Click for next page ( 28


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 27
IV. PERFORMANCE INDICES FOR ASSESSING INVENTORY MANAGEMENT 4.1 INVENTORY MANAGEMENT PERFORMANCE OBJECTIVES Inventors management theory defines two cor~icting objectives of inventor management: minimize the amount of inventory and maximize (or maintain) the availability of inventory items. These objectives are cor~icting since increasing inventory can have the effect of increasing availability and reducing inventory can lead to a reduction in availability. The primary task of inventory management is to effectively balance these two objectives so that inventory is available to sufficiently support the demand for inventory items, while at the same time controlling the dollars tied up in inventory. Inventory management performance indicators usually measure the performance of one of the two primary objectives. For example, "total inventory dollars" measures the amount of inventory, while the "percent of demand" that is filled measures the availability (or service level) provided by the inventory. A third category of performance indicators measures the level of effort and cost required to manage inventory. Minimizing the cost of managing inventory is also an objective of inventory management. Examples of performance indicators for this category include the personnel cost per inventory dollar of personnel assigned to inventory functions and the percent inventory carrying cost. The fourth category of performance indicators measures the accuracy of inventory records, such as the percent of items whose perpetual balance is incorrect. Section 4.2 of this chapter defines the inventory management performance indicators that are commonly used by the public transit industry, based on our survey results. In addition, Section 4.2 defines other useful performance indicators that can be derived Dom the survey data. Section 4.3 of this chapter examines the effects of agency characteristics (such as the population served, operating budget, and number of annual passenger miles) and fleet characteristics (such as the number of foreign manufactured vehicles, the average age, etc.~. These characteristics can be found in Sections ~ and ~ of the survey, respectively. Section 4.4 presents the values of the performance indicators, based on the survey responses. These values can be used for benchmarking by individual transit properties. Section 4.5 compares the values of the performance indicators for the public transit industry with the values nonnally found in other industries. 27

OCR for page 27
4.2 PUBLIC TRANSIT INVENTORY INDICATORS MANAGEMENT 4.2.1 Use of Inventory Management Performance Indicators -- Survey Response PERFORMANCE Our survey shows that most public transit properties use a small number of indicators to monitor inventory performance. Moreover, many properties, particularly those with less than 50 vehicles, do not formally monitor inventory performance. These properties merely set minimum and maximum levels for inventory items to control replenishment and address parts shortages as they occur. The table below shows the inventor perfonnance indicators named in the survey and the percent of survey respondents in each size category that use the indicators. The size of the properties is based on the number of vehicles as follows: Small Properties Medium Properties Large Properties Vely Urge Properties 50 or fewer vehicles (3 ~ survey respondents) 51 - 300 vehicles (33 survey respondents) 301 - 2000 vehicles (15 survey respondents) over 2000 vehicles (7 survey respondents) __ . . . . . . . - . . . . . . . - . ;. . . . . . . ......... . . Resonance lnolmton - . . bma 1 . MmIum: ~e: ~ ; very . . :Iotal ~ . : . , .. . . .; ....... ...... .... .. . . . . ... ........ .. . ;. . . . ~ .; . .. .,., ........ ,., . ,.,, . . . , . . , .:: ; . . . . . Inventory Amount Indicators 19% 48% 47% 57 /o 41% Toad Inventory Dollars ~19% T 39% ~27% T 57% ~31% DoDars per Vehicle | 0/O | 9% | 20% r 0% | 10% Inventory Turnover | 0% | 9% ~7% | 29% | 7% Availability/ServiceIndicators | 10% | 45% 1 47% | 86% 1 36% % rhxnand Feed 1 0% ~3% 1 13% 1 57% 1 8% Number of Stockouts 1 0% | 21% | 27% | 0% 1 13% Number of Back Orders 1 60/0 T 210/0 ~27% ~860/0 ~22% TirnetoF~Backorders | 3% | 0% | ~ 0% | 0% | 1% Vehicles Out of Service | 0% | 6% | 7% | 14% | 5% _ Inventory AccuracyIndicators ~32% ~73% T 60% ~86% r 57% Dollar Vananoc ~10% 1 42% 1 27% 1 29% 1 27% Item Vanance 1 23% 1 54% 1 47% 1 71% 1 43% Management Cost Indicators | 0% | 9% | 7% | 0% | 5% % Inventory Carrying Cost ~0% ~9% | 7% | 0% | 5% . . . . . .. 28

