service parts optimization

��

Service Parts Optimization

Inventory vs. availability

Table of Contents

Executive summary ....................................................................................................... 1 Just in time vs. just in case .......................................................................................... 2 Service complexity ........................................................................................................ 3

The service parts reward.............................................................................................. 4 The service parts challenge ......................................................................................... 5 Root questions driving service parts challenges.......................................................... 6 Making a service parts system work ............................................................................ 8

The SAS methodology for service parts optimization............................................... 8 1) Forecast generation ................................................................................................. 9 2) Automatic optimization of inventory policies............................................................ 9 3) Alert list and reporting ............................................................................................ 10 4) Forecast review and simulation: demand manipulation......................................... 10 5) Recalculation of inventory and order quantity requirements ................................. 10 6) Inventory simulation: supply manipulation ............................................................. 11

A case study: SAS® Service Parts Optimization....................................................... 11 Basic steps used in the case study............................................................................ 11 Case study results...................................................................................................... 12

The SAS difference...................................................................................................... 13 Conclusion ................................................................................................................... 13 About SAS .................................................................................................................... 13

The primary content provider for this white paper was Bob Davis, a product manager for SAS in Cary, North Carolina. He is responsible for development of demand intelligence and inventory replenishment planning solutions, with a focus on service optimization and the manufacturing industry. Bob has been a featured speaker and writer on the topic of demand-driven supply chains. He has spoken at such conferences as the Council of Logistics Management, Logicon, BetterManagement Live and Frontline’s Supply Chain Week.


1

Executive summary

“Profit margins for aftermarket service and parts range from 25 to 1000% higher than margins for initial products, making service a key focus in corporate agendas.” 1

“Most enterprises do not have adequate systems in place to handle the inherent differences associated with planning and forecasting service parts versus production parts.” 2

Aberdeen Group

The past thirty years have brought innovation, analytics and automation to production and delivery supply chains. From Just in Time (JIT), lean initiatives, and collaborative planning, forecasting and replenishment (CPFR) to sales and operations planning, the supply chain has morphed from a low-level business process to a critical lynchpin of corporate success. Companies like Toyota, Dell and Cisco are leading the way with innovative techniques that both streamline and speed up supply chains to exploit them for savings and process improvement. Their supply chains now react as a synchronized process to deliver the right product at the right time in the right place.

However, the fact that these world-class organizations understand the production/delivery supply chain is only the first part of the equation. They have taken the extra step to recognize that two supply chains exist: one that delivers and one that services. By recognizing the difference and building ways to optimize service, they have turned their service supply chains into powerful competitive differentiators.

This paper will examine how a service supply chain differs from a delivery supply chain. We will look at ways organizations are developing service supply chain techniques to solve their unique challenges rather than falling short of optimal performance with traditional production/delivery methods. Finally, we will explore which business methodologies will foster the best return on investment in service supply chain improvements.

1 Vigoroso, Mark, and Rachel Gecker. The Service Parts Management Solution Report. Aberdeen, 2005. Pg. 6. 2 Ibid.


2

Just in time vs. just in case

Research from the Aberdeen Group states that about eight percent of the gross national product of the United States is tied up in aftermarket parts and services related to repair.3 This statistic means approximately $700 billion are being spent yearly to service assets already purchased. Another study by Bearing Point and APICS shows more than 60 percent of all companies that deal with service parts are handling the tasks with manufacturing supply chain planning systems or enterprise resource planning systems that offer limited means to analyze the unique facets of service parts supply chains.4 Obviously a significant investment has been made to set up, integrate and maintain these systems, and there is a need to generate an acceptable return. There are some companies that have realized ROI by managing service parts with manufacturing or production planning systems, but these results are limited and can cause difficulties.

Here are some tell-tale signs that you should look beyond production technology in managing your service supply chain:

• It is difficult to forecast inventory for slow-moving parts with intermittent demand.

• Shipments frequently have to be expedited because of improper product placements.

• Product or part obsolescence is a problem.

• Service level agreements are getting harder to support.

• The service parts organization is viewed as a liability, not an asset.

• It is necessary to service multiple product life cycles, from new to discontinued.

• The service parts supply chain is more complicated than the system can handle.

