Operational Excellence: Know Thyself with Predictive Modeling

The pharmaceutical industry, with its extraordinary discoveries and rigorous processes has contributed enormously to longer and improved quality of life around the world. And yet it faces significant challenges ahead due to changes in the global marketplace.

“We must act to contain the growth of health care costs to ensure our economic stability; to help American businesses deal with the health care challenge; and to make sure we are getting our money’s worth”
-US Senators Kennedy & Baucus, 4/20/2009

Ageing populations across the developed world are profoundly affecting the healthcare market. They are increasing the demand for drugs while the increasing number of retirees simultaneously creates significant pressure for lower costs. Insurance companies and national health care systems are using their large market share to actively push the use of lower cost generic drugs. Over time, as more drug patents expire, even more of the industry’s profits will be subject to intense pricing pressure. The global recession is straining government, corporate and personal finances and exacerbating these long term pricing trends.

In May, 2008 Wal-Mart expanded its $10 generic drug program to cover 350 generic medications. With its global reach and market presence Wal-Mart can be expected to become a major distribution channel. Wal-Mart is no stranger to operational and supply chain efficiency. They work with their suppliers to drive efficiencies up their value chains.

In this environment, there are opportunities for innovative pharmaceutical firms to gain market share and protect their profitability. Suppliers with shorter lead times, more reliable delivery and competitive operating costs can permanently gain market share in the shift to mega-market channels. Successful companies with best in class operational efficiency, quality and financial metrics will dominate this marketplace transformation.

Studies [1] showing that business process innovations deliver more business ROI than product innovations have interesting strategic implications in the global pharmaceutical marketplace. Companies with a superior order to delivery process can gain market share when there are low barriers to product substitution. Large mega-channel buyers like Wal-Mart demand a combination of low cost, on-time delivery and LEAN inventory management processes from their preferred suppliers. Relentless market volatility means a company’s ability to efficiently and precisely respond to customer demand is vital in a business environment where economic stress and consolidation pressure is a daily reality.

Existing approaches that plan and schedule production capacity to demand forecasts do not deliver best in class results. Innovative customers are combining predictive intelligence with probability models and optimization engines to dramatically improve the operating capability of their order to delivery process, including production planning and inventory management. New predictive control solutions allow a company’s improved capability to be explicitly allocated to meet operational targets for customer service, production efficiency and capital allocation.

Elements of Operational Efficiency
Leading pharmaceutical companies recognize the strategic long term necessity of operational excellence. Stated simply, operational excellence is minimizing waste while maximizing customer value. Value is defined from the perspective of both the external and internal customer in a measurable manner.

According to Womack and Jones, there are seven types of process waste – rework, overproduction, excess inventories, non-value added process steps, excess people movement, excess material transportation, waiting, and non-value added goods of services. The Lean movement has developed important key performance indicators for measuring an operation’s performance and bench marking their performance against the achievements of best-in-class organizations.

Key operational performance metrics for customer value are on time delivery (OTD) service levels and elapsed time from customer demand notification to order fulfillment. From the recent threat of a global swine flu pandemic, one can easily appreciate the advantage of quick response times. Production processes take time, but so do waste activities within those processes.

The two KPIs for internal process efficiency are overall equipment effectiveness (OEE) and total inventory (DSI). OEE [2] is a composite metric accounting for availability, performance and quality. OEE and inventory measure asset efficiency. A mid size pharmaceutical manufacturer estimates that for every billion dollars in pharmaceutical revenue, each percentage point improvement in OEE% is worth $7 million in savings.

In high gross margin businesses, inventory carrying costs are usually insignificant relative to process efficiencies or high product availability. However, high inventory levels (i.e. less than 3 turns per year) is a very strong indicator of inefficient enabling operational processes.

Bench Marking Pharmaceutical Operations
The pharmaceutical industry has been focusing on improving its manufacturing agility through improvements in set up and cycle times and improving its OTD performance. [3]

Informance [4] provides comprehensive manufacturing benchmarking services and they have published OEE comparisons between the pharmaceutical industry and consumer packaged goods (CPG) industries that show a significant performance gap. It is often assumed, due to regulatory and public safety requirements, that cross-industry comparisons of operational effectiveness are not meaningful indicators of performance capabilities. However, within the pharmaceutical industry, there is also a wide gap between best-in-class and the rest of the industry.

Every process has its own capability limits. The key to achieving operational excellence is to determine a pharmaceutical value chain’s physical process capability and identify the gaps between actual and its true capability. Without an understanding of the physical system’s capability, improving performance becomes more complex, especially with unreliable external benchmarks. The inherent complexity of production and fulfillment procedures with inter-related, multivariate, non-linear sub processes means that these systems cannot be adequately analyzed with conventional tools. Digital modeling and simulation is recognized as the approach of choice for these types of complex systems.