OCR for page 27
As the above table shows, only 41% of the survey respondents use inventory amount indicators, 36% use availability or seance level indicators, 57% use inventory accuracy indicators, and only 5% use indicators for the cost of managing inventory. Many of the respondents stated that inventory performance is tracked less formally, using annual persome1 perfonnance reviews and complaints or comments Dom inventory "customers". Although many of the survey respondents do not regularly track the above indicators, most were able to provide values in response to the survey questions. The respondents also provided information to calculate and examine additional indicators, such as the percent of obsolete inventory and the number of inventory personnel per vehicle. Therefore, the survey provides sufficient information to analyze and benchmark inventory performance indicators, even though some indicators may not currently be in widespread use in the public transit industry. 4.2.2 Definition and Calculation of Performance Indicators In some cases, the survey respondents defined or calculated perfom~ance indicators differently. addition, there are some indicators that were not explicitly solicited in the survey, but can be calculated Dom other survey data. This section presents standard definitions and calculation methods for the inventory management perfonnance indicators and describes the inventory management attributes that each measures. Inventory Amount Indicators Total Inventory Dollars "Total inventory dollars" is the total cost to the transit agency of ad items held in inventor at a given point in time. It is calculated by multiplying the number of units for each item times the item's unit cost, and summing across ad items. This indicator measures the size of inventory in terms of the dollars that the transit agency has tied-up in inventory assets. It is best used to monitor changes In the size of inventory (increases and/or decreases) by exaniLriing the value at different points in time. Inventory DolIa" per Inventory dollars per vehicle is the average amount of inventory dollars Vehicle on-hand at a point in time to support a vehicle in the transit agency's fleet. It is calculated by dividing the "total inventory dollars" by the number of vehicles using items Dom the inventor. This indicator measures the size of inventor, in dollars, that the transit agency holds to support a vehicle. It eliminates the effect of fluctuations in Beet size when monitoring inventory levels across time. It is also useful when comparing the relative size of inventory across Beets with different numbers of vehicles, different makes and models, different modes, etc. Inventory Turnover Inventory turnover is the number of times the "total inventory dollars" is used by inventory customers in a given period of time. For example, 29

OCR for page 27
annual turnover is calculated by dividing the total dollar value of the items used Mom inventory dunna the year by the average total dollar value or items held in inventor during the yew. We average tote dollar value of items held In inventor for the year can be calculated by taking the average of"total doDar inventor' levels measured at different times during the year (for example, at the end of each month, or the beginning and ending levels for the years. Turnover can be calculated for any time period. For example, monthly turnover is calculated by dividing total monthly dokar usage bv the average tow doLar inventory for the month. ---cat- - ~- -~- -- - Inventory turnover indicates inventor size, In dollars, relative to the amount of inventory that is used dunng a given time period. For example, an annual inventory turnover of 2.5 means that the transit agency uses two and a half times the amount of dolDars it holds In inventory. In other words. inventory is "turned overt' 2.5 times dunna the year. Since an objective of Inventory management Is to maze Inventor levels, the higher the inventory turnover, the more efficiently the inventory level is managed relative to the demand for inventory items (usage). As an indicator, inventory turnover attempts to compensate for the size of demand when mon~tonng inventory levels, and is widely used to compare inventory performance across time and between different organizations. Note: The survey respondents provided inventory turnover values based on a variety of time periods and average "total inventor dollars". To ensure consistency, the inventory turnover was calculated for each agency using the 1993 total inventory dollars and the average monthly usage times 12. Months on Hand "Months on hand" is the number of months that a transit agencies inventory will last if no additional items are added to inventor. Months on hand is the inverse of monthly inventory tumover. It is another way to measure the size of inventory relative to the demand for inventory over a specific tane period. It is calculated by dividing the average "total inventory dollars', for a month by the total dollars used - from inventor during the month. The fewer months that a transit agency must keep on hand to support the demand for inventory items, the better the perfonnance relative to minimizing inventory levels. This indicator can also be calculated for different time periods, such as "days on hand" or "years on hand". Months on hand is a figurative rather than a literal indicator in that it is a measure of how long an agency's inventory dollars grill last. This measure assumes that the items on hand are exactly the items that will 30

OCR for page 27
be used during the time period. Like turnover, it attempts to compensate for the effects of demand levels on the size of inventory. Note: Sconce months on hand and inventory turnover are mathematically the inverse of each other, only inventor turnover is benchmarked and examined in this report. Any inferences regarding turnover are also valid for months on hand, and benchmark values for months on hand can be determined using the inverse of the benchmark values for monthly turnover (or annual turnover divided by 12~. AvailabilitY/Service Indicators % Demand Filled (Fill The percent of demand fired, or the inventory fib rate, is the percent of Rate) items requested from inventor,, that are provided from inventory at the ~ . _ ,, ~ ~ time of the request. It is calculated by dividing the total number of items requested from inventor into the total number of items issued from inventory at the time of request during a given time period. This indicator measures the level of availability of inventory items. It also defines the probability that an item will be available from inventory when it is needed. The fill rate is used to monitor how well the items held in inventory match the items that are needed over a given period of time. It is also used to compare inventory management performance, regarding availability, between organ cations. Number of Stockouts The number of stockouts is the number of unanticipated times that active inventory items reach a zero balance on hand during a specified time period. This indicator measures the exposure of inventory to potential unfilled requests. Only unanticipated stockouts are counted because, at times, some items are carried at zero balance on a planned basis (such as seasonal items or items that are ordered only on request). Unanticipated stockouts are counted regardless of whether there is an outstanding request for the item. The fill rate measures the ultimate availability of inventory matenal, but the number of stockouts indicates the degree to which fate is tempted. % of Items Stocked The "percent of items stocked out" measures the percent of the total Out number of inventory items (i.e. part numbers ) that reach a zero balance during a Even Period of time. The total number of inventors nart - ~ r~~ numbers are called "stockkeeping units" or SKUs. This indicator is calculated by dividing the number of unanticipated stockouts during a given period of time by the total number of SKUs held in inventory. This indicator provides a measure of exposure to unfilled requests that is relative to the size of the inventory, in SKUs. It provides a measure that can be compared across time regardless of the number of items 31

