Data Products: 9 Best Practices To Minimize Risk
Data generation, collection, and analysis are making their way into more types of products and services. The trend is creating new opportunities for innovation, some of which are so impactful, they're causing some companies to revisit their business models. The path to success isn't always obvious, however, so here are a few best practices to keep in mind.
![](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/blt00ce722489ddf341/64cb4075ae1f5e4d098f3577/binary-world_000077899919_Small.jpg?width=700&auto=webp&quality=80&disable=upscale)
Data is finding its way into just about every type of modern product and service. As a result, some companies are necessarily rethinking their business models, product strategies, customer engagement strategies, and supply chain strategies. Meanwhile, entrepreneurs and intrapreneurs are discovering entirely new solutions to age-old problems.
"Our traditional business model, the way we provide products and services, is being disrupted because people -- especially Millennials -- do not look at a big book of codes," said Nataniel Lin, analytics and strategy lead at the National Fire Protection Association (NFPA), in an interview. "We're in the process of becoming a 120-year-old startup. Essentially, we're leveraging all the data that's available out there and aggregating data to create unique value and solutions that up until today were not possible."
In NFPA's case, data is flowing in from connected IoT systems in homes and commercial buildings, insurance companies, and other sources. Lin is working with 26 different property and casualty insurance companies with the goal of anonymizing and aggregating data in a way that benefits all of the companies without exposing them to privacy or security risks. That way, the companies can have a more objective view of revenue, profitability, and risks than would be possible using only their own data.
[Looking to make your analytics agile? Read Agile Analytics: 11 Ways to Get There.]
In the agriculture industry, farmers, crop consultants, commodity traders, and food security analysts are consuming agriculture and weather data to make more informed decisions. Agronomic data provider and data management company aWhere is working with agriculture software companies, commercial agricultural research groups, and information providers, including mobile network operators, to change how agricultural decisions are made. Its data comes from a number of sources, including satellite sensors, ground radar stations, weather stations, and other sensor-based devices.
"[Originally], our founders were working on agricultural research projects in Africa where the challenge was a lack of data. We developed techniques and methodologies to model weather across the landscape using a variety of other data sources to model virtual weather stations everywhere and give a complete picture of the climatology," said Stewart Collis, cofounder and CEO of aWhere, in an interview.
Those initial efforts enabled the aWhere founders to assess agricultural risks and understand the impact of growing different crop varieties. Farmers now use the same data to more effectively manage the impact of weather variability and assess the impact of weather events on global agricultural production.
While it's becoming more common to offer data products as part of a company's overall business strategy, the attractiveness of potential opportunities can sometimes blind founders and executives to some of the common risks and pitfalls that cause their companies to stumble. To minimize their potential effect, consider these best practices.
Organizations tend to underestimate the amount of effort, patience, and open-mindedness that may be required to create and maintain successful data products. Similarly, it isn't always apparent what degree of impact a data product may have on the business.
For example, a consulting company morphed into an application company because an investor perceived an opportunity the founders had not considered. At the time, the consulting company was working with a hospital to improve the spend management associated with implantable devices used in knee, hip, spine, and cardiac surgeries. To do its job more effectively, the consultancy developed a system that was used for internal purposes. Just as the company was closing a round of seed funding to fuel its growth, a new investor noted that the funding was adequate to grow the consulting business but inadequate to build a SaaS product. As a result, the founders decided to pursue the new direction. They ultimately launched Kermit, a spend management analytics platform that's built on the Mendix application platform. The full-featured version of Kermit allows surgeons to see what the implants will cost, and allows hospital executives to better understand profitability. It also automates invoice auditing and contract compliance.
"When a surgery is performed, the sales rep for the implant company uses a piece of paper to track the items that were implanted," said Richard Palarea, CEO of Kermit, in an interview. "What started out as an internal need to calculate savings [by] negotiating better price points turned into an opportunity to create an electronic, intelligent version of the paper bill sheet being used -- one that could capture overt and covert billing errors before they are submitted to the hospital for payment."
Some companies realize the data they are amassing has value they hadn't considered previously. That data may be productized and monetized directly or indirectly, which generally means that the data is sold, made available via a SaaS subscription, made available via an API, or simply provided as a value-add to differentiate the company and its product or service. However, what may appear to be an attractive market opportunity may prove to be a costly or unprofitable endeavor if the idea is not continuously tested, validated, and fine-tuned to meet the actual needs of customers.
"Be clear about the value you can provide to your market, even if that vision is small to begin with," said Richard Palarea, CEO of spend management platform provider Kermit. "As you grow in customers and users, they will show you the next steps. You need to show them you care about improving the quality of their daily experience and they can trust you."
Companies that are in the business of selling data products may expand their offerings with the goal of entering new markets. As they enter these new markets, they may find that the strategies and tactics that made them successful in one set of circumstances doesn't work as well when applied as-is to another set of circumstances. Meanwhile, some manufacturers building intelligence into their products are discovering that they have to think differently about their core competencies.