The Power of Predictive Analytics and Digital Modeling
Predictive analytics is statistical data analysis to capture relationships used to predict future trends and behaviors. Credit card companies use predictive analytics in sophisticated fraud detection systems and leading retailers and e-tailers use predictive analytics to suggest additional items that may be of interest to the customer based on past behaviors and demographics. In process manufacturing, Pavilion Technologies, a division of Rockwell Automation, uses predictive control models that enable production control systems to adjust more quickly and accurately than with actual measurements. Applying these predictive control techniques significantly improves production yield and quality for many process manufacturing companies.

A digital model of the pharmaceutical operations must be detailed enough to show the interrelationship of customer response times, OTD, production changeovers and inventory levels. That is, the model must be at the customer order – SKU level of detail and operate over continuous time for the period of interest, usually a year. This task is daunting in that the customer orders or forecasts must be converted accurately to production schedules without human intervention. These schedules must then be digitally “executed” to produce product inventory that is stocked and shipped to customers.

SherTrack’s research into demand patterns has enabled the successful application of predictive analytics to the order-to-delivery process. These predictive analytics create an accurate, synthetic demand signal for the operational time horizon of 2 to 3 production cycles that enable a LEAN Pull process. Combining predictive analytics with innovative probabilistic inventory and scheduling solvers completes the requirements for effectively modeling complex manufacturing operations.

Predictive models of the order-to-delivery process make digital simulations over a statistically significant time period such as an entire year a practical option. This predictive modeling approach allows continuous improvement teams to measure both the physical process capability and quantify which sub-processes limit overall performance. Six Sigma tools like Design of Experiments (DOE) can then be used to quantify the performance of order-to-delivery processes and isolate the impact of both physical and enabling processes in realistic operating scenarios.
 
Predictive Manufacturing Case Study: Bayer MaterialScience
Bayer MaterialScience is among the world’s largest polymer companies focused on the manufacture of high-tech polymer materials and the development of innovative solutions to important customer problems. Their compounding business wanted to improve customer response time while not degrading service levels or their cost position. These are common goals of many complex manufacturers.

Connie Conboy, Vice President of Quality and Business Excellence realized that Bayer’s manufacturing facilities with their complex interaction of constraints and operating processes would be very difficult to analyze using traditional LEAN/Six Sigma approaches. Running controlled trials in live operations to measure their response to changing inputs would be prohibitively expensive and pose unacceptable operational and cost risk. A Six Sigma team was chartered to leverage SherTrack’s innovative predictive manufacturing application to help: 
• Reduce customer lead times,
• Reduce production costs,
• Improve Working Capital, and
• Enhance capacity utilization.

Using the predictive models, hypotheses were tested and complex cause and effect relationships between project inputs (I), business and production processes(X) and expected outcomes(Y) were explored. “This methodology provided the Bayer project team with a very robust set of quantified and qualified analyses associated with every manufacturing hypothesis within the Six Sigma set of criteria and our CTQs”, said Bayer’s Senior Six Sigma Lead, Rick Baxendell.

What the Bayer team learned from this novel new DMAIC (define, measure, analyze, improve, control) methodology was that:
• Lead times can be reduced by > 50%,
• Service levels can be improved more than 5%,
• OEE and capacity utilization rates can be raised by 10%,
• Cash flow and working capital can be improved by as much as 20%, and
• Tremendous insight can be gained into production issues that impact performance.

Summary
The pharmaceutical industry faces significant challenges from, and increased complexity in the global marketplace. Process complexity is a critical barrier to improving order-to-delivery performance, and predictive modeling at the operating process level of detail can be a very powerful tool for understanding a pharmaceutical value chains true capability limits. Equipped with a detailed understanding of complete order-to-delivery process capabilities, targeted improvement initiatives can yield significant performance gains that are truly important to the business.

New predictive modeling capabilities are now available that use proven data based decision support methods to quantify process capabilities and identify improvement opportunities critical for business success. In the pharmaceutical arena, process or enabling process changes can be evaluated without affecting your customers or operations and before seeking regulatory approval if required.

Innovations in the enabling processes of complex manufacturers drive significant improvements in responsiveness, reliability and lower manufacturing costs.

References:
[1] http://www.doblin.com/IdeasIndexFlashFS.htm
[2] http://www.industryweek.com/articles/oee_the_heart_of_the_matter_18211.aspx
[3] http://www.pharmamanufacturing.com/articles/2006/096.html
[4] http://www.pharmpro.com/ShowPR.aspx?PUBCODE=021&ACCT=0000100&ISSUE=0801&RELTYPE=PR&ORIGRELTYPE=ATO&PRODCODE=0000&PRODLETT=G&CommonCount=0