OCR for page 27
added or removed Tom Inventory, or between organizations with different numbers of SKUs in inventory. Number of Open Back The "number of open backorders" is the number of unfilled requests for Orde.rs inventory material that exist at a given point in time. It is calculated by counting the number of inventory items that have been requested and are currently unavailable Dom inventory. It is used to focus inventory management activity and to monitor the status of the availability of inventory items at a given point In time. ~ en e ae Time to Fill Backorder A smear Indicator is the "average number of open back orders". This Indicator measures the typical status of Inventory availability by averaging the number of open back orders at several points in time. The "time to fig backorders" is the average time it takes to provide an Inventor item that is unavailable at the time of request. The time to fib a backorder is the time period beginning with a request for an unava~iable Inventory item and ending at the time the item is provided to the requester. These instances are averaged over a period oft~me to yield the "average time to fib backorders". This indicator measures inventory management performance in resolving unavailable inventory items. It can also be used to compare performance across organizations. Vehicles Out of Service "Vehicles Out of Service" is the number of times a vehicle is held out of service due to unavailability of inventory items. A vehicle is counted each tune it misses a service run, even if the same part is unavailable. This indicator measures the eject of inventor availability on transportation service provided by a public transit agency over a specified period of time. % of Fleet Out of The "percent of fleet out of service" is the percent of a transit agency's Service Beet that is held out of service due to unavailable inventory items. It is calculated by dividing the ``vehicles out of service,' for a service run by the total number of service vehicles. This indicator can be averaged over a period of time to provide the average percent of Beet out of service. This indicator measures the effect of inventory availability on transportation service, relative to the total fleet size. It can be used to compare inventory performance across different fleet sizes and different - orgamzatlons. 32

OCR for page 27
Inventory Accuracy Indicators Dollar Vanance The "dollar vanance" is the difference between the "total inventory dollars" based on a transit agency's inventory records (book value) and the "total inventory dollars" based on a physical count of inventory items. This indicator measures the effect of inventory accuranv bat _ _ 41 _ _ _ _ . ~ ~t ~ . ~ on one aggregate Sonar value the inventory. It teds how inaccurate an agency's records are based on total inventory dollars, and is used to adjust the book value of inventory. Absolute Dollar The "absolute doDar vanance" is the sum of the doLar variances for Variance each individual inventory item (SKU). It is calculated by summing the absolute value of the difference between the dollar value of each item based on inventory records and the dollar value based on a physical count of the item. In using the absolute value, variances that are negative and positive will not cancel each other out. This indicator provides a more comprehensive picture of the accuracy of Individual inventory item values and an overall measure of the accuracy of "total inventory doLar,' records. % Absolute Dollar The "percent absolute dollar variance" is the "absolute dollar variance" Variance divided by the "total inventory dollars". It measures the overall accuracy of inventory values relative to the size of inventory, in doLars. ~ . . .. . . . . .. . 1 IS indicator can be used to compare the overall accuracy of inventory value between organizations or over time, regardless of the total value of inventory. Item Variance The "item vanance" is the difference between the number of units on hand of an individual item based on a transit agency's inventory records and the number of units based on a physical count of the inventory item. This indicator measures the accuracy of each item's perpetual balance records, and is used to adjust the "quantity on hand" records for each inventory item. This measure is not nonnaby summed to give a variance for the total number of units, however it can be averaged to give the average variance for an inventory item. This "average item vanance" measures the average number of units that an item's physical count varies from the inventory records. Number of Items Out The "number of items out of balance" is the total number of inventory of Balance items (SKUs) for which an "item variance" exists. It is determined by counting the number of items for which the physical count does not match the inventory records for quantity on hand. While the average item variance measures the average size of the discrepancy between the physical count and the inventory records, this indicator measures the actual number of items that have a discrepancy. 33

OCR for page 27
% of Items Out of The "percent of items out of balance" is the number of items (SKUs) Balance out of balance divided by the total number of items in inventory. It measures the overall accuracy of perpetual inventory balances relative to the size of inventory, in SKUs. This indicator can be used to compare the overall accuracy of inventory balances between orgar~tions or over tune, regardless of the number of items stocked in inventory. Management Cost Indicators % Inventory Carrying The "percent inventory carrying cost" is the cost of maintaining Cost inventory divided by the total dollar value of the inventory. The cost of maintaining inventor includes the following components: storage cost, the cost of storage space and equipment insurance cost, the cost, if any, of Ensuing inventory obsolescence, the cost of items that become obsolete (e.g. due to changes In Beet series) shrinkage, the cost of inventory items that become missing, damaged, spoiled, decayed or otherwise unusable capital cost, the opportunity cost associated with investing dollars in inventor rather in other assets This indicator measures the "overhead" costs Involved In maintaining inventory. One or more of the above components may be excluded Dom the calculation if it does not apply. For example, in some cases storage space is absorbed by other transit functions (such as vehicle maintenance). % ObsoleteInventory The "percent obsolete inventory?' is the cost of obsolete inventory items divided by the total dollar inventory value. Although this indicator is also a component of canying cost, many organizations track it separately. Items may become obsolete to a transit agency due to changes in the mix of vehicle series in the fleet, changes In parts design, changes in part quality specifications, etc. If these items remain in inventory, the transit agency will incur the cost of carrying items that it cannot use. This indicator assists in measuring the degree to which inventory management anticipates and reacts to changes in fleet mix and parts storage requirements. Inventory Dollars per "Inventory dollars per person" is the total inventory dollars divided by Person the number of people with inventory management and control 34

