The Oakland experiment centered on reassessing different facets of baseball, including management, gameplay, player suitability, and rationale. The implementation of computer technology greatly minimized the costs associated with collecting and analyzing extensive baseball data. Consequently, players’ salaries rose, further enhancing the benefits gained from this knowledge.
‘If we’re going to pay these guys $150,000 a year [1977] to do this…we should at least know how good they are’… If this sounded compelling when baseball players were getting paid $150,000 a year, it sounded one hundred times more so when they were paid $15 million a year…” — Michael Lewis, Moneyball1 Graham Gardner, CEO of Kyruus, scanned the room and observed his executive team consisting of the heads of the technology, analytics, product, marketing, and legal departments of the company.
He and co-founder Julie Yoo were strong advocates of hiring exceptional individuals with diverse backgrounds and empowering them to collectively guide the company. This was precisely the circumstance in which they needed such a team. In the spring of 2012, the company was preparing for their upcoming Board of Directors meeting and the team had important choices to make. The team unanimously agreed that Kyruus must determine its ‘killer app’ – the key product or feature for the company. However, there was a debate on how to effectively allocate the company’s resources while identifying and pursuing its most prospective customers and markets.
Gardner and Yoo established Kyruus in August, 2010, with the aim of enhancing collaboration between organizations and their physicians’ networks. (Refer to exhibit 1 for management biographies.) Kyruus offered data and analytics services to various clients, including hospitals, payers, and pharmaceutical companies, enabling them to comprehend and oversee the risk and value associated with their networks. Over time, the company accumulated data on the 800,000 physicians employed in the United States.
The data contained details about the education, licensing, publication and patent history, industry interactions, and prescription and medical device usage patterns of each physician. ________________________________________________________________________________________________________________ This case was prepared by Senior Lecturer Robert F. Higgins, Research Associate Penrose O’Donnell, and Mehul Bhatt (MBA 2013). HBS cases are intended for class discussion purposes only. Kyruus’ board of directors is chaired by Robert F. Higgins.
Cases do not serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. © 2012 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545-7685, write Harvard Business School Publishing, Boston, MA 02163, or visit www.hbsp.harvard.edu/educators. This publication may not be digitized, photocopied, or otherwise reproduced, posted, or transmitted without Harvard Business School’s permission.
This document is authorized for use only by Farhan Habib in Information Systems Development – Spring 2014 taught by Anna Sidorova from January 2014 to July 2014. For the exclusive use of F. Habib 813-060
Kyruus: Big Data’s Search for the Killer App
Kyruus had already developed and sold its first product, a regulatory compliance software package, to certain leading academic medical centers.
Kyruus was currently in negotiations to sell the next version of the product to three other major hospitals.
The product assisted academic institutions in adhering to regulations concerning the interactions between physicians and the industry.
Preliminary market research showed that there were several buyers interested in physician data. Hospital CFOs stated that Kyruus could assist in improving revenue and cost management by addressing patient “leakage” from their system. Executives from the pharmaceutical industry showed interest in Kyruus developing more focused marketing strategies. Various stakeholders in the healthcare industry were enthusiastic about Kyruus’ potential to enhance their businesses. However, one of the challenges presently confronting Kyruus is managing the high level of excitement from potential customers.
The data set powered various applications, leading the team to work tirelessly on creating multiple products for different healthcare industries. Nevertheless, merging the product, engineering, and sales teams across those numerous products was proving challenging. While Gardner desired to pursue multiple industries, doing so would add to Kyruus’ expenses and potentially result in disappointing clients. As Gardner heard his team discussing their strategy, he identified three potential choices for Kyruus. The first option was to focus on the regulatory compliance market.
The team decided to concentrate on academic medical centers as their client base while refining their market positioning and product features. They were making progress in this market segment, with the American Hospital Association (AHA) endorsing Kyruus’ solution for meeting compliance requirements (see exhibit 3). Although pursuing the compliance market was the least risky choice, some internal executives doubted the size of this market opportunity and its ultimate potential.
The company considered two options for growth. One option was to expand its presence in hospitals by developing tools for physician network development and patient referral operations. Initial discussions with hospital CEOs and CFOs indicated that optimizing patient referrals was a major concern for hospitals. It was believed that reducing patient leakage to other institutions would have a significant impact on a hospital’s financial performance. The market potential for this product was expected to be much larger than the compliance market.
The team had a lot to learn to develop and sell the referral and network products. Finally, the company could focus on selling to the life sciences industry, such as pharmaceutical, medical device, and diagnostics companies. Some team members believed that these organizations had bigger budgets compared to hospitals and shorter sales cycles, making them ideal clients. Kyruus had already created a few successful products that industry sales forces liked, and many on the Kyruus team saw this as a potentially lucrative market.
Underlying all these options, there was a fundamental question about whether, when, and how to involve physicians themselves. During the initial year of developing products and entering the market, Gardner and Yoo had decided to assist hospital administrators in managing their physician staff. However, Gardner believed in the significance of “feedback loops” and believed that improving physician performance would eventually necessitate directly involving physicians through Kyruus’ technology platform. He recognized that they only had one opportunity with doctors, as any mistakes in the data could immediately damage Kyruus’ credibility.
While listening to the progress update from the data scientist, he questioned if it would be better to prioritize applications for physician engagement along with the other options. This document is only authorized for use by Farhan Habib in Information Systems Development – Spring 2014 taught by Anna Sidorova from January 2014 to July 2014. Exclusive use of F. Habib Kyruus: Big Data’s Search for the Killer App 813-060 Some team members were clearly in favor of the first option, pointing out that Kyruus had recently gained significant momentum in this area.
The marketing team was creating a comprehensive series of campaigns focused on the recent AHA endorsement. At the same time, the data team was prioritizing the collection and management of compliance-related data, which is highly valuable and ready to be utilized in the database. One team member expressed the importance of excelling in our initial product and target market, emphasizing that we have an entirely unique and game-changing product for this expanding market. Therefore, achieving success in this area should be our top priority.
“As was the style of the executive team, another member pushed back in order to drive the conversation. The hospital compliance market is a solid $50 million market and I agree that Kyruus is well on its way to providing an excellent product in this area. But, the real goal for our company is to identify the $500 million product. I worry that the compliance office is a difficult place to start if you want to expand into other parts of the hospital. As hard as it may be for the Kyruus team, I think we really have to prioritize our understanding of the patient referral market and aim our product team in that direction.”
In support of the second option, one team member brought up previous research on the referral optimization opportunity. They mentioned that redirecting even just 1% of patient flow, which would otherwise be lost to outside referrals, could result in $10 million in revenues for a $1 billion healthcare system. These revenues could potentially contribute to the bottom line due to the fixed cost nature of hospitals. Considering the narrow margins that many of these hospitals operate on, referral optimization is essential for their success.
Some team members believed that a referral optimization application could be a natural extension of the company’s data asset. They thought it would be beneficial for the sales team to approach potential hospital clients from different angles, although the audience and product might need to be adjusted. Another group suggested an even more aggressive approach, targeting both the life sciences industry and hospitals. After all, Gardner and Yoo had designed the company to quickly iterate with multiple clients and find the most impactful products.
Conversations with a medical device company led to the development of the concept of referral management as a product. Initially, an iPad application platform for referral management was created specifically for a medical device sales team. By working with various customers, the team has been able to enhance their existing products and generate innovative ideas. “In addition to the revenue generated from the industry, our collaboration with different healthcare stakeholders is also providing us with valuable knowledge.”
We have made great strides and attracted numerous investors who are keen on supporting our expansion. Therefore, it would be wise to intensify our endeavors and enlarge our team in order to meet the soaring demand for our products,” Gardner observed as he checked the time. Despite having a productive discussion, he and his team had to reach a decision to present to the Board. He often likened this scenario to “moneyball,” recalling how baseball had embraced statistics and analysis as the sport progressed a generation ago.
During a time when healthcare reform required hospitals, payers, and life science companies to have a deeper understanding, Gardner recognized the importance of Big Data in advancing technology. However, there was a question about whether they should prioritize one primary application or if the true value was in the underlying platform that supported multiple products. Additionally, what about doctors? It could be beneficial to encourage physicians to fully utilize the data.3 This document is authorized for use only by Farhan Habib in Information Systems Development – Spring 2014 taught by Anna Sidorova from January 2014 to July 2014.
For the exclusive use of F. Habib 813-060 Kyruus: Big Data’s Search for the Killer App
The Founders Graham Gardner was born in Boston to Scottish parents who immigrated to the United States for higher education. After completing their studies, Gardner and his family moved to Scotland where they resided for two years. Subsequently, Gardner’s father secured a job with a pharmaceutical company in France, leading them to reside there for six years before returning to the United States. As the oldest of three sons, Graham grew up surrounded by family friends who were physicians and developed a desire to become a doctor from an early age.
Gardner realized the extensive influence of the pharmaceutical industry on healthcare during a dinner with his father. This realization expanded his understanding of medicine’s scope and the significant impact doctors can have in multiple ways. Motivated by this epiphany, Gardner joined Brown University’s BA/MD program, pursuing dual degrees in History and Biology. To gain a comprehensive comprehension of physicians’ diverse roles, Gardner actively sought out various experiences. These experiences included participating in basic science research, working in the marketing division of a pharmaceutical company, and taking on consulting projects while attending medical school during his summers.
Gardner’s passion for medical education led him to fully immerse himself in the clinical setting, with a strong desire to acquire extensive knowledge. He chose to pursue a residency program in internal medicine at a prestigious hospital affiliated with Harvard University in Boston. During this time, his fascination with cardiology grew, igniting a deep interest in the field. The training he received had a significant impact on Gardner, expanding his understanding of different medical specialties and easing any apprehensions he may have had towards unfamiliar areas.
As an internal medicine resident, I used to have concerns about the kidney’s role in overall health. However, my transition to becoming a cardiologist allowed me to solely focus on my expertise in cardiology. This change had a significant impact on me and made me comfortable with not having all the answers. Even now, as a CEO, I maintain transparency with my team and Directors by acknowledging my strengths and areas where I need support.
In addition, I have realized the importance of recognizing and utilizing individuals’ strengths in the medical field. Understanding this concept empowers us to provide high-quality and cost-effective care for our patients.
Matching physician styles to specific patients and conditions can potentially decrease healthcare spending by trillions of dollars, enabling physicians to excel in their field. Gardner realized that his true passion may be at the intersection of business and medicine as his expertise in medicine reached a plateau. Meanwhile, as my training concluded, my parents moved to Boston to assist in establishing a new company with our long-time family friend, Bob Langer. It was during this period that I became acquainted with venture capital and entrepreneurship.
After realizing that there were professionals being paid to start and develop new businesses and innovations, I decided to attend business school. My goal was to increase my understanding of the industry and explore potential opportunities for myself. During my time at Harvard Business School, I had the chance to intern at both Boston Scientific and Highland Capital Partners. Luckily, after completing my degree, I received an invitation to join the Highland team. Originally, I intended to pursue a traditional venture capitalist role.
However, during my first year at Highland, I realized the incredible potential it had as a platform for forming companies. It provided opportunities to connect with amazing individuals and cultivate ideas that can shape the future. Highland gives you all the necessary resources to create the next successful company.
This document is authorized for use only by Farhan Habib in Information Systems Development – Spring 2014 taught by Anna Sidorova from January 2014 to July 2014. For the exclusive use of F. Habib Kyruus: Big Data’s Search for the Killer App 813-060
Gardner played a key role in the establishment of Generation Health, a company aiming to become a genetic testing benefit management company, during his time at Highland. Initially serving as a board member, he later assumed the position of Chief Medical Officer (CMO) during the company’s inaugural year. His primary responsibilities in this role included overseeing the development of the clinical team and the initial product. It was during this period that Gardner recommenced his relationship with Julie Yoo, who had been a teaching assistant in one of his business school classes. Due to Yoo’s expertise in big data and her previous work in medical genetics, Gardner extended an invitation for her to join Generation Health as the head of clinical informatics.
