Monday, December 01, 2014
Big Data Analytics Business Model: Revenue Sharing Strategy and Productized Services
Healthcare companies are living in a threatening environment: costs are increasing at higher rates than inflation rates, and new technologies and therapies are expensive and take too long to bring to market. This situation has moved the marketplace to adopt analytics as a way to prioritize population health strategies, chronic disease therapies, and claims processing efficiencies. A strategy to reduce costs and to accelerate these analytic solutions is to embrace a revenue-sharing strategy with IT and analytics vendors.
Some healthcare business analytics vendors will have a problem adapting to this strategy. Steve Jobs said it best in 1995, when referring to IBM: “Companies get confused. When they start getting bigger they want to replicate their initial success. And a lot of them think, well, somehow there’s some magic in the process of how that success was created. So they start to try to institutionalize process across the company and before very long people get confused that the process is the content. I mean that’s ultimately the downfall of IBM. IBM has the best process people in the world; they just forgot about the content.”
The content in healthcare analytics is a measurable business lift that is timely, reduces costs, and improves clinical outcomes. This measurable business lift is different from the traditional IT ROI. Analytic vendors should be able to stand by their products or solutions, not just sell the software and provide training. This is the reason that fifty percent of all analytic solutions are failing; they do not take into consideration the measurable business results of their solution.
In order for a revenue-sharing analytics solution to work, the analytics vendor and the healthcare companies need to be financially invested in the same business metrics. Large companies like IBM, Microsoft, Oracle, or GE need to change their strategy in the area of big data analytics. This shift from technology and analytics, to solving a business problem with measurable results in a timely manner, is going to be the keystone of success in the 21st century. In the Internet industry we see the revenue sharing model with Google AdSense, and, in my opinion, it would work in the healthcare industry, as well (with some industry-specific modifications). Also, the Banking and Financial industry have seen the benefits of a revenue-sharing model in mutual funds, which resulted in the innovation know as high frequency trading.
We are at an exciting time when technology, big data, and advanced analytics are merging to produce and improve measurable business insights. Let me suggest a business model (which includes technology, analytics and domain expertise) to help streamline solving business problems: a combination of industry-specific predictive models products and services to improve the implementation and accuracy of those models is what I refer to as a Productized Service business model.
I have been in the analytics industry for over twenty years, and in today’s environment we require not only new approaches to big data analytics, but a different business model strategy as well.
First, companies are designing analytics platforms, but there seems to be some disconnect between the analytics platform and the business problems that analytics should help solve. The main reason is that the IT team is used to solve technology problems and not specific business problems. This is a “translation” issue. The business and technology teams have different language and motivation. IT thinks that this bridge is overcome with the position of business analyst, but in reality the BA is mostly collecting requirements but not explaining the business need.
Second, IT and data scientists speak a different language. For example, a predictive analytics segment is a part of the population that has statistically similar characteristics, while a segment (think database indexing/partition) for IT is an optimized procedure to allow queries to run faster.
Third, data scientists and business owners also speak a different language. I am sure that we all have had the experience of putting both groups in the same room and the business owners want to talk about specific issues, while data scientists want to talk about distribution, correlation, and predictive models.
I have found that a way to overcome these translation issues is to have a productized business model. An immediate is to have an organizational alignment from the top down that brings these groups together to start overcoming these language issues at all levels within a company. In other words, the strategy and alignment start from the executive management and moves down throughout the organization. Metrics are extremely important in this step. Companies must include in their metrics the number of revenue sharing programs and how are they performing.
The second step is to create learning opportunities, knowledge transfer, for each group to summarize both their basic skills and motivation. This requires planning and the ability to deal with new information not previously known.
Third, companies need to have analytics products that deal with both domain specific issues and operational efficiency issues. For example, in the healthcare industry there are operational issues (i.e., subrogation, correct payment of claims, and utilization of nurses for care management and disease management programs), as well as clinical issues (i.e., coordination of care, quality of care, treatment, and diagnosis). This is crucial since it allows for companies to talk about specific business issues that can be explained and modified as needed.
Fourth, companies need to develop specific business and technical (IT and statistical) metrics up front to diagnose whether an analytic product or solution is working or not. This process of designing metrics is a good time to start overcoming language barriers among different groups.
Fifth, companies need to develop libraries that allow business, IT, and analytics terms to be captured and disseminated throughout the company. These dictionaries allow people to have a standard way to reference each other’s languages. It is important that these dictionaries have examples using common language for examples and not just technical language. In healthcare, disease specific population health management segmentations should be the same for marketing, claims processing, and CM/DM programs.
Sixth, the productized service needs to be cost efficient. One of the best ways to do this is a revenue sharing arrangement between all the stakeholders that they will share in the costs and revenues arrived through analytics. This is going to be challenging in today’s financial models for some companies, but something that can be experimented and refine as needed.
Seventh, already developed products need to be implemented quickly. Companies want an analytics product to be operationalized and implemented within three months, if possible, to start reaping its benefits. This will involve compromises by all the different groups. Business, IT, or data scientist should not expect the first product that is implemented to be perfect. Nevertheless, they should have high expectations that the analytics product have a measurable impact in their operations.
Lastly, the maintenance, refreshing, and update of any analytic product must be established up front as part of the project plan. These services are expensive and require planning and budgeting, hence requiring time to prepare. Depending on the company fiscal year, a one-year notice may be required.
In conclusion, healthcare companies are going to need to influence these strategic changes for IT and analytics vendors by taking the lead in the way that they do business. Ultimately, revenue sharing strategy and productized services will force the corporate changes necessary to quickly adapt in this challenging environment.
- advanced analytics (2)
- analytics (5)
- analytics tools (2)
- big data (3)
- buisness analytics (4)
- business analytics (4)
- business plan (1)
- center of excellence (1)
- classification (1)
- companies (2)
- data mining (3)
- framework (3)
- game theory (1)
- innovation (3)
- leverage (1)
- marketing analytics (1)
- predictive modeling (4)
- prioritization (1)
- priority (1)
- projects (1)
- recession (1)
- robotic surgery (1)
- segmentation (1)
- social media (2)
- trade promotion (1)
- trends (2)
- web analytics (2)
- what if scenarios (2)
- ► 2008 (103)
- ► 2007 (55)