OCR for page 27
responsibility. Inventory management and control personnel are the entire inventory staff, including stores personnel, inventory planners, clerical personnel, etc. This indicator provides a measure of Inventory management and control staffing levels relative to the size of the inventory, In doDars. It can be used to compare staffing levels across different organizations. Inventory Dollars to "Inventory dollars to personnel dollars" is the ratio of total inventory PersonnelDoRars dollar value to the total cost (salary and fringe) of the personnel charged with managing and controlling the inventory. This indicator provides a measure of the cost of inventor management and control personnel relative to the size of the inventory, in doldars, that is being managed. It can be used to c~mnare Tiffing Alec Urn Ai~PrP"t - organizatio'L. ~;, ~__ _~_ _~ Inventory Transactions "Inventory transactions per person" is the average number of inventory per Person transactions (issues, receipts, transfers, returns) per person for individuals with inventory management and control responsibility. This indicator measures the activity level of inventory material now relative to the number of people in the inventory organization. It can be used to compare relative workload of inventory personnel across different organizations. 4.2.3 Inventory Performance Indicators Used for Benchmarks Many of the inventory performance indicators defined above are absolute measures that are best used to monitor a specific inventory over a period of time, e.g., total inventory dollars. Other indicators measure performance relative to a standard factor, such as inventory dollars per vehicle. These indicators better lend themselves to comparing performance between different inventories and are more meaningfi~! as benchmarks. In addition, the survey data was unavailable or inconsistent for some indicators, such as carrying cost and percent of fleet out of service for parts. Further, the survey quantifies some indicators by mode (bus, rail, and other). Bus and rail responses were acceptable; however. responses in the "other'' category were veal inr.On~iCtPnt ATOP `^ Honda A;~O ;- , ~ ~ J ~ J ~---I__- Am_ ~ Van me- ~ ~1 ~11~O ~1 . . . . . Interpretation by respondents. As a result, benchmark values for the following inventory management performance indicators will be quantified. 1. 2. 3. 4. 5. 6. 7. 8. Bus Inventory Dollars per Vehicle Rail Inventory Dollars per Vehicle Annual Bus Inventor Tumover Annual Rail Inventory Turnover Bus Percent Demand Filled Mill Rate) Rail Percent Demand Filled (Fill Rate) Percent of Items Stocked Out per Week Average Days to Fill Bus Inventory Backorders 35

OCR for page 27
Average Days to Fib Rail Inventory Backorders Percent of Items Out of Balance Percent Obsolete Bus Inventory Percent Obsolete Rail Inventory Total Inventory Dollars per Person Total Inventor Dollars to Personnel Dollars Total Inventory Transactions per Person 4.3 TE1: EFFECT OF TRANSIT AGENCY AND FLEET ClIARACTERISlICS ON PERFORMANCE INDICATORS 4.3.! Objectives for Analyzing Agency and Fleet Characteristics Inventory management indicators can be used to monitor and evaluate inventory management performance. Furthermore, benchmark values for the indicators can serve as a yardstick for comparing inventory management performance between departments, organizations, and entire industries. As part of this comparison, it is important to identify any effects that relate to the characteristics of the organizations being compared. In particular, for comparing benchmark values between public transit agencies, the characteristics of the public transit agency or the agency's fleet may have an impact on inventory performance. For example, the population of the agency's service area, the number of annual passenger miles, the percent of foreign manufactured vehicles, or the average age of the Beet may have identifiable elects on inventory perfo~nance indicators. These effects should be taken into account when using benchmarks. The objectives of analyzing the effects of agency and Beet characteristics are to: (~) identifier which characteristics, if any, affect which inventory management indicators; (2) isolate and quantify the effects; and (3) develop a decision modeling guide to assist a public transit agency In identifying the appropriate benchmark values that pertain to the agency, based on the agenc,r's characteristics and Beet profile. For example, an agency with a Beet size greater than 1000 vehicles that serves urban and suburban ridership may have different benchmark values for inventory turnover than an agency with less than 50 vehicles serving a suburban and rural area. 36

OCR for page 27
4.3.2 Methodology for Analyzing Agency and Fleet Characteristics The data provided In survey Sections I and II, Agency Profile and Fleet Profile respectively, were used for agency and fleet characteristics. Data from survey Section V, Inventory Management Performance, were used to calculate values for inventory management indicators. Two primary statistical methods were used for the analysis, based on whether the characteristic being examined was numerical (e.g. annual passenger molest or divided into "categories" (e.g. rural, suburban, urban, etc.~. 4.3.2. 1 Correlations for Numencal Data The conception coeff-cienf and the associated coe~fcienf of determination were used to analyze numerical data. The correlation coefficient (r) measures the degree to which variables vary together. The value of r ranges from -1 for variables that are negatively correlated (as one variable goes up, the other goes down), to O for variables that are not correlated at all, to +1 for variables that are positively correlated. Values In between indicate the degree of correlation. The closer to +1 (or 1), the higher the correlation between the variables. The coefficient of determination (d) equals r2 and measures the percent of variance in one variable that can be associated with variance In the correlated variable. For example, annual bus passenger miles can be correlated with bus inventory dollars per vehicle to quantify the degree of the relationship between these two variables. The resulting value of r, 0.253, measures the degree of the relationship. Although the variables are positively correlated, the degree of correlation is low, since the value of r is not close to I. The coefficient of detenrunation, d, is rid, or .064. This indicates that only 6.4% of the variance in bus inventory dollars can be associated with the variance In annual bus passenger miles. This correlation is not high enough to consider annual bus passenger miles as a factor when monitoring bus inventory dollars per vehicle. 4.3.2.2 l-Tests for Category Data Some sunrey questions asked the respondent to check a category, such as whether the agency's service area was urban, suburban, rural, or some combination. In addition, some numerical responses were grouped into categories indicating ranges, for example, service area population between 500,001 md 1,000,000. Since the number of survey responses for most individual categories was less than 30, the standard normal distribution could not be used to test the difference between category averages. In addition, the number of responses in each category were different and do not represent "paired" data. Therefore, the t statistic for small, unequal sample sizes was used to test whether average values of inventory performance indicators are different for different response categories. The methodology for the t-test is to hypothesize that there is no difference in inventory performance indicators between categories (null hypothesis), calculate the t statistic, and determine the 37