• The CEO brags to customers, “We will pull parts off the assembly line to fix your machine if we have to.”

What is different about the service parts supply chain that makes manufacturing-oriented planning methodologies so inefficient? In a nutshell, the focus in service parts supply chains has shifted from just-in-time inventories to just-in-case availability. This seemingly insignificant change alters all aspects of the chain so that it is weighted heavily toward the customer and product rather than spaced evenly and structured solely for a timely flow.

Just in time: Emphasis is on having just enough of an appropriate inventory position to handle the expected need. Information that promotes efficient production runs and even distribution flow is paramount. The key driver is customer orders, and performance is based on inventory turns.

3 Minahan, Tim. The Service Parts Management Benchmark Report. Aberdeen, 2003. Pg. 11. 4 Bearing Point. Service Parts Management: The Focus is the Customer. 2004. Pg. 7.


3

Just in case: Emphasis is on having inventory close enough to the service need to ensure minimum down-time for each asset. Information that enables efficient support of service level agreements (SLA) will be rewarded. The demand cue is an equipment failure, and performance is tied to fill rate and SLA compliance.

Service complexity

Over the past several years, supply chain executives have begun to understand the different analytic needs of production planning and distribution planning systems. Production planning methods have often been applied to distribution situations, but the random events that characterize distribution systems severely limit the usefulness of production planning. Indeed, this challenge has prompted the use of highly sophisticated stochastic algorithms that help better forecast the seemingly unpredictable occurrences inherent in distribution.

Service parts then add a whole new set of complexities. Due to various offerings, constraints and product life cycles, service parts exponentially heighten the need for advanced analytics. The root of this need, as reported by Aberdeen Group, lies in the haphazard way in which service parts are managed.5 The causes are:

• Lack of visibility into service parts inventory levels. Located in hundreds of depots, trucks, shops and warehouses, the service parts supply chain is not really part of the organizational supply chain. In many cases, service part inventories are viewed simply as a liability and are lumped together. Since service parts do not conform to normal KPI’s, they can be difficult to grade.

• Disparate sources of service parts data. Information about service parts can be located anywhere and everywhere in the organization. Manufacturer, distributor, third-party logistics or repair centers can be owned by multiple sources, all with different systems.

• Inconsistent naming conventions for parts and suppliers. Not only is service parts information stored in multiple places, the same parts tend to be named differently in each location. This inconsistency creates confusion and causes demand and supply cues to be missed.

• Disconnected inventory planning and execution procedures. The service chain tends to be planned in silos rather than a responsive, connected chain of distribution. Given the multiple owners and/or locations found in service, it is not surprising that each tries to optimize their own system. As a result, inventory tends to accumulate very close to the end of the chain.

5 Minahan, Tim. Service Parts Management. Aberdeen, 2003. Pgs. 28-31.


4

• Insufficient and fragmented use of automation and advanced analytics. Much of the emphasis for automation and analytics has been placed on the traditional supply chain and not service parts. In turn, the adaptation of manufacturing/production inventory methodologies in service parts have met with limited success. In some cases a spot solution might be implemented that ultimately harms other parts of the chain.

The service parts reward

Organizations have seen more and more commoditization of finished goods. With intense price competition, suppliers have been looking for an edge or a way to differentiate. Taking a page from service-based industries like banking and telecom, the way a supplier can differentiate its offerings is with better after-sale service. Forward-thinking companies have researched this opportunity and found a treasure trove of potential cost savings and competitive differentiation. This decision is consistent with recent studies that have shown an average of 25-30 percent of revenue and 40-50 percent of profit is attributed to after-sales service while IT investment for the service supply chain is typically below 20 percent.

Service and service parts management provide an excellent opportunity to create competitive differentiation and increase the bottom line. They are low risk and develop a long-term revenue stream fueled by customer retention. Moreover, they help with sales acquisition, since loyal customers will recruit others. Lastly, service and service parts management have a huge potential for product development collaboration because they foster almost constant communication and documentation of product usage with the field. In essence, the integration of service and service parts management into the production/distribution supply chain completes a closed-loop supply circle that enhances growth, profitability and brand excellence.

