considerations here are: timing (real-time vs historical), delivery (push vs. pull), and format (json vs. xml vs. pdf vs. csv). Organizations should choose the characteristics that will yield the highest overall appeal to the market and exceed costs incurred to produce it. Once all these considerations have been solidified, then, voila! A data product is born!
3 Price the Data Product
Based upon decisions in the previous step, a data vendor should now hopefully be able to assess the cost to make data available for sale. This effectively serves as the price floor for the data product. From there, the task of pricing data is a balance of art and science.
First, vendors must recognize the rule of economics with regards to price elasticity and willingness to pay. Price too high and products wont sell. Price too low and run the risk of leaving money on the table (also known as consumer surplus). With this in mind, vendors must understand what really drives value for their data products. At DataStreamX, weve coined these drivers as Three Rs of Data Value:
- Richness Accurate, robust, granular, detailed, frequent
- Reliability Authentic, dependable, complete, consistent
- Rarity Uncommon, irreplaceable, extraordinary
Keep in mind that the same dataset can be cut several ways and command different prices. For example, a company with data on gas stations across Europe might track several data points such as locations, number of pumps, and hourly car traffic. The company with this dataset needs to understand who needs the possible buyers of the data are as this will help dictate the structure and the pricing. For example, the company could create data products for each country or aggregate all of Europe. Separately, the company could chose to provide real-time, weekly, monthly or yearly traffic to each of the pumps. With just these parameters alone, there are at least 400 combinations for data products, each with a distinct value to the market.
4 Create Demand For Data Products
If the primary consumers of a data product are easy to define, segment and access, then vendors may find it useful to employ a direct sales force. This is especially true if data sales are a primary component of the vendors business model. Twitter, for example, trades in social media data and the flow of high-paying customers thirsting for the Twitter Firehose is seemingly endless at present. In this situation, the high value of each Twitter transaction properly justifies a deep, and expensive, sales force.
In most cases however, vendors find it difficult to justify employing or building a direct sales force for data products, since selling data may still be far from their core capabilities. For this reason, indirect channel partners, such as a data marketplace, are better suited to build awareness, extend the sales reach, expedite transactions and quickly scale up. The marketplace also lets vendors test new products and offerings quickly, without a large institutional investment.
The primary trade-offs with any indirect partner is that the partner typically takes a modest commission to fund infrastructure, marketing, business development and transaction processing. At the same time, customers typically only have a direct relationship with the partner.
Placing data products on a marketplace also allows vendors to reach new geographies and buyers in other industries who might not have previously expressed a need for the data. For example, weve seen an interesting case where produce price data in South America was used to make automotive production decisions in Western Europe. With a data marketplace, decision makers worldwide have access to robust data sets necessary for enhanced causation modelling and assigning prescriptive actions.
Step 5 Delivering On The Service
Like any commercial agreement, data product sales are bound by terms and conditions between a data vendor and data buyer. These T&Cs include service level agreements for uptime and response time, authorizations for use and resale, and technical support for bug fixing and API integration. If a vendor is new to the data economy, these T&Cs may be a daunting task and require significant institutional change.
Buyers will have high expectations, and the need to set up a proper data product delivery mechanism is key. Yet another reason to consider a channel partner with which to offload the service and delivery responsibility.
Recognizing these steps inspired us to create DataStreamX. We believe that humanity has a vast wave of progress on the horizon. Before DataStreamX, selling data products was not an easy activity to scale. However, with the right partner’s reliable and trusted system, data owners can now offer their data products to smart, clever and eager organizations worldwide.