OCR for page 27
confidence level with which the nub hypothesis can be rejected. In other words, if the confidence level is high enough (e.g. 90%), one can conclude that there is a difference In performance indicators for different categones. For example, the average bus inventory turnover for the 24 responding transit agencies serving only urban areas is 2.00, with a standard deviation of 2.43. The average for the 1 5 responding agencies that serve both urban and suburban areas is 2.62, with a standard deviation of 2.~. The nub hypothesis is that there is no significant statistical difference between the category averages, 2.00 and 2.62. In other words, the nub hypothesis implies that the difference between the average bus inventory turnover from the survey sample is due to chance and does not represent a real difference between transit agencies. The resulting confidence level for rejecting the nud hypothesis indicated by the t- statistic is 59.4%. Therefore, there is only 59.4% confidence that the difference between the sample averages, 2.00 and 2.62, represents a real difference in bus inventory turnover. Conversely, there is a 40.6% confidence that the difference observed from the survey is due to chance. A minimum confidence level for rejecting the null hypothesis is usually 90%. Therefore, based on this test, we cannot conclude that there is a difference between bus inventory turnover for agencies serving only urban areas and those serving both urban and suburban areas. Even though there is an observed difference of 0.62 in the category averages, the large standard deviations indicate that there is enough variance In both categories to prevent this difference Tom being statistically significant. 4.3.3 Effect of Transit Agency Characteristics on Inventory Performance Indicators The following transit agency characteristics from survey Section I (Agency Profile) were analyzed: Service Area Characterized as Urban/Suburban/Rural Service Area Population Agency Operating Cost Annual Material Purchases by Mode (Bus, Rail) Annual Passenger Miles by Mode (Bus, Rail) 4.3.3. I Service Area - Urban/Suburban/Rural The survey responses for urban/suburban/rural were analyzed four ways. First, l-tests were conducted using each unique category of response. Then three sets of l-tests were conducted to exaIT,ine the inclusion or exclusion of urban, suburban, and rural areas. The following categories were used: Category Set ~ Urban Only Suburban Only Rural Only UrbarJSuburban UrbanlRural 38

OCR for page 27
Suburbar~ral Urban/Suburban/Rural Category Set 2 Category Set 3 Category Set 4 Urban No Urban Suburban No Suburban Rural No Rural Average values ofthe inventory performance indicators listed in section 4.2.3 for each category within a set were tested against others in the set. This resulted in 24 l-tests for each of 15 inventory performance indicators, for a total of 360 t-tests. In order to conclude that the average values of the inventory performance indicators were different between categories, a category had to have at least three (3) responses and the t statistic confidence level had to be at least 90%. Based on these criteria, only a few isolated categories have different average inventory performance indicators. The performance indicators and the categories are listed below with the number of responses (in parentheses) and the confidence level. CategoIy Set 1: Bus dollars/vehicle ~ Urban Only (31) and Suburban Only (5) - 98.7/O Bus dollars/vehicle-Urban Only (3 1) and Urban/Suburban/Rural (9)-94.3/O Bus dollars/vehicle-Suburban Only (5) and Urban/Suburban/Rural (9) - -99.7% Bus dollars/vehicle Urban/Suburban (18) and Urban/Suburban/Rural (9) - 97.~% % items out of balance-Urban Only (20) and Urban/Suburban (14) - 94.~% InventoIy transactions/person-Urban Only (22) and Suburban Only (6) - 94.4% CategoIy Set 2: Category Set 3: Bus % obsolete items - Urban (41) and No Urban (7) - 96.3% Inventory dollars/person-Urban (57) ant! No Urban (a) - 92.6% Inventory transactions/person-Urban (49) and No Urban ha) - 99.6% Rail dollars/vehicle-Suburban (10) and No Suburban (3) - 90.~% Stockout % of SKUs-Suburban (25) anti No Suburban (15) - 96.~% % items out of balance ~ Suburban (27) and No Suburban (23) - 98.4% 39