Some benefits to embracing service parts management

Early adopters of service parts management have realized dramatic returns. But the reward for service excellence goes far beyond percentages or numbers on the bottom line. Being a service leader drives the organization to become customer focused:

• SLAs can focus on key metrics critical to the customer.

• Customers who help improve metrics can be rewarded.

• Contracts can be established to provide higher service levels to customers who need it most.


Better fill rates, lower inventories, higher customer retention and SLA compliance are just a few of the other results that can be gained from focusing on service parts management. A poll of service professionals at an Interlog Conference even showed that 75 percent of the companies surveyed now view service and service parts as a profit center, not a cost center.6 If the benefits are so obvious, why doesn’t everyone do it?

The service parts challenge

Figure 1: Opportunities for improvement in service parts management.7

The Bearing Point and APICS survey on service parts management concluded that demand/forecast planning and inventory management are two areas in service parts that provide the best opportunity for improvement. Transactional activities like order management, procurement or warehouse management were far down the list. This result implies that current transactional ERP or supply chain management systems are doing their job.

But the difficulty in service parts is not transactional. The problem lies in the data being delivered to those systems.

5

6 Ibid. 7 Bearing Point. Service Parts Management: The Focus is the Customer. 2004.


6

Industry analysts have noted that companies trying to adapt transactional MRP or ERP systems into service supply chain management tools have had poor results. There are fundamental differences in a manufacturing production system and a service support system, and MRP/ERP solutions were not designed for the latter. MRP systems are fine for transactional control, but for analytically-based business processes that depend on data integration, analytics and data integration are obviously required. A sound strategy, then, is to map your solution to the key business issues of the service supply chain.

Root questions driving service parts challenges

• What will future demand levels be?

• How should inventory be replenished to reduce costs and increase turns?

• When should orders be placed to restock inventory?

• What is the appropriate inventory level?

• What is the projected customer service level?

Understanding how Just in Time intertwines with “just in case” means highlighting the key shortcomings of ERP systems. How you take advantage of these shortcomings will increase the efficiency and effectiveness of the service parts supply chain. There are two key ways to answer the questions above:

Create accurate demand plans and forecasts

Service parts do not act like production parts. Service parts, by their nature, are needed when something breaks down. The window of anticipation is very narrow, and the demand cues come from wide-ranging segments of the business. The demand signal could come from any combination of the following:

• Call centers.

• Customer service reps.

• Warranty claims.

• Service bulletins.

• Third-party providers.

• Parts depots.

• Sales contacts.

• Field technicians.


7

The key is the reaction time to the demand signal. In production the reaction is anticipated and predictable with “known” lead time. Time constraints placed on service parts, such as warranty contracts, service level agreements or even angry customers with a costly asset down time, push service parts to move within 24 to 48 hours of notification.

Two critical dynamics are occurring in this process. A large portion of the service parts inventory will necessarily migrate to the end of the supply chain because of the reaction time needed for contract compliance. The second dynamic is that production-style forecasting techniques feed incorrect volume and placement cues. This is due to the fact that production forecasts rely on robust historical data. The result is a glut of product or potential out-of-stock situations that force expensive expediting of supply chain processes.

Accurate demand planning means service parts forecasting must address intermittent or lumpy demand signals. Techniques exist to overcome this hurdle, but the critical need for service parts forecasting is to have the ability to select automatically the methodology that gives the smallest mean absolute percentage error (MAPE).

Augmenting or adding forecasting functionality inherent to service parts will help align parts with the proper quantities and position your company to better serve its customers. An increase in accuracy of 5 to 15 percent can have a huge effect on both the supply chain and the bottom line. These benefits are neither theoretical nor anecdotal. They are based on large empirical studies, including one conducted on consumable spare parts for the British Royal Air Force.8

Creating accurate inventory plans and forecasts

Building an inventory strategy for service parts requires the ability to measure a lot of places at the same time. Service supply chains tend to have a “distribution explosion” at the end. Rather than having a finite number of warehouses, linearly spaced, that ship to final customers, the service chain ends with shops and/or trucks that are expected to have the right parts to fill an order immediately. The service chain becomes a constant battle of repair speed, storage cost and storage space.

