‘Predictive analytics’ has become a buzzword today, and there isn’t an organization that isn’t thinking about it, or hasn’t implemented their definition of it. The key is to not debate whether predictive
Successful business outcomes with a high ROI are being attributed to predictive models that are integrated with business models. Many industries are looking at the possibility of improving their effectiveness in the market by employing predictive analytics. Essentially, they are looking to replace informed choices (which were largely based on historical models) with informed intuition (which can be assisted with predictive models looking into the future). Mature organizations define a universal set of business outcomes, create dashboards with key performance indicators, and track them diligently. There is a shift being observed in progressive organizations which are adopting dashboards that include a key performance forecaster (KPF); these organizations are able to improve their decision-making capability. Several applications such as SAS and SPSS are available; these incorporate complex mathematical/ statistical models to provide insights to decision makers.
Predictive analytics will gain importance in India as organizations have become aggressive in seeking data from beyond the scope of their own organizations to integrate it with internal data.
Surmounting current limitations
Let’s take the example of an organization in the life-sciences sector that seeks to improve its KPF in key business areas: supply chain, finance, performance, of products and quality. For the purposes of this discussion about predictive analytics, let’s assume that it’s a progressive organization which seeks to integrate market intelligence (third-party data) that provides performance benchmarks to achieve the outcomes mentioned below.
The primary purpose is to align data (internal and external) with their current actions to achieve these key outcomes:
- Identifying and developing pipeline of new products to be introduced in the market
- Customer retention, greater market penetration, and improved customer acquisition
- Marketing and pricing strategies (determining discount rates, promotions, etc.)
- Overall sales, revenue, operation profit margin, year-on-year comparison of these measures
- Reducing overall cost (operational, marketing, sourcing, etc.)
- Improved collaboration with suppliers and distributors for operational efficiency
- Reduced spend rates—effective budget planning
- Optimal inventory—effective supply chain processes
Relevance of predictive analytics
The traditional business intelligence approach will address the outcomes listed to a certain degree, and will provide insight to the decision makers. All the outcomes mentioned above would be achieved through developing data warehouses and making decisions that are driven by historical information provided through dashboards and other formats. To surmount the challenge of resolving the complex issues listed below, predictive analytics would need to be adopted.
- Identify the risk of developing certain health conditions such as diabetes, heart disease, and other lifetime illnesses.
- Model market changes for the next five years, based on demographics, innovation, health care trends.
- Identify the source for specific products and proactively resolve issues related to complaint rates, customer satisfaction, and non-conformance, etc.
- Identify exit strategy for old products and introduce new products that are likely to produce higher margins.
Most of these scenarios are solved through predictive analytics by deriving patterns from millions of transactions, that would have been recorded by the organization—sales, supply chain, finance, quality, etc,—to develop key performance forecaster with reliable levels of accuracy.
This will help organizations to command higher premiums and maintain leadership positions. From a data management perspective, this can be achieved by mature organizations which have developed robust data warehouse systems.
This is critical to the success of developing predictive modeling as they have rationalized, normalized sets of data which are stored in a consistent format. There is an increasing awareness and trend to this effect that is being witnessed in the life sciences sector.
As organizations move up the maturity curve of honoring information requests by business users through different mechanisms, including effectively implementing enterprise datawarehouse systems, the cross-relationships between functions or business areas within the organization will provide rich data for predictive analytics to yield results of great value.
In the absence of this, it could still provide benefit if the warehouse is limited to a function. However, there is no benefit if data governance and management are not rigorously followed by the organization.
There is a seismic shift from scarcity to a world of abundance of data perspective. As a result, the ability to extract the necessary data from large databases and make meaningful decisions is extremely challenging and time consuming. Organizations should consider that predictive analytics may increasingly become necessary to sustain its leadership position and to avoid unintentional erosion in brand value, market share, and profitability.
About the author: Chandu Mukkavalli is director at Deloitte Consulting India. Chandu has more than 13 years of IT experience, with a primary focus on information management, business intelligence, and data warehousing. Chandu has developed and led the business intelligence/ data warehousing practice at large consulting organizations.
Disclaimer: This presentation contains general information only and is based on the experiences and research of Deloitte practitioners. Deloitte is not, by means of this presentation, rendering business, financial, investment or other professional advice or services. As used in this document, ‘Deloitte’ means Deloitte Consulting LLP, a subsidiary of Deloitte LLP. Please see www.deloitte.com/us/about for a detailed description of the legal structure of Deloitte LLP and its subsidiaries.