The two individuals worked closely together for the following year during the acquisition by CVS Caremark. Julie Yoo had parents who were both professors – her father specializing in physics and materials science engineering, and her mother teaching English as a Second Language to students wanting to study in the United States. Yoo, herself, had aspirations of pursuing a career in medicine.
When she was admitted to MIT, Yoo initially enrolled as a Biology major. However, being in school during the dot-com boom of the late 90’s, she witnessed the rising enthusiasm for technology and software companies. Intrigued by this, Yoo decided to give Computer Science a try. She found that it aligned well with her passion for problem solving and subsequently changed her major to Computer Science, while also satisfying her pre-medicine requirements. Throughout her time in college, Yoo engaged in consulting projects with local start-ups to gain practical knowledge in application development and information architecture.
In the summer after her junior year, she interned at Endeca Technologies, a venture-backed software company specializing in datadriven search and business intelligence for enterprises. Thoroughly enjoying her experience there, she decided to join them full-time after graduation in a software engineer role, becoming the 37th employee at the firm. Over the course of her tenure at the company, Yoo was exposed to various industries, but her foray into the field of healthcare was the one that made the strongest impression.
According to Yoo, during their time at Endeca as a sales engineer, they gained experience in a range of industries including retail, financial services, manufacturing, and healthcare. Out of all these sectors, it was healthcare where enterprise IT solutions that had proven successful in other industries were not being easily embraced or effectively implemented. This presented an interesting challenge: what unique factors in healthcare and life sciences hindered the efficient utilization of data and technology?
This question prompted me to delve deeper into this field. Consequently, after spending 6 years at Endeca, I chose to enhance my knowledge in the healthcare industry by enrolling in the Biomedical Enterprise Program. This program, a collaboration between MIT and Harvard Medical School, provided me with a chance to explore the integration of fundamental science, clinical medicine, and entrepreneurship. During my time in this program, I focused my thesis research on genomics and personalized medicine. This field presented numerous possibilities for developing big data solutions applicable to both research and clinical use.
During this time, I resumed my practice of working with local startups in the healthcare informatics and analytics field. I discovered a great opportunity to effectively utilize data for the benefit of physicians and organizations working with them. After completing her graduate studies, Yoo was recruited by Gardner to join Generation Health (GH). GH required someone with expertise in genomics and clinical applications of genetic testing to assist payors in establishing reimbursement policies for genetic tests. Yoo was an ideal candidate for the role.
This document is only meant to be used by Farhan Habib for the Information Systems Development – Spring 2014 course taught by Anna Sidorova from January 2014 to July 2014. It is exclusively for the use of F. Habib 813-060.
In the winter of 2009, Generation Health was acquired by CVS Caremark and Highland Capital saw a 6x return on its investment in just 13 months. Shortly after, Gardner resigned from his role as CMO. As the field of genetic testing benefits progressed, a significant portion of its business was focused on oncology. Graham believed that an oncologist would be best suited for the CMO position.
In early 2010, Graham rejoined Highland Capital as a venture capitalist. Yoo departed from Generation Health in July of the same year. She desired the flexibility and entrepreneurial atmosphere offered by small organizations and was keen to venture into new prospects within the expanding Health IT and Analytics sector. During that summer, Gardner and Yoo convened to exchange information, wherein Yoo reminisced: We concurred that numerous healthcare IT companies were solely focused on establishing the infrastructure for data accumulation, resulting in a limited presence in the analytics field. The significance of interpreting the data remained largely unexplored.
In August 2010, Gardner and Yoo established Kyruus with the purpose of improving performance in the healthcare industry through the utilization of data. Despite being a significant producer of data, the healthcare industry has struggled to effectively utilize it. Kyruus aimed to fill this gap by becoming the primary provider of physician data and physician engagement tools, helping organizations optimize their interactions with physicians. At the time, other healthcare data companies were examining specific aspects of physician behavior, such as IMS’ physician prescribing data, but none were consolidating all physician interactions in a single platform.
Yoo described the perfect data set: Physicians have various roles and responsibilities such as patient care, research, clinical trials, publishing studies, consulting, education, and training. There is a wealth of data that captures these activities, but it is not easily searchable in one centralized location. Kyruus’s initial goal is to provide basic access and easy navigation of this information for various stakeholders in the healthcare system. However, before creating any products, they sought to gain insights into the market for physician information.
Gardner held meetings with individuals from medical device companies and hospitals to identify their requirements. It was discovered that there was a significant demand and a strong willingness-to-pay for information concerning physicians. Gardner explained that his previous experience in the business world taught him that companies who interacted with clients in the early stages of product development had a higher likelihood of success in creating market-desired products. Consequently, he reached out to anyone who would lend an ear to his ideas. Following these discussions, Julie would create a prototype according to the company’s specific requirements.
Having discussions and being able to respond promptly to customer needs was crucial in obtaining our first pilots. Customers were eager to share their requirements and were willing to pay a high price. These conversations informed various applications for physician data, but the initial idea for Kyruus’ product emerged when Gardner and Yoo participated in a pharmaceutical regulatory compliance conference in October, 2010. It was during this conference that Gardner and Yoo recognized the significant role physician interactions played in the healthcare ecosystem.
Kyruus has the potential to create a comprehensive data platform for organizations to efficiently manage risk and optimize value within their relationships. Gardner explains that this document is only authorized for use by Farhan Habib during the Information Systems Development course taught by Anna Sidorova from January 2014 to July 2014. This exclusive use is granted to F. Habib. Kyruus: Big Data’s Search for the Killer App 813-060 reports that executives from major pharmaceutical and medical device companies appeared visibly shaken after their companies faced severe penalties for engaging in illegal interactions with physicians.
Despite recognizing the importance of physician-industry interactions in driving healthcare innovation, these companies admitted that they lacked the necessary data and tools to engage with physicians in a transparent and compliant way. While we observed numerous law firms offering their services, there were no companies offering data to support compliant relationship management for these organizations. This realization prompted Julie and me to understand the potential of our revolutionary data platform, which could enhance physician interactions with companies, hospitals, and fellow professionals. The context of our discovery was compliance, healthcare IT, and the utilization of ‘Big Data’.
Physician-Industry Interaction: The Physician Payment Sunshine Act (PPSA) was a crucial element of the 2010 healthcare reform law, the Patient Protection and Affordable Care Act. The PPSA requires the reporting of all payments exceeding $10 to physicians or hospitals from pharmaceutical, medical device, or medical supply companies to the government. Failure to report such payments can result in penalties as high as $1 million.
The motivation behind implementing the PPSA was the concern over the lack of transparency surrounding the large sums of money that physicians received annually from the pharmaceutical industry.
Despite some of the payments being used to fund research, there were concerns about potential conflicts of interest arising from multiple interactions and financial relationships between these entities in the clinical setting (See exhibit 4). Unfortunately, academic medical centers had to face the consequence of a public relations crisis as a result of this regulation. Although many of these institutions had strong conflict of interest policies in place, they mostly depended on physicians to disclose their industry relationships.
The hospital compliance officers faced inquiries from journalists, who gathered data from pharmaceutical companies and released stories about physicians receiving substantial payments from life science companies. Often, the compliance officers at the physician’s academic institution were unaware of these interactions because they were not disclosed or because they happened after the last disclosure cycle. Insufficient reporting of potential conflicts of interest could jeopardize federally-sponsored grant awards to institutions.
At large institutions, these awards totaled hundreds of millions of dollars per year. Gathering the information published by pharmaceutical companies proved to be challenging. The companies were uploading the data in various formats, using different names to refer to the same physician (such as John Reynolds vs. Jonathan Reynolds, MD), and employing diverse terms to describe their relationships with physicians (Refer to exhibit 5). Several academic institutions had started employing analysts to access each company’s website and sift through hundreds of thousands of transactions to find their staff members’ names.