The key is to develop optimized, individual inventory policies for each part of the chain. This means chief parameters like reorder points, order-up-to points and lot sizes must be aligned with the corresponding fixed order and holding costs, then matched with other variables like delivery lead times and lead time variation. While it seems daunting, today’s advanced analytics can achieve this kind of process in a relatively short period of time. The result of these kinds of advanced policies is that you can link a highly accurate service parts demand forecast to a finely tuned supply model. The linkage of supply and demand is then translated to the required service level. As a result, inventory levels will be appropriate to comply with contracts and SLAs.

8 Power, R, S. Reynolds, et. al. ”Expert Provisioner: A Range Management Aid.” Applications and Innovations in Expert Systems. Eds. A. Macintosh and R. Milne. BCS Expert Systems Group: Cambridge, 1997.


8

By understanding and executing on optimized inventory policies, the service chain can be aligned to take advantage of such things as inventory placement. Moreover, if multi-echelon optimization is used, the entire chain can act and react as one entity rather than waiting for down-stream demand to work its way through each link of the chain. Indeed, one of the advantages of multi-echelon optimization is identification and elimination of redundant inventory that results from the infamous “bull-whip effect.”

Making a service parts system work

Most organizations already have the building blocks in place for a better service parts supply chain. Success is a question of shifting the emphasis from just in time to just in case. This simple, yet sophisticated, change in point-of-view will yield huge benefits. But the complexities of service part supply chains—including their many points of origin, vaguely named parts, multiple points of distribution and odd demand patterns—require advanced analytics in order to run at peak performance. It has been proven that production planning techniques do not work well.

By adding superior analytics to optimize the demand and supply differences found in service parts, existing ERP, warehouse management systems (WMS) or supply-chain management systems (SCM) can perform the dual duties of both production and service supply chain work.

The SAS methodology for service parts optimization

SAS Service Parts Optimization is specifically designed to help service organizations predict demand for parts and calculate optimal replenishment policies to reduce inventory costs without compromising service levels. This solution operates according to a sophisticated methodology (see Figure 2) and incorporates industry-leading analytics for the most accurate results. The methodology involves the following six steps.


Figure 2: The SAS methodology for service parts optimization.

1) Forecast generation

Demand data comes in from various parts of the organization. Included in this process are both short- and-long term inputs for improved accuracy. For instance, data mining and clustering techniques can be used to determine demand for long-term forecasting needs. This technique is far superior to time series forecasting.

Once the demand data has been staged, the system automatically assesses which forecasting techniques are needed for best accuracy. Then a demand forecast is fed by batch process to the inventory optimization system.

2) Automatic optimization of inventory policies

As the demand forecast is entered into the inventory optimization engine it is matched on an individual part or product basis to the relevant warehouse and procurement cost data. This data will come from many disparate sources along with normal transactional data found in ERP, SCM and WMSs. This information is then used to produce optimized polices and give the proper inventory inputs, such as reorder points, order-up-to points and order lot sizes. This is done for all inventory points and all parts. Given that service parts tend to be driven by intermittent and/or lumpy demand, the review process for inventories is done on a continuous basis.

9


10

Once optimized inventory policies are placed against the demand forecasting inputs, the system is ready to send the information back to the organization’s appropriate ERP, WMS or SCM system for execution. If there are no alerts that need to be reviewed, the system will run automatically.

3) Alert list and reporting

The alert and reporting system provides performance information to the needed product/part/chain hierarchy. The kinds of benchmarking provided are:

• Average backorders.

• Average cost – holding, ordering, backorder penalty.

• Average inventory.

• Average order frequency – replenishment orders.

• Inventory ratio – average inventory/average demand.

• Turnover – average demand/average inventory.

• Service rate metrics – backorder ratio, fill rate, ready rate.

Depending on the parameters loaded by the organization, alerts are triggered at predetermined over-stock or under-stock levels that may require manual intervention. However, there may be times when further analysis will be needed. This can be done by manipulating the demand or supply information. In addition, this review and simulation alternative can be used to better understand supplier decisions before implementing them.

4) Forecast review and simulation: demand manipulation

From the alert list there are parts in various distribution points that have either high or low inventory positions projected. It might be that the forecast chosen by the system as the best alternative does not match known or unknown causal factors. By changing the forecast model, the user will be able to reduce the MAPE for better results.