OCR for page 27
Category Set 4: Bus turnover -- Rural (13) and No Rural (42) -- 94.9/0 Bus dollar~ve~cle -- Rural (13) and No Rural (53) -- 96.0% In summary, ondy 14 of the 360 l-tests resulted In categories that had statistically significant differences In perfo~ance indicators. Among the 14 categories, there was no consistent pattern. Although bus dollars per vehicle was affected by five categories, the results were not sufficient to draw a conclusion. 4.3.3.2 Service Area - Population Survey responses for each of the following service area population categories were analyzed against the others for each ofthe 15 inventory performance indicators, resulting In 225 t-tests: Over 1,000,000 500,001 - 1,000,000 200,001 - 500,000 100,001 - 200,000 50,001 - 100,000 Less than 50,000 Using the same cntena, a minimum of three responses per category and a 90% confidence level, the following differences were found: Bus dolIa~IveWcle - 2-500K (15) and 50-1OOK (6) - 92.6% Bus % obsolete - >1~1 (14) and 1-200K (10) - 99.7% Bus % obsolete - 50~100K (4) and 1-200K (10) -- 97.5% % out of balance - 500k-1~! (7) and 1-200K (10) ~ 90.1% Inventory dollardperson ~ >154~4 (21) and 2-500K (15) -- 95.3% [nventor~v dollar~person -- >1~1 (21) and 1-200K (13) ~ 98.8% [nventor~v dolIardperson -- >1~! (21) and SO-NOOK (4) -- 92.1% Inventory transactiondperson -- 2-SOOK (13) and SO-NOOK (3) ~ 92.7% Only ~ of the 225 l-tests resulted in statistically significant differences In perfonnance indicators. Of these 8. inventory dollars per person appears to be different for agencies serving populations over 1 million than for the other categories. However, upon closer exan~mat~on, this difference was due to the fact that most properties with rail service are In this category. (As will be shown later, raid properties have higher inventow levels.] When the analysis was corrected for the eject of rail properties, there was no significant , _ in, . _ . as. difference due to the population area. 40

OCR for page 27
4.3.3.3 Agency Operating Cost , , Survey responses for total transit agency operating cost serve as one measure of the size of a transit agency. The total operating cost responses were correlated with each of the 15 inventory performance indicators with the following results: Number of Correlation Coefficient of Variable Responses (n) Coefficient (r) Determination (r2) Bus inventory turnover 52 .0747 .0056 Rail inventory turnover 13 .0578 .0033 Stockout % of SKUs 35 .1368 .0187 Bus inventory dolIars/vehicle 59 .0002 .0000 Rail inventory dodars/vehicle 12 -.2695 .0726 % items out of balance 47 -.0654 .0043 Bus % fill rate 46 .0756 .0057 Rail % fill rate 10 .3946 .1557 Bus % obsolete items 42 .0333 .1100 Rail % obsolete items 9 .5783 .3344 Bus days to fill backorders 42 -.0995 .0099 Rail days to fill backorders 6 -.2506 .0628 Inventory dollars per person 59 .2368 .0561 Person dolIars/Invento~y dollars 38 .0629 .0040 Transactions per person 52 .0599 .0036 The above correlation coefficients reveal that there is no significant correlation between agency operating cost and inventory perfonnance indicators. 4.3.3.4 Material Purchases by Mode Bus material purchases and rail material purchases were correlated with bus and rail inventory ~ . perfonnance Indicators with the toDow~ng results: Number of Correlation Coefficient of Variable ResponseS(n! Coefflcienttr! Deterrninationfr Bus inventory turnover 44 .2912 .0848 Bus inventory dolIars/vehicle 52 .0043 .0000 Bus % fin rate 40 .1 157 .0134 Bus % obsolete items 37 .0150 .0002 Bus days to fig backorders 35 -.0687 .0047 41

OCR for page 27
Rail inventory turnover 10 .083 ~ Rail inventory dollars/vehicle ~ ~.01 19 Rail % fill rate 7 .2293 Rail % obsolete items 7 .9237 Rail days to fill backorders 4 -.421 ~ .0069 .0001 .0526 .8532 .1773 Only one inventory perfonnance indicator, the percent of obsolete items for rail inventory, has a significant correlation with material purchases (r = .92~. The small sample size of seven reduces the significance of this isolated result. (Removing one data point would reduce r to .74.) There is no consistent pattern of correlation between material purchases and inventory performance indicators. 4.3.3.5 Annual Passenger Miles by Mode Selected "Section 15" data was gathered using the survey for reference and comparison purposes. The data most likely to affect inventory performance was the annual passenger miles. The bus and rail annual passenger miles was correlated with bus and rail inventory performance indicators with the following results: Number of Variable Responses (n) Correlation Coefficient of Coefficient (r) Determination (r2) Bus Inventory turnover 42 .0697 .0049 Bus inventory doDars/vehicle 49 .2534 .0642 Bus % Fill rate 37 . ~ 198 .0143 Bus % obsolete items 3S .2840 .0806 Bus days to fill backorders 36 -.OSS3 .0078 Rail inventory turnover ~ ~-.0353 .0012 Rail inventory dollars/vehicle 12 -.2356 .0555 Rail % fill rate ~.3313 .1098 Rail % obsolete items ~.6682 .4465 Rail days to fib backorders ' 5 -.2819 .0795 The only indicator showing a correlation of any size is the percent of obsolete items for rail inventory (r = .66). As with material purchases, this small sample size reduces the significance ofthis result. Based on the above, there is no significant pattern of correlation between annual passenger miles and inventor performance indicators. 4.3.3.6 Summary of Agency Profile Effects on Inventory Performance Indicators The analysis of the effects of agency profile data (survey Section I) against the inventory performance indicators shows only scant isolated effects with no consistent patterns. Only 22 out of 42