Neither a reliable nor scalable solution, Kyruus’s Big Data approach was perfect for dynamically aggregating and “disambiguating” physician identities, contextualizing information, and enabling institutions to benchmark themselves and their physicians against peer institutions nationwide. Gardner and Yoo recognized an opportunity to aid institutions with data management solutions for regulatory compliance. Studies showed that physicians frequently did not sufficiently disclose their industry relationships, leaving them vulnerable to public relations and regulatory risks.6, 7
This document is authorized for use only by Farhan Habib in Information Systems Development – Spring 2014 taught by Anna Sidorova from January 2014 to July 2014. For the exclusive use of F. Habib 813-060 Kyruus: Big Data’s Search for the Killer App They saw an opportunity to collect data proactively and help physicians avoid reporting errors. At the same time, they could develop individual portfolios of professional activities that would contextualize industry interactions. Gardner and Yoo hoped that these comprehensive narratives would protect the reputations of physicians and institutions as they interacted over the long term with industry.
7 “Big Data” is the term used to describe extremely large data sets that cannot be managed, organized, or analyzed using traditional database software tools. 8 These data sets are found everywhere and their volume is rapidly increasing. In 2011, a total of 1.8 zettabytes of digital information was created globally, with a projected increase to 35 zettabytes by 2020. a,9 Our ability to store this data has doubled every three years since 1980. 10 All industries have been collecting massive amounts of data. For example, Facebook has 30 billion new pieces of content on its website each month. 11 Additionally, Walmart processes 1 million transactions per hour, which are stored in a database containing 2.
Big data refers to enormous amounts of information, which is equivalent to over ten times the content found in the books of the entire US Library of Congress. One common framework for describing these datasets is known as the 3V’s: volume (the quantity of data generated), velocity (the speed at which data is produced), and variety (the different formats from which the data originates, including databases, blogs, and websites). The analysis of these vast datasets aims to provide improved insights into potential customers, facilitate the development of innovative products and services, and assist in decision-making processes to mitigate risks.
According to Gardner and Yoo, they believed that healthcare was a suitable field to apply the advantages of big data. Gardner provided an explanation, stating that Big Data can be defined as when data itself becomes a part of the problem that needs to be solved. Instead of needing more data, people needed to integrate the data they already had, examine new intersections of information, and use the insights gained to achieve better outcomes. Information regarding physicians, who are key individuals in these institutions, was typically scattered among numerous separate systems such as HR departments, credentialing systems, electronic health records, and supply chain databases.
The data was stored in various formats, making it impossible to join using a single key. Despite this challenge, critical insights could be obtained by exploring the intersections of these data, which organizations could utilize to improve and optimize their operations in the Healthcare IT Market. There are four main categories of healthcare data: pharmaceutical R&D data, claims and billing data, clinical data, and patient behavior data. By integrating this diverse range of data, significant value could be generated for all participants in the healthcare value chain.
16 Recently, the government has been promoting the integration of information technology into the healthcare industry. The American Recovery and Reinvestment Act (ARRA) of 2009 aimed to allocate $20 billion in stimulus funding over 5 years to incentivize the adoption of electronic health records by physicians and hospitals.
17 By 2010, the healthcare IT (HCIT) market was valued at around $25 billion. Among the different sectors within healthcare, the HCIT market was projected to experience the highest annual growth rate of over 10% for the next five years.
The pharmaceutical, medical device, and diagnostics markets were predicted to experience single digit growth. In addition, more VC funding was being directed towards the Health IT sector due to the rising importance of big data. In 2009, out of a total US venture funding of $21.4 billion, $7.73 billion was allocated to the healthcare sector. This document is authorized for use only by Farhan Habib in Information Systems Development – Spring 2014 taught by Anna Sidorova from January 2014 to July 2014. For the exclusive use of F. Habib Kyruus: Big Data’s Search for the Killer App 813-060.
Investment in healthc