5) Recalculation of inventory and order quantity requirements

Once the demand forecast has been reviewed and approved, the new inputs are matched to the inventory policies and recalculated. This data now provides an updated input to the execution system. It is reloaded to the execution system, and the forecast and replenishment system is revised accordingly.


11

6) Inventory simulation: supply manipulation

Over time, parts and products change their delivery patterns. It might be that they are shipped in different lot sizes or a supplier’s lead time is questionable. For your organization to make the right decisions, information regarding the changes can be loaded and recalculated with ease in the simulation area. The information is then reoptimized and reviewed. Once the information is deemed acceptable, it is saved as a new policy parameter for the next inventory optimization batch run.

A case study: SAS® Service Parts Optimization

In this case study, the client’s service supply chain was providing good support for SLAs, but management deemed the inventory too high. SAS used a portion of the parts list and ran the SAS methodology against the organization’s historical data to evaluate the optimal inventory levels for that period.

Original data

• More than 400,000 parts (SKUs).

• Demand and inventory history over 118 weeks.

• Weekly inventory amount: $11.5 million.

Case study data

• 11,605 SKUs.

• Demand activity occurred over 118 weeks.

• Weekly inventory amount: $785,000.

Basic steps used in the case study

For each month from week 53 to 118, SAS:

• Forecast the demand for the next month using demand history prior to the current month.

• Computed the weekly optimal inventory policy using the demand forecast.

• For each week in the current month, evaluated the optimal policy with the actual weekly demand and inventory balance carried over from previous weeks.


Case study results

0

200000

400000

600000

800000

1000000

1200000

53 57 61 65 69 73 77 81 85 89 93 97 101 105 109 113 117

Week

Am

ount

ActualOptimal

Figure 3: Inventory profile over time for the case study.

Actual profile Optimal profile

Weekly inventory cost $785,148 $571,823

Non-stockout percentage

96.3% 98.8%

Figure 4: Results of the case study – actual vs. optimized inventory profiles.

As stated above, the client had strong service levels, but the inventory required to maintain that compliance was extremely high. By using the forecasting techniques included in SAS Service Parts Optimization, the client could have reduced the MAPE, matched it with individually optimized inventory policy parameters, seen a further increase in service and still managed a 27 percent reduction in average weekly inventory over the life of the test period. In addition, the client could have anticipated lump demand signals.

12


13

The SAS difference

• Forecasting depth – Accurate, fast and automated hierarchical forecasting (down to SKU level).

• Causal forecasting – Causal forecasting techniques for more accurate prediction of demand.

• Intermittent forecasting models – Models specific to service parts demand forecasting.

• Data mining for long-term forecasting – Data mining and clustering techniques to determine demand for long term forecasting needs. This technique is far superior to time series forecasting.

• True multi-echelon optimization – State-of-the-art simulation based on stochastic optimization algorithms to calculate optimal, multi-echelon inventory policies for complex supply chain networks with complex business rules.

• Scalability – Scalability to encompass thousands or even hundreds of thousands of items makes this solution especially well suited for large service networks.

Conclusion

We are seeing a transformation of the service parts supply chain. As profits are squeezed out of finished goods, companies are looking at ways to generate new, long term profit opportunities and fuel growth through customer lifecycle management. The service parts and service management area has emerged from being a largely ignored cost center to being a high-value business process capable of generating profits and distinguishing a brand. SAS Service Parts Optimization is designed to shift the focus of the service supply chain from just-in-time inventory to just-in-case availability.

About SAS

SAS is the leader in business intelligence software and services. Customers at 40,000 sites use SAS software to manage and gain insights from vast amounts of data, resulting in faster, more accurate business decisions; more profitable relationships with customers and suppliers; compliance with governmental regulations; research breakthroughs; and better products. Only SAS offers leading data integration, intelligence storage, advanced analytics and traditional business intelligence applications within a comprehensive enterprise intelligence platform. Since 1976, SAS has been giving customers around the world The Power to Know®.

��

��

��

��6�� 102517US_378537.0506

service parts optimization

Documents

service optimization

sas service parts optimization

service parts challenges

service parts reward

aftermarket service

service complexity

service parts system

parts range