OCR for page 27
585 l-tests resulted in statistically significant differences in inventory performance indicators, and only 2 out of 35 correlations resulted in high correlation coefficients (.66 and .92). Furthermore, the significant differences appear random and may be the result of chance. For example, at the 90% confidence level used for the t-tests, one could expect that as many as 10% of the tests (58) could randomly result in significant differences. Therefore' agency profile characteristics have no significant effect on inventory performance indicators. 4.3.4 Effect of Fleet Profile Characteristics on Inventory Performance Indicators The following fleet profile characteristics from survey Section ~ (Fleet Profile) were analyzed: Service Mode (Bus versus Rail) Fleet Size ~ Number of Vehicles % of Fleet that is Manufactured in the US Number of Different Vehicle Models Average Vehicle Age -- Years Average Vehicle Age ~ % of Expected Life Average Annual Miles 4.3.4. 1 Differences Between Bus and Rail (Mode! The average value of bus and rail inventory performance indicators were compared using t tests with the following results: Confidence Level Perfonnance Indicator Bus Rail for Difference Annualinventoryn~nover 1.74 0.71 99.998% Inventory dollars per vehicle $5,027 $37,497 99.1% /0 fin rate 89.0% 86.1% 35.3% %obsoleteinventory 9.2% 6.1% 74.8% Days to fig backorders 16.4 2S.3 64.C9/o As the results show, transit agencies carry significantly more inventory per vehicle to support rail service than bus service. In addition, rail inventory is turned over significantly fewer times per year than bus inventory. The differences in other inventory performance indicators are not statistically significant at the 90% confidence level. However, the confidence levels are also not low enough to conclude that there is not a difference. 43

OCR for page 27
4.3.4.2 The Effect of Fleet Size -- Number of Vehicles Correlations were performed between the number of vehicles and inventory performance indicators. Those indicators relating to bus and rail were correlated with the number of buses and rail vehicles, respectively. The results were as follows: Number of Correlation Coefficient of Variable Responses (n! Coefficient fr) Determination (r2) Bus inventory turnover 55 .1533 .0235 Bus Inventory doLar~ve~cle 66 .1231 .0152 Bus TO fig rate 52 .1182 .0140 Bus TO obsolete items 48 .2494 .0622 Bus days to fin backorders 48 -.0897 .0081 Rod inventory turnover 1 1 -.0557 .0031 Red Inventory doDar~veWcle 13 -.3015 .0909 Rod /0 fig rate 9 .5429 .2948 Rod /0 obsolete items 9 .5477 .3000 Rod days to Backorders 6 -.3401 .1157 Stockout % of SKUs 40 . ~ 117 .0125 /0 items out of balance 50 -.0349 .0012 Inventory dollars per person 65 .2385 .0569 Person doDar~r~nventory dollars 40 .0534 .0029 Transactions per person 61 .1422 .0202 The above results show that there is no significant correlation between the number of vehicles and inventory performance indicators. To test this conclusion further, separate correlations were run between inventory performance indicators and subsets of the survey respondents based on the four categories defined by number of vehicles: Small Properties Medium Properties Large Properties Very Large Properties 50 or fewer vehicles (3 ~ survey respondents) 51 - 300 vehicles (33 survey respondents) 301-2000 vehicles (IS survey respondents) over 2000 vehicles (7 survey respondents) In addition to correlations, l-tests were performed on the average values of inventory performance indicators for each of these categories. The results were the same within these categories as for the entire survey sample. There is no significant correlation between the number of vehicles and inventory performance indicators within any of the above size categories. Furthermore, there is no pattern of statistically significant differences between average inventory performance indicator values for the above groups. 44

OCR for page 27
4.3.4.3 The Effect ofthe TO of Fleet that is Manufactured in the US Correlations were performed between the percent of the agency's fleet that was manufactured in the United States and the inventory performance indicators. Those indicators relating to bus and rail were correlated with the percent of US manufactured buses and rail vehicles, respectively. All correlation coefficients are below .35. Correlations were also performed for the four categories based on number of vehicles with similar results. There is no significant correlation between the percent of US manufactured vehicles and the inventory perfonnance indicators. 4.3.4.4 The Elect ofthe Number of Different Vehicle Models Correlations were performed between the number of different vehicle models in an agency's Beet and the inventory performance indicators. In addition, l-tests were run for the average value of the indicators for each value of the number of vehicle models. Most of the correlation coefficients were very low. Only two were over .35; the percent fill rate (.47) and the dollars per vehicle (.56) for rail inventory. These coefficients are not high enough to be significant. In addition, the l-tests did not show a statistically significant difference at the 90% confidence level between the number of models and the inventory performance indicators. Therefore, there is no statistically significant relationship between the number of vehicle models in an agency's fleet and the inventory performance indicators. 4.3.4.5 The Effect of Average Vehicle Age The effect of the average age of a transit agency's fleet was examined separately for two indicators of age; average years old and percent of expected life expended. In both cases, there were no significant correlations between the average age of an agency's fleet and the inventory perfommance indicators. 4.3.4.6 Average Annual Miles Average annual miles was the final fleet profile characteristic that was examined. As with the others, there was no significant correlation between average annual miles and inventory performance Indicators. 4.3.4.7 Summary of Fleet Profile Effects The results of the analysis of fleet profile characteristics (survey Section ~ against inventory performance indicators are that fleet profile characteristics, except for mode, have no statistically significant effects on inventory performance indicators. The difference between the bus and rail inventory performance indicators for turnover and dollars per vehicle are significant at the 99% confidence level. 45

OCR for page 27
4.3.5 Benchmark Decision Modeling Guide The development of a benchmark decision modeling guide was contingent upon isolating the relationships between agency and fleet profile characteristics and the inventory performance indicators. The goal was to develop different benchmark values based on significant relationships. For example, if inventory turnover were significantly correlated with agency operating cost, the benchmark value for Hanover would depend on the agency's operating cost. However, as the analyses summarized in this section show, there were no significant effects between an agency's characteristics and fleet profile and the inventory performance indicators. Therefore, there is no need for a benchmark decision modeling guide. The differences that were found relating to mode (bus and rail) win be accommodated by using different benchmark values for bus and rail. 4.4 BENCHMARK VALUES FOR PERFORMANCE INDICATORS The benchmark values for the inventory performance indicators are based on the survey responses. Rather than using the average value alone as the benchmark value, the following are presented for each inventory performance indicator: Mean - the average value Median-the middle value (equal number of respondents above and below) Maximum-the highest value Minimum-the lowest value 20th percentile-the value greater than 20% of the responses SOth percentile ~ the value greater than 80/0 ofthe responses ~ Performance Indicator. ~ ~ ~..~ Crimean ~ ~:~ Median ~ Max i. ~ ~ ~. Min Act: ~ ~ ~: 20% :. : ~ ~ .~ 80% ~ : i- ~ _ . . . . . . _ . . .. : Bus inventory turnover 1.74 1 1.43 7.36 0.13 0.75 2.54 ~: PI inventory turnover 0.71 ~0.56 1.43 0.29 0.51 0.99 Stockout % Of SKUs 1.52% 1 0.17% 20.0% .013% .047% 1.54% 1 1 Bus inventory Stvehicle S5,027 S4,604 S 15,384 S281 S2,566 S7,234 ~1 PI inventory S/veWcle S37,498 S27,418 $ 139,286 S6,785 S 12,660 S47,688 1 /Oitems out ofbalance ~783% ~5.0% ~: 60% 11 005% 1~ 1.42% H 10% us%fillrate 11 89.0% nH 95.0% 11 100% . 10% r 85% T 98% 1 1 . Rai1%fillrate 1 86.1% 1 90.2% 1~ 100% 1 40% 1~ 84.4% 1 98.3% L Bus % obsolete items 9.2% 5% 60% .01% 2% 13.8% 1 . . Rail%obsoleteitems H 6.1% H 5% 11 20% 11 1/o 1 1.12% ~10% 46

OCR for page 27
~ , :. . . : . . , . . . . . 1 . ; . ; .: 1 .. :; . .; :: . : . : . ~ Penance indicator: ~ ~ 1l ~ ~Mea:n~-~1l~: ~Median:: 1l : ~ AM ~ ~i ~Mimi ~:~-1t 20%:~ ~- ~1l - -801% --of Bus days to fill backorder 16.4 10 90 | 1 3 30 ~ ~ Rail days to fill backorder ~25.3 ~18 ~56 ~1 ~14 ~45 Inventory S per person 11 S217,980 ~ ~S146,000 11 1,300,000 11 S32,418 11 S84,302 11 S250,5?8 Person S/Inventory S 11 SO.31 11 S0.25 11 S1.05 11 S0.05 ~SO.15 11 S0.44 ~ ~798 _ 11 5 11 61.1 11 225.6 ! 4.5 INVENTORY PERFORMANCE INDICATORS IN OTlIER INDUSTRY S The following table shows innovation on inventor perfonnance indicators in other industries as compared to transit. This information was extracted Dom the articles included in the literature search conducted at the beginning ofthis study. Electnc Utility Fleet Spare Parts Beverage Distnbutor Fleet Spare Parts -- All Beverage Distributor Fleet Spare Parts ~ Sof`Drink Beverage Distnbutor Fleet Spare Parts-Beer Retail Disinbution Center Product Inventor American Airlines Public Transit-Bus Public Transit-Rail Fleet Spare Parts Inventory (mixture of all fleet types) I~25 vehicles Fleet Spare Parts Inventory (mixture of all fleet types) 2~99 vehicles Fleet Spare Parts Inventory (mixture of ad! fleet types) 100 or more vehicles Annual Inventory Tumover 3.3 6.1 7.8 3.4 52 1.74 0.71 2.7 Inventory Doldars per Vehicle $15,793 $13,271 $22,710 $1,800,000 $5,027 $37,498 $26,040 -$65,100 5.1 $10,010 -$38,1 15 5.7 $37,330 Maximum The above chart compares public transit bus and rail annual inventory turnover and inventory dollars per vehicle to the same indicators in other industries. The "fleet spare parts" data is from a survey of Beet managers who manage a variety of public and private sector fleets. The annual turnover for a 47

OCR for page 27
ram dl~u110n Operas pawn Memos is not compile 10 Spay pans Memos but Is Include far Salon. ~-se, 1be down per veDlcle far ~edc~ Edges Is Include. In the new chapter me discuss the rel~lons~ps between pe~=ce measures Ad o~a~z~loni1 profiles. Included are discussions of the appropd~e thresholds far mode o~s~=lonal Lectures or Hang other Reps to lucrease account~dby Ad contact 48