"Very few companies are organized in a way that can leverage the changes that a smart product requires," said Nathaniel Lin, analytics and strategy lead at the National Fire Protection Association. "Smart products tend to cut across data silos. They have to be promoted differently, marketed differently, and delivered differently."
Individuals and companies creating data products can sometimes get over-enamored with technology or data to the point of becoming distracted by it. As always, the hottest technologies aren't appropriate for every use case -- an important fact that is often forgotten amid the industry excitement and hype. Similarly, vast amounts of data can inspire analytics, modeling, and trending efforts that are academically interesting but have little value, practically speaking.
"You start with a business question, then you do a data audit based on the analytics you do, and you test it to see if it created a business opportunity or solved a problem," said Nathaniel Lin, analytics and strategy lead at the National Fire Protection Association. "It's like peeling an onion. Once you get through one layer, you discover other layers -- [sometimes] opportunities you would not know existed unless you asked the right questions and used the right tools."
Business as usual is a stale proposition when it is obvious that the business model could be improved or a new opportunity could be exploited. For example, when mobile advertising solution-provider UberMedia was an app company building social media apps, it realized it had a wealth of previously untapped consumer data that was being generated by mobile devices across its platform.
"We started to apply sophisticated data science and machine learning to unlock consumer audience preferences, interests, and intent cues," said Michael Hayes, CMO and chief revenue officer at UberMedia, in an interview. "[It] changed the course of our business in terms of revenue creation via advertising dollars."
Investment decisions hinge on information. Fortunes are made and lost because one party has access to information competitors lack, faster access to information than their competitors, or both. Knowing that, text-analysis startup Prattle developed an algorithm that analyzes and quantifies central bank communications to produce tradable data from market chatter. Its clients use the data to anticipate market movements and make investment decisions faster than their competitors can.
"[Many] capital markets firms [including] hedge funds, mutual funds, or investment banks still rely on very qualitative analysis of assets," said William MacMillan, cofounder and CTO of Prattle, in an interview. "Our perception is that there are a number of spaces where a technically sophisticated approach to previously unquantified information would create interesting market opportunities."
The financial services industry as a whole, and the capital markets sector specifically, are more diverse than Prattle's founders originally anticipated. As a result, the company now offers several products aimed at different audiences.
The difference between evolution and revolution is a matter of degree. Some data products provide incremental improvements over what exists. Others may completely disrupt entire industries.
"A few years ago, it was impossible to effectively deploy ads to target consumers who had been to a specific retail location and then measure the foot traffic resulting from the ad campaign," said Michael Hayes of UberMedia. "Today's consumers expect companies to understand their preferences and provide relevant personal experiences across every touch point. This seamless omnichannel experience is proving to be incredibly challenging for marketers."
UberMedia synthesizes more than 1 billion real-world location data points per day to provide ad targeting and measurement accurate within three feet of a retailer's physical location. The geolocation accuracy enables marketers to re-engage customers who have been to their stores, target consumers who are shopping at competitors' stores, and gauge the performance of their campaigns based on actual foot traffic to their stores.
More data and more types of data create an opportunity to build entirely new data products. In the healthcare industry, organizations are combining data from hospital networks, anonymized patient medical claims, healthcare provider data, public records, and socioeconomic data to drive predictive analytics for population health outcomes, quality scores, fraud indicators, waste, and abuse, according to Theresa Greco, vice president, life sciences at LexisNexis Risk Solutions, in an interview.
"The types of opportunities that are being presented today require access to real-time data that spans many varied sources including public, professional, individual, corporate, and relationships," said Greco. "The speed of business requires high availability and high accuracy of information to deliver current and comprehensive data in order to make decisions."
More companies are taking advantage of data flowing in from different sources including customers, partners, public sources, and social media networks. Meanwhile, product manufacturers and service providers are monitoring usage and other information to innovate and improve their offerings. Quite often, data that is of value in one part of the supply chain can benefit at least some of the other players.
"Value in healthcare generally is shifting away from pure physical products to the information housed in patient medical records, in clinical development systems, and other supporting systems," said Barry Blake, VP of Research for research institute SCM World, in an interview. "As reimbursement and payment shifts towards improvement in outcomes (such as reduced length of stay, medical adherence, or reduced recidivism), the data providing this evidence becomes extremely valuable to manufacturers as they develop new products or services to lessen the burden on healthcare delivery organizations."
More companies are taking advantage of data flowing in from different sources including customers, partners, public sources, and social media networks. Meanwhile, product manufacturers and service providers are monitoring usage and other information to innovate and improve their offerings. Quite often, data that is of value in one part of the supply chain can benefit at least some of the other players.
"Value in healthcare generally is shifting away from pure physical products to the information housed in patient medical records, in clinical development systems, and other supporting systems," said Barry Blake, VP of Research for research institute SCM World, in an interview. "As reimbursement and payment shifts towards improvement in outcomes (such as reduced length of stay, medical adherence, or reduced recidivism), the data providing this evidence becomes extremely valuable to manufacturers as they develop new products or services to lessen the burden on healthcare delivery organizations."
-
About the Author(s)
You May Also Like