Perhaps the greatest value businesses gain from data comes from extricating deep insights that can influence business strategies, and then forecasting and optimizing future performance based on these insights. Leveraging current and historical data to predict trends and optimize processes can deliver a significant competitive advantage. In the era of big data, companies need high-level computational science and expertise as well as the right technology to facilitate the work. But many businesses today may not have these necessities at their disposal. In a recent survey conducted by Ventana Research,1 results show that 59% of companies cite a lack of resources as the most common barrier to implementing changes in predictive analytics programs.
SAP ensures its customers have access to these resources. For example, the Performance and Insight Optimization (PIO) organization at SAP works collaboratively with customers on their most complicated computational data issues, leveraging both deep expertise in computational science and the power of SAP HANA’s in-memory technology to solve complex business challenges. The PIO organization is comprised of advanced analytical experts, mathematicians, industry experts, and developers who combine their knowledge with technologies like SAP’s analytics applications and SAP HANA to analyze massive amounts of data for business value in specific industries.
“We minor in technology and major in the science of business,” says David Ginsberg, Vice President and Global Head of SAP’s PIO organization. “Business executives aren’t thinking about software when they’re trying to answer complicated questions about out-maneuvering the competition and improving profits. We can help them do this using the right computations and technology. Everything we do starts with data, and today, our work is inextricably tied to SAP HANA. Value can be attained when businesses needs are matched with advanced analytic approaches, which is an art that requires imagination, experience, and rigorous training.”
The PIO team helps companies maximize data by creating mathematical models and providing actionable business insights for solving the C-levels’ most complex issues. In fact, the team is building a track record of high-ROI SAP HANA use cases. Let’s look at four examples.
“Business executives aren’t thinking about software when they’re trying to answer complicated questions about out-maneuvering the competition and improving profits. We can help them do this using the right computations and technology.”
— David Ginsberg, Vice President and Global Head of SAP’s PIO organization
Example #1: Forecasting Sales in High Tech
SAP’s PIO organization is working with high-tech manufacturers and distributors to help forecast future sales and demand for products. For large global companies, forecasting can be a complex challenge given the highly seasonal nature of the business as well as regional differences in the high-tech market. For example, electronic components can be sold alone in one region and as part of a bundled solution in another. And distributors must deal with many suppliers, each with its own pricing and product strategies that vary over the course of a year.
SAP’s PIO organization has helped address these forecasting challenges by:
- Linking disparate data systems to feed into SAP HANA and leveraging large data sets for improved data analysis in real time
- Creating forecasting models based on granular details, with significantly improved accuracy over traditional approaches that capture the daily, quarterly, and annual seasonal variations
For one high-tech firm in particular, forecasting accuracy increased by more than 300%. For businesses with revenues in the billions, this level of improvement can dramatically affect cash flow and overall business management. System performance is never sacrificed for this improved forecasting accuracy. On the contrary, these companies experience response times that are 50 times faster when compared to their existing analytics platform, making the capabilities more useful for business executives.
Example #2: Bringing Operations Research to a New Level
Operations research — the science of better decision making by means of analytical methods — has come a long way since its origins in military planning. Today, many large companies apply sophisticated algorithms to optimize their production schedules, maintenance intervals, supply chain setups, and other strategic or operational decisions.
Given the complexity of such systems and the large number of decision variables and constraints, computation power and system performance has always been a major issue in operations research. Consider the example of large airline companies. Historically, among the front-runners of operations research, airline companies have planes, crew members, and passengers that need to be scheduled or re-scheduled, and are subject to thousands of complex business rules.
SAP’s PIO organization is working with the operations research department of a large SAP customer to improve the performance of its optimization solutions and enable completely new use cases. Helmut Linde, Head of Science and Technology for the EMEA region for PIO, explains, “Their key idea is to take certain pieces of the algorithms — those that are extremely performance-critical and are executed repeatedly — and implement them directly in an SAP HANA database. In this way, a large part of the computations is pushed down into the database, instead of having to pull data out for processing.”
By bringing the algorithms to the data, instead of doing it vice versa, complex computations become possible at the actual speed of business and lead to actionable results. Certain applications of operations research can thus move from tactical planning down to a real-time optimization of operational decisions.
Example #3: Proactive Compliance in Government
Big data is not only about the size of the data set, but also about the combinations. Detecting and acting on fraud, waste, and abuse is a major data challenge for government organizations.
For example, finding and prioritizing businesses and individuals that have not paid their taxes can involve potentially trillions of combinations, making it nearly impossible for some agencies to track. Governments have approached this problem in a reactive, ad-hoc manner. If a person committing fraud fills out the necessary tax form correctly, his or her crime cannot be easily detected. Also, with a limited number of analysts looking for cases of fraud, only a small sample of data can be analyzed.
By combining the right computational science with SAP HANA, the PIO organization developed a proactive method that allows governments to analyze combinations in under 10 seconds and gain actionable information. Now, fraudulent persons can be identified quickly as the data is analyzed holistically.
Example #4: Global Regulations in Manufacturing
In the chemical industry, assuring product compliance to legal rules can be a great challenge — even from a pure data and performance perspective. “For example, if you are designing a new chemical product, a certain substance may only be used up to 0.05% in one country and 0.03% in another,” explains Linde. “There are a large number of these rules because every country and substance has a number of regulations and rules for various products. You end up with millions of compliance rules for hundreds of substances in each product.”
SAP is working with a specialty chemical maker to address this issue by ensuring all of the ingredients used in its custom-designed products are compliant with all global regulations. It’s difficult to quickly compare all of the ingredients with all of the applicable compliance rules. And the product development process involves constantly trying new ingredients to get the recipe right, which means a continuous cycle of compliance checks.
The key to success lies in pushing the algorithms from the application layer down into the database. Instead of retrieving product bill-of-materials and compliance rules from the database and then performing calculations, all the compliance checks are implemented within SAP HANA and the database simply sends back the results of a check to the requesting application. Benefits include decreased waiting times during the product innovation process and the possibility to re-check large numbers of products in a short time when the composition of a certain ingredient changes, for example.
A Long List to Tackle
These examples are just some of the unique, industry-specific challenges that SAP’s PIO organization is tackling with SAP customers. Many of these problems have gone unsolved for years, but the combination of PIO’s modeling and computational capabilities, along with the power of SAP HANA, have opened up a number of new doors — making these previously impossible obstacles achievable.
“For us, everything starts with the data and goes forward from there,” says Ginsberg. “We’re just beginning to unlock the power of this combination, and SAP customers will benefit from our previous work as we progress to solve the most complex business problems our customers face. From how we develop sales strategies based on the behavior of our customers’ customers to forecasting demand and optimizing inventories and supply chains, PIO is driving top-line performance at a time when competitive advantage is hard to find.”
1 Ventana Research, “Predictive Analytics: Improving Performance By Making the Future More Visible” (2012). [back]