Break-Even Costs for Traditional Versus Quick Response Apparel Suppliers by A.D. Pinnow and R.E. King Using the Sourcing Simulator (aka ARMS) the break-even cost for a replenished product with a higher wholesale price can be compared to a traditional one-shipment product at a lower wholesale price where the resulting retail gross margin or GMROI are the same. Read full report below... March 2004

www.techexchange.com

Break-even Costs for Traditional versus Quick Response Apparel Suppliers

A.D. Pinnow R.E. King

NCSU-IE Technical Report #97-4 Industrial Engineering Department North Carolina State University

June 23, 1997

www.techexchange.com

1. Introduction Wholesale cost has been the primary concern for merchandise buyers for apparel retailers when sourcing decisions are made. This has led to higher and higher levels of imports due to lower labor costs. Typically for seasonal garments, offshore sourcing has taken the form of one delivery prior to the start of the selling season with the volume and assortment of goods based upon the retail buyer’s plan. Several analyses comparing the performance of traditional and Quick Response (QR) sourcing strategies have been undertaken, each of which have shown the superiority of a QR vendor for in-season inventory replenishment in terms of performance measures such as GMROI, inventory turns, service level, lost sales percent, and joboff percent over a traditional (often offshore) supplier [1, 2, 3, 4, 5, 6]. However, because traditional or off-shore vendors typically charge lower wholesale costs, they still remain attractive to buyers. We illustrate that higher prices may be paid to QR vendors to yield the same performance obtained from purchasing traditionally at a lower cost1. We focus on break-even analysis - specifically, the percent change in wholesale cost over that of traditional at which a garment may be purchased from a QR vendor and still provide the same performance. We consider two measures of performance: Gross Margin and GMROI, the latter being the Gross Margin divided by the average dollar investment in inventory. Gilreath et al [1] used a simple economic calculation to show the advantage of in-season drops over 100% up-front loading of the merchandise in terms of inventory turns. This led to the observation that a retailer can afford to pay from 30 to 50% more to a vendor that provides inseason replenishment and yield an equivalent GMROI. King and Hunter [5] using ARMS 2, a detailed retail store simulation model, found similar break-even costs in terms of GMROI for a QR vs. a traditional supplier. They also showed that even when using a simple measure like the Gross Margin to Sales Revenue ratio, a retailer can still afford to pay more to a QR vendor. In this paper we expand the analysis of King and Hunter [5] and explore break-even wholesale cost points between QR cases under varying lead times and corresponding traditional 1

Quick Response sources allow reordering and delivery throughout the selling season. With a QR vendor reorders are based on re-estimations of demand at the SKU level from point of sale data. Traditional sources, however, require that the buyer purchase exactly what is specified in the buyer’s plan developed prior to the selling season. 2 Apparel Retail Modeling System  North Carolina State University, 1997.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 1

cases where 100% of the buyer’s plan is received at the beginning of the season. In addition, we consider a traditional vendor with in-season drops of the buyer’s plan. The impact of markup level, joboff revenue (residual inventory salvage value), season length, and demand volume per SKU is also examined. 2. Methodology The ARMS Seasonal model developed at North Carolina State University was used to simulate all of the scenarios considered. This model stochastically simulates the selling season of a line of apparel given a set of inputs describing the characteristics of the line of goods and the merchandising plan, and provides standard performance measures. It is specifically designed to determine the impact of a specified level of forecast error in terms of overall volume and SKU (Stock Keeping Unit) mix error. Customers arrive to the store according to a non-stationary Poisson process and select garments according to the underlying SKU demand preferences. If the selected item is in stock a sale is recorded, otherwise, a stockout is recorded and an alternative choice may be selected based upon a model of consumer behavior. Depending upon the type of vendor, replenishment stock may be shipped to the store during the season. Under QR, the quantity and mix of replenishment stock is based upon a re-estimation of consumer demand using point of sale (POS) data. Details of the modeling assumptions and operation of ARMS is described extensively in the literature [2, 3, 6]. Each scenario is replicated 25 times to account for variability, and the results reported are an average of the replications for a given scenario. The standard deviation of the mean is less than 0.5% of the mean of the 25 replications for both Gross Margin and GMROI. The breakeven wholesale cost for the QR cases relative to various retail markup percentages3 and joboff prices at different levels of plan volume error and SKU mix error are calculated using Gross Margin and GMROI as the output measures.

3

Retail markup is expressed as a percent of wholesale cost. The retail price is equal to the wholesale cost plus the product of the markup percentage and the wholesale cost. For example, a garment with a wholesale cost of $12.50 with a 120% markup would sell for $12.50 + (1.2)($12.50) = $27.50.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 2

Plan volume error refers to the amount by which the buyer is wrong in the estimation of the total season demand across all SKUs in the line. For example, if the planned number of units to sell were 800 and the actual demand turned out to be 1000, the volume error would be -20%, indicating that the buyer underestimated the total demand by 20%. The SKU mix error represents the total error associated with the estimation of the style, color, and size distribution within the buyer’s plan [4]. It is arbitrarily defined as the sum of the absolute differences between the planned and actual style, color, or size percentages. A mix error of 20/20/10 indicates a 20% total style error, 20% total color error, and 10% total size error. For example, suppose there are 4 styles with a planned distribution of 10%, 25%, 35%, and 30% respectively. Suppose the actual demand over the season turns out to be 15%, 22%, 28%, and 35% for each style respectively. The absolute errors for each style then are |10% - 15%| = 5%, |25% - 22%| = 3%, |35% - 28%| = 7%, and |30% - 35%| = 5%. This sums to a total style error of 5% + 3% + 7% + 5% = 20%. It has been shown that the distribution of the error across each individual style, color, or size has little significance in overall retail performance; rather it is the total error that matters [6]. All scenarios considered cover a 20-week selling season where the consumer demand follows the mid-peak seasonality pattern shown in Figure 1. The total actual demand for the season in all scenarios was 4800 units, and a total of 24 SKUs were included in the buyer’s plan. An individual SKU represents a particular style, color, and size combination. For this analysis the 24 SKUs are comprised of the possible combinations of 2 styles, 3 colors, and 4 sizes. The buyer’s planned percent distribution of sales by style, color, and size is shown in Table 1.

1 2 3 4 Style 60% 40% Color 45% 30% 25% Size 10% 25% 35% 30% Table 1. Buyer’s Planned Percent Distribution of Styles, Colors, and Sizes.

For the QR cases the minimum order quantity was set at 1 per SKU. For any given lead time, orders were made beginning at the end of the first week and placed at the end of each week thereafter such that the receipt of the last order would be two weeks prior to the end of the season. Cost parameters for the model were held constant for all runs. Although the wholesale

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 3

cost was set at $12.50 and the garment sold for $25.00 retail (100% markup) with no joboff value at the end of the season, these parameters are varied in the analysis of the results. No markdowns were made, performance-based or otherwise, and no price premiums were included. The various scenarios considered included QR cases with order lead times varying from 1 to 8 weeks in increments of 1, plan volume errors from -30% to +30% in increments of 10%, and SKU mix errors of 0/0/0, 20/20/10 and 40/40/20. A parallel traditional case with 100% of the buyer’s plan received at the beginning of the season and no reorders was run for each volume and SKU error combination to compare to the QR cases under the varying lead times. 2.1 Initial Inventory The traditional case is representative of a typical off-shore scenario where the entire buyer’s plan is purchased and shipped ahead of the selling season. Therefore, the initial inventory (initial delivery percentage) was set at 100%. For the QR scenarios it was desired to set the initial inventory level such that the service level, defined as the percentage of customers who find their first choice garment, would be 95% over the weeks prior to receipt of the first reorder. Setting the initial inventory in this manner biases the results toward the traditional cases where 100% of the inventory is received up front and all customer demand is intended to be satisfied It is assumed that the demand in any given week follows a Poisson distribution, thereby allowing the initial inventory required to be calculated directly by a convolution of the presumed demand over the lead time weeks using the Poisson density function to calculate probabilities. For a given SKU, the target inventory level is set to k where k is smallest integer such that k

∑

i=0

e− λ

λk ≥ 0.95, k!

where λis the arrival rate over the period of interest according to the preseason plan. Based on the summed SKU-level requirements for the 95% target service level, the values in Table 2 show the total inventories used (in terms of percent of total forecast volume) and the resulting service levels over the supply lead time weeks. Notice that because of integrality the values are not exactly 95%.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 4

Lead time 1 2 3 4 5 6 7 8

3360 Inv. SL % % 10.0 95.2 14.3 94.9 19.0 94.9 23.8 95.0 28.9 95.0 34.2 95.2 39.5 95.0 45.1 95.0

3840 Inv. SL % % 9.4 94.9 13.9 95.0 18.5 94.9 23.4 95.1 28.4 95.0 33.6 95.0 39.0 95.0 44.5 95.1

4320 Inv. SL % % 9.3 95.1 13.6 95.0 18.3 95.1 23.4 95.0 28.5 95.0 33.7 94.9 39.0 94.9 44.5 95.0

Forecast Volume 4800 Inv. SL % % 9.3 95.0 13.6 95.2 18.1 95.1 23.0 94.9 28.1 95.0 33.3 95.0 38.7 95.0 44.2 94.9

5280 Inv. SL % % 9.1 95.1 13.6 95.0 18.0 95.0 22.9 95.0 28.0 95.1 33.3 95.1 38.8 95.0 44.1 95.0

5760 Inv. SL % % 9.1 95.1 13.4 94.9 18.2 95.1 23.0 95.1 28.0 95.0 33.2 95.0 38.5 95.0 44.1 95.1

6240 Inv. SL % % 8.9 95.0 13.1 94.9 17.6 94.8 22.4 95.0 27.4 95.0 32.8 95.0 38.2 94.9 43.7 94.9

Table 2. Initial Inventory Requirements (percent of total plan volume).

2.2 Break-even cost equations For Gross Margin and GMROI, equations are derived for calculation of break-even wholesale costs at varying retail and joboff prices. Gross Margin (GM) is defined to be the difference between total revenue (sales and joboff) and the cost of goods. Specifically,

GM = PS + (U − S ) J − WU = ( P − J ) S + ( J − W )U where J = job off price, P = retail price, S = units sold, U = units purchased, W = wholesale cost.

The derivation of the equation for calculation of the break-even wholesale cost based on GM is as follows. GM Trad = GM QR GM Trad = PS QR + J (U QR − S QR ) − WQR U QR GM Trad = ( P − J ) S QR + ( J − WQR )U QR WQR =

PS QR + J (U QR − S QR ) − GM Trad U QR

If the joboff price is zero, this simplifies to the following.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 5

WQR =

PS QR − GM Trad U QR

GMROI, Gross Margin Return on Inventory, is defined as the ratio of the Gross Margin to the product of the wholesale cost and average inventory carried over the course of the season, i.e.,

GM IW ( P − J ) S + ( J − W )U = IW

GMROI =

where

I = average inventory. Break-even wholesale cost calculation from the simulation runs is derived as follows.

GMROI Trad = GMROI QR GMROI Trad =

( P − J ) S QR + ( J − WQR )U QR I QR WQR

GMROI Trad I QRWQR = ( P − J ) S QR + ( J − WQR )U QR GMROI Trad I QR WQR + U QRWQR = ( P − J ) S QR + ( J − WQR )U QR WQR =

( P − J ) S QR + JU QR GMROI Trad I QR + U QR

This simplifies to the following for joboff price equal to zero. PS QR WQR = GMROI Trad I QR + U QR In the analysis that follows the results reported and illustrated are the percent change in wholesale cost that may be paid to a QR source over a traditional source. This percent change is calculated by subtracting the traditional wholesale cost from the corresponding QR break-even wholesale cost, then dividing by the traditional wholesale cost and multiplying by 100% to obtain a percent, i.e., D=

WQR − WTrad WTrad

×100%

where D = percent change in wholesale cost between a traditional and QR vendor.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 6

A negative percent change represents a situation where the QR wholesale cost needs to be lower than the traditional wholesale cost to achieve the same value of either GM or GMROI, and a positive percent change indicates that the QR case can demand a greater wholesale cost than the traditional for the same resulting GM or GMROI. 3. Break-even Analysis using Gross Margin 3.1 Effect of volume error and varying markup percentages For the following it is assumed that the joboff revenue is $0 and the SKU mix error is 40/40/20. This level of mix error has been discussed extensively with retailers and is believed to be representative (if not conservative) for that of seasonal merchandise. For any level of volume error, and even when the overall demand volume is predicted with 100% accuracy, the retailer can purchase QR items at a significantly higher price than traditionally supplied items. The percent increase in wholesale cost for QR-supplied garments ranges from a low of 20% (8 week lead time, total demand volume underestimated by 30%, and 80% retail markup) to a high of 52% (1 week lead time, 30% overestimation in demand volume, and 120% retail markup). Figure 2 shows the percent change in wholesale cost at varying lead times and levels of overall demand volume error for three retail markup percentages. The percent decreases with increasing lead time for a constant volume error. This illustrates the benefit of shorter reorder lead times. The power of using demand re-estimation in placing reorders during a selling season is also demonstrated. In all cases, the retailer was able to sell more garments with the QR supplier. Since the QR reordering strategy is based upon actual sales, when the buyer’s plan is in error either in overall volume, plan mix, or both, the buyer is able to capture this error and correct it. As orders are placed, adjustments can be made in the SKU mix and overall volume, allowing more demand to be satisfied. When purchasing from a non-flexible, traditional supplier, performance depends upon the buyer’s ability to forecast demand at the SKU level ahead of the selling season. As shown in Figure 2, the percent increase in the QR break-even price increases with an increase in retail markup percentage. Since more garments are sold compared to the traditional case, the revenue earned under QR increases with an increased profit per garment.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 7

With a QR supplier the ability to re-estimate demand and place orders accordingly during the season is reflected in the overall service level and percent jobbed off for the season. The overall service level is greater than the service level prior to the receipt of the first reorder. Thus, as the season progresses what is being purchased and stocked corresponds more closely to actual customer demand. The joboff percentages similarly show that QR is better at estimating customer demand. Under all of the QR cases, regardless of the volume error, the percent jobbed off was much less than traditional, as shown in Table 3.

Traditional

Quick Response

Lead Time 1 2 3 4 5 6 7 8

Volume Error -30% -20% -10% 0% +10% 10.4 14.7 19.0 22.6 26.3 0.5 0.5 0.5 0.6 0.6 0.5 0.6 0.6 0.7 0.7 0.6 0.7 0.7 0.8 0.9 0.8 0.9 0.9 1.0 1.1 1.1 1.3 1.4 1.5 1.6 2.5 2.5 2.6 2.6 2.8 1.5 1.8 2.1 2.5 3.0 2.8 3.1 3.5 3.9 4.8 Table 3. Percent jobbed off.

+20% 29.9 0.6 0.8 1.0 1.2 1.7 3.1 3.6 5.7

+30% 33.1 0.6 0.8 1.0 1.3 2.0 3.4 4.3 6.7

In the traditional case, the greater the volume error the greater the amount of residual inventory. Even when overall season demand was accurately estimated (0% Volume Error), 22.6% of the offering was jobbed off due to the SKU mix error. When 30% more than is needed is ordered (+30%), the job off percent is closer to the known volume error (3.1% higher), but it hardly seems reasonable to make it a practice to order 30% more than is expected to sell simply to try to capture possible plan mix error, especially in a situation where there is no salvage value for the garments. Under QR the percentage of stock jobbed off is relatively stable across all volume errors, increasing only slightly with an increase in volume error. This increase is more pronounced with longer lead times, which is explained by the larger initial inventory stocked in those cases. This gives further evidence to the strength of demand re-estimation, as even with a demand volume

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 8

error of +30% compounded with the SKU mix error and a long lead time (8 weeks), the percent jobbed off under QR is only 6.7%.

3.2 Traditional sourcing with in-season drops In terms of Gross Margin, it is intuitive that stocking 100% up front is the best that can be done with a traditional vendor. Although allowing in-season drops with a traditional supplier will improve the inventory investment, the Gross Margin is at its highest when the entire buyer’s plan is stocked at the start of the selling season.

3.3 Effect of job off revenue In order to assess changes in the break-even wholesale cost with changes in joboff prices, results were analyzed by varying the joboff price from $0 to $12.50 (the traditional wholesale cost) in increments of $2.50. The retail price was set constant at $25.00 (100% markup). When the traditional case and the QR cases with varying lead times are compared with no salvage value for the garments, the results are biased toward QR because the traditional case is receiving the maximum penalty for any overstocking. The other extreme, where the results are biased toward the traditional source, is when the joboff price is equal to the wholesale cost, i.e., there is no penalty for residual stock. If the retailer can recover the entire cost of the garment, essentially nothing is lost by overstocking4. As shown in Figure 3, for any joboff price, the price paid to a QR source may be greater than the price paid to a traditional source regardless of lead time. As the volume error increases, the differences between the break-even wholesale cost at the different joboff prices increases. This is due to the increasing percentage of garments jobbed off in the traditional case, as previously shown in Table 3. When demand is underestimated, the percent jobbed off in the traditional case is relatively low, and therefore changing the amount received for the remaining garments does not impact the results significantly. When demand is overestimated, however, the amount that can be received for the garments remaining at the end of the season has a greater impact on the results. As the joboff price increases, the traditional case has an opportunity to generate more revenue because those garments remaining at the end of the season can be sold for 4

In this case we are neglecting the cost of capital.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 9

a greater price. In fact, when demand is overestimated by 30%, to yield the same Gross Margin there is only about a 5% difference between the QR and traditional wholesale costs when the joboff price is equal to the traditional wholesale cost. However, the traditional case is generating revenue by jobbing off 33% of the garments ordered at the end of the season. The QR cases, on the other hand, are generating revenue by selling garments to customers.

3.4 Effect of SKU mix error So far the discussion has centered on the situation where the SKU mix error is 40/40/20. This was the greatest level of error considered, and is representative of errors actually seen in practice as consumer demand for apparel can be difficult to predict. As the level of SKU mix error decreases, the price that can be paid to a QR source to yield the same Gross Margin decreases. However, only when there is no forecast error in the buyer’s plan does the traditional source compete with the QR scenarios. Figure 4 compares the three levels of SKU mix error considered at different joboff prices. In each case the retail is $25.00 (100% markup) and there is no error in the buyer’s estimation of the total season demand (no volume error). The maximum of 36.6% change occurs with a one week lead time and SKU mix error of 40/40/20. As the joboff price increases, the percent change decreases. However, for each of the cases with SKU mix error, more can be paid for the QR garment to achieve the same Gross Margin. In the cases where there is no SKU mix error the QR wholesale costs actually fall below the traditional wholesale cost for the same Gross Margin to be attained. This is explained by the initial inventory level. For QR, the initial inventory levels were set such that one would expect to satisfy only 95% of the demand over the weeks prior to the receipt of the first reorder. The QR cases are essentially set up to perform worse than a traditional case when demand is perfectly known. As the lead time increases, and thus the time of the receipt of the first reorder, the percent difference in the break-even wholesale cost increases, since a greater portion of the demand is being satisfied at only 95%.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 10

4. Break-even Analysis using GMROI 4.1 Results for all scenarios Gross Margin is inadequate in and of itself when making comparisons between sourcing strategies [1,5]. In this section, break-even wholesale costs are analyzed for a measure that also incorporates inventory investment: GMROI. GMROI (Gross Margin Return on Inventory) is the ratio of Gross Margin to the product of the Average Inventory and the wholesale cost per unit. It basically represents the amount of money made for every dollar tied up in inventory. For all scenarios the wholesale cost paid for a QR garment may be significantly higher than that paid for a traditionally-supplied garment to achieve the same GMROI. Figure 5 shows the percent change in wholesale cost at an SKU mix error of 40/40/20 and joboff price of $0 for demand volume errors of -30%, 0%, and 30% at varying retail markup percentages. When the volume is severely overestimated by the buyer (volume error of 30%), at a 100% retail markup the QR wholesale price can be up to 92.23% higher than the traditional cost to yield the same GMROI. For a traditional cost of $12.50, the corresponding QR cost would be $24.03. Because the retail price is $25.00, this means that the QR garment only needs to make a profit of $0.97 per garment to achieve the same GMROI. Even in the worst case (demand underestimated by 30% and lead time of 8 weeks) the QR break-even point is still 53.34% higher than the traditional cost for a retail markup percentage of 100%. The impact of considering inventory investment in break-even wholesale cost calculations is significant. Changes in joboff price do not have a significant impact on the percent change in wholesale cost paid between a QR or traditional supplier, although, as with Gross Margin, the impact of the joboff price increases with an increase in forecast error. Similarly, as the SKU mix error increases the impact of changes in the joboff price slightly increases. The maximum difference between joboff prices of $0 and $12.50 occurs when demand is overestimated by 30% and the QR lead time is 8 weeks. For SKU mix errors of 0/0/0, 20/20/10, and 40/40/20, the difference in the percent change in wholesale cost is 7.5%, 9.3%, and 12.7%, respectively. This difference increases with increasing SKU mix error because the percentage jobbed off in the traditional case increases, as previously discussed in relation to Gross Margin.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 11

For all scenarios the trend in percent change in wholesale cost for GMROI is the same as in Figure 5. To summarize results of a representative case, Table 4 presents the percent increase in wholesale cost that may be paid to a QR supplier with various lead times to achieve the same GMROI as with the traditional supplier where the joboff price is $7.50 (60% of wholesale) and the retail price is $25.00 (100% markup).

Volume Error 0%

-30% Lead Time

Quick Response

2 4 6 8

0/0/0

79.2 73.9 62.4 46.9 Table 4.

20/20/10

40/40/20

0/0/0

20/20/10

+30% 40/40/20

79.4 81.8 80.3 82.3 84.8 73.8 75.9 74.8 76.3 78.5 62.5 64.7 65.2 66.6 67.9 48.3 51.8 52.0 53.9 55.5 Percent Change in Wholesale Cost.

0/0/0

20/20/10

40/40/20

84.6 78.7 69.2 58.2

85.3 79.0 69.5 57.7

86.8 80.3 69.2 55.8

The reason that QR can demand a wholesale cost so much greater than the traditional supplier is because the traditional case is stocking 100% of the line up front, and therefore average inventories are much greater than the corresponding QR cases. Table 5 presents a comparison of inventory turns for traditional and QR cases. Under QR the retailer carries as little as 8% and at most 41% of the average inventory carried in the traditional case. As both the number of garments ordered increases and the SKU mix error increases, the inventory turns decrease for all scenarios. For the traditional case this is due to carrying a greater total volume and not being able to sell as much of it with an increasing SKU mix error. For the QR cases, the inventory turns similarly decrease because of less demand being satisfied prior to the receipt of the first reorder with an increasing level of SKU mix error. As the volume of garments ordered and the lead time increases, the number of garments initially stocked increases, meaning that a greater volume is carried for those weeks, which leads to decreasing inventory turns. Volume Error 0%

-30%

Traditional Quick Response

+30%

Lead Time

0/0/0

20/20/10

40/40/20

0/0/0

20/20/10

40/40/20

0/0/0

20/20/10

40/40/20

2 4 6 8

2.8 24.8 18.9 12.6 8.1

2.6 23.1 17.6 12.0 7.9

2.3 20.2 14.8 10.1 6.7

2.0 18.8 14.2 9.7 6.7

1.9 17.8 12.9 8.8 6.0

1.8 16.0 10.9 7.3 4.9

1.6 14.6 10.2 6.8 4.7

1.6 14.2 9.7 6.4 4.4

1.6 13.1 8.5 5.5 3.8

Table 5. Inventory Turns.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 12

4.2 Traditional vendor with in-season drops When in-season drops are used with a traditional supplier, the GMROI is improved, although it still does not compare to that of QR. Table 6 compares the GMROI for traditional cases where in-season drops are made to QR scenarios with the same delivery schedule. The results reported are for an SKU mix error of 40/40/20. Even though the GMROI value for inseason drops is significantly increased over that of 100% up front loading, the increase in GMROI break-even cost for QR still ranges from about 10% (8 week lead time, 30% overestimate in demand volume) to 75% (1 week lead time, 30% underestimate in demand volume).

Volume Error -30%

0%

+30%

Lead Time 2 4 6 8 2 4 6 8 2 4 6 8

Traditional, Traditional with 100% initial stock In-Season Drops 1.8 8.6 1.8 7.7 1.8 6.4 1.8 5.1 1.0 3.4 1.0 3.3 1.0 2.9 1.0 2.4 0.5 1.7 0.5 1.6 0.5 1.4 0.5 1.2 Table 6. GMROI.

Quick Response 20.0 14.6 9.6 6.3 15.8 10.7 6.9 4.5 12.8 8.3 5.1 3.3

4.3 Service Level A third measure of considerable interest to retailers is customer service. As stated in the Methodology section, initial inventories were set so that the results would be biased toward the traditional case, at least over the weeks prior to the receipt of the first reorder. Despite this bias, the QR cases yielded service levels comparable to, and usually better, than traditional sourcing. Because the initial inventory was set assuming that the buyer was not in error in either volume or SKU mix, when the scenarios were run with various levels of plan volume error and SKU mix error, the resulting service levels prior to the receipt of the first reorder were in general less than 95%. Table 7 presents the overall service level for the traditional case where 100% of the

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 13

inventory is received at the start of the season and for QR cases with varying lead times. In the traditional case the service levels are a result of both SKU mix and plan volume error, as well as end of season stockouts. The service levels in the QR cases, however, are affected more by the initial inventory percentages and the stockouts that occur during the weeks prior to the receipt of the first reorder.

Volume Error 0%

-30% Lead Time

Traditional Quick Response

2 4 6 8

0/0/0

20/20/10

40/40/20

0/0/0

20/20/10

69.0 66.2 58.3 96.6 86.2 82.6 80.1 75.0 91.9 88.7 79.6 77.6 72.4 90.9 87.4 78.7 77.2 73.1 90.8 87.4 79.6 77.8 72.3 91.6 87.7 Table 7. Overall Service Level (percent).

+30% 40/40/20

0/0/0

20/20/10

40/40/20

73.5 82.2 81.2 81.5 81.3

100.0 94.8 95.1 95.9 96.2

95.5 92.9 92.6 93.0 93.5

83.7 87.4 86.9 87.2 87.5

When demand is underestimated by the buyer, the service levels under QR are much better because additional garments are ordered and therefore more customer demand is satisfied. As the volume error increases, the overall service level increases regardless of vendor type. When the traditional service level surpasses that of QR it is largely due to 95% service level target prior to receipt of the first reorder. At any level of plan volume error, however, the QR service levels become increasingly better than traditional with an increase in SKU mix error. Again, because QR is able to reorder based on observed demand over the course of the selling season, it is able to correct for error and better serve customer demand. It is interesting to note that the trend in service level is not constantly decreasing as the QR lead time increases. Rather, it is nearly constant. This suggests that the service level is more dependent on the amount of error in the buyer’s plan than on the lead time before the first order is received. It should be noted that as the initial inventory level is increased, QR service levels improve avoiding potentially poor performance prior to the receipt of the first reorder [2, 3, 4, 5, 6]. (This is directly addresses in Section 7.) When combined with prior discussion it is evident that with QR a retailer can buy at a greater price, and achieve the same Gross Margin and/or GMROI with better customer service.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 14

4.4 GMROISL Instead of taking all of these measures independently, a measure has been suggested [7] which incorporates Gross Margin, average inventory, and in-stock percentage. It is GMROISL, and is the product of GMROI and the in-stock percentage. In this analysis, when break-even wholesale costs are calculated using GMROISL as the measure, the results do not significantly change from those obtained with GMROI as the measure of interest. The resulting break-even wholesale costs differ by at most about $0.50 between GMROI and GMROISL. The addition of the in-stock percentage as incorporated in this measure does not have a significant impact on the break-even wholesale cost between Quick Response and traditional suppliers. However, the advantage of using GMROISL as a performance measure, in general, is that since it explicitly includes the in-stock level, it penalizes the case when a high GMROI is obtained simply by maintaining very low inventories. Thus, GMROISL may be a better measure to judge buyer performance.

5. Effect of Season Length With shorter season lengths, a QR supplier may still demand a greater price than a traditional supplier to achieve the same Gross Margin or GMROI. Figure 6 shows the impact of season length on the percent change in wholesale cost at various lead times and forecast volume errors with Gross Margin as the measure of interest. The retail markup is 100%, the joboff price is equal to the wholesale cost, and the SKU error is 40/40/20. The initial inventory for each of the QR cases is set to achieve a 95% service level prior to the receipt of the first reorder as in the prior analysis. There is a decrease in the percent change in wholesale cost with a decrease in season length for each volume error, which is less pronounced at shorter lead times. When the forecast for the overall volume is overestimated by 30%, the percent change in wholesale cost falls to between -1% and -2% for the 2 and 4 week lead times at a season length of 8 weeks. As described in earlier sections, this is because the joboff price is equal to the wholesale and the initial inventory is set to satisfy 95% of the demand in the QR case.. However, even at a joboff price equal to the wholesale cost, the price paid to a QR supplier would only be 1 to 2 percent less than the price paid to the traditional source.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 15

Figure 7 illustrates the effect of season length on the percent change in wholesale cost using GMROI as the measure of interest. The percent change in wholesale cost for a 20 week season ranges from 75% to 87%, and for an 8 week season from 12.5% to 68%. In both cases the high occurs at a volume error of +30% and 1 week lead time, and the low occurs at a -30% volume error and 4 week lead time. Like with Gross Margin, as the season length decreases the percent more that may be paid to a QR vendor decreases. As the lead time increases and the season length decreases the percent change drops more significantly because the initial inventory carried increases. For an 8 week season and lead time of 4 weeks, the initial inventory covers the first 5 weeks, representing 62.5% of the season.

6. Effect of Weekly SKU Volume With a lower demand per SKU during a week, a QR supplier may still demand a greater price than a traditional supplier to achieve the same Gross Margin or GMROI. Thus far in the analysis a total of 24 SKUs were considered with an average weekly demand of 10 per SKU. As the number of SKUs in the line increases, the volume demanded per SKU per week decreases. Figures 8 and 9 show the percent change in wholesale cost for varying weekly demand volumes per SKU at various lead times and forecast volume errors for Gross Margin and GMROI, respectively. The retail markup is 100%, the joboff price is $12.50, and the SKU mix error is 40/40/20. The season length considered is 20 weeks. The initial inventory was set to satisfy a 95% target service level prior to the receipt of the first reorder. For each number of SKUs offered the initial inventory percent required to satisfy a 95% service level changes because of the necessity to consider a greater variability in customer demand with a greater number of SKUs. For both Gross Margin and GMROI, the percent change in wholesale cost that may be paid to a QR vendor is relatively constant across the weekly SKU demand volumes considered. For Gross Margin, the difference between the largest and smallest percent change in wholesale cost is within 5%. Similarly for GMROI, for weekly demand volumes per SKU between 3 and 10 the maximum difference in percent change is within 5%. There is a more significant change between demand volumes per SKU of 1 and 3 because of the larger inventory carried to satisfy a 95% target service level with a larger number of SKUs. However, for both Gross Margin and

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 16

GMROI, the QR supplier can demand a greater price at any weekly SKU volume to achieve the same performance as a traditional supplier.

7. Effect of Initial Inventory Percentage Previously the initial inventory percentage was set at the minimum level to attain a 95% service level, based upon the buyer’s pre-season estimate of demand over the weeks prior to the receipt of the first reorder. Often the required presentation stock, however, would be even larger. In this section the impact of increased initial inventory levels is examined. For order lead times of 1 and 8 weeks, break-even costs were determined for four initial inventory levels at demand volume errors of -30%, 0%, and +30% with an SKU error of 40/40/20. The retail markup is 100% and the joboff price is equal to the traditional wholesale cost of $12.50. Setting the joboff price at its highest possible level favors the traditional case especially in the case where the overall demand volume is overestimated by 30%. Tables 8 and 9 show the results for Gross Margin and GMROI for lead times of 1 and 8 weeks respectively.

Gross Margin GMROI Initial Inventory -30% 0% 30% -30% 0% 30% 10% 21.63 12.96 6.26 82.63 84.53 85.36 20% 32.54 20.51 11.89 72.79 74.32 74.84 30% 35.47 21.80 12.45 63.54 64.21 64.11 40% 36.21 21.93 12.45 54.62 54.42 53.14 Table 8. Percent Change in Wholesale Cost for Lead Time of 1 week.

Gross Margin GMROI Initial Inventory -30% 0% 30% -30% 0% 30% 40% 13.34 5.69 1.44 54.78 56.26 55.19 50% 19.89 11.05 5.57 47.84 47.36 44.00 60% 23.77 14.12 7.10 40.62 38.21 33.63 70% 26.66 15.42 8.06 33.64 29.06 24.53 Table 9. Percent Change in Wholesale Cost for Lead Time of 8 weeks.

As the initial inventory percentage increases, the percent more that may be paid to a QR vendor over a traditional vendor to achieve the same Gross Margin increases for both lead times. This is a result of having more stock on hand to satisfy the demand over the weeks of the season Break-even Wholesale Cost Analysis

www.techexchange.com

Page 17

prior to the receipt of the first reorder. Recall that when using a QR vendor, the retailer is able to order more of what is needed, so starting with a larger inventory level does not limit their ability to get more stock. The trade off, however, is the increase in average inventory which leads to a decrease in the percent change in wholesale cost when using GMROI as the measure of interest. For both lead times the percent change in wholesale cost for GMROI decreases by as much as 30%, indicating the significant impact that the initial stock level has on the average inventory for the season. However, even when 70% of the buyer’s plan is stocked up front with an order lead time of 8 weeks, the QR vendor can still charge up to 26.7% more than the traditional supplier to make the same Gross Margin and up to 33.6% to achieve the same GMROI.

8. Conclusions When considering different sourcing options for apparel, it is important to consider more than just wholesale cost. •

QR sources can demand a higher price than traditional sources and yield a better Gross Margin because more customer demand is satisfied and less garments are jobbed off. The greater the markup level and the less the residual value of inventory, the greater the QR advantage.

•

GMROI better captures the true benefit of using a QR supplier because it considers the dollar investment of inventory. When using GMROI as the measure of interest, breakeven costs are much greater than those obtained when considering Gross Margin.

•

Under any amount of forecast error, because a QR supplier enables the buyer to reorder based on observed demand, service levels and inventory turns improve.

•

Quick Response suppliers can still demand greater prices than traditional suppliers, even when the season lengths are shorter and weekly SKU demand volumes are smaller.

•

Even with an initial inventory as high as 70% with an 8 week lead time, more can be paid to the QR vendor to break-even on both Gross Margin and GMROI.

•

When comparing to importing, duty and higher shipping costs (not considered in this analysis) lead to an even greater advantage for domestic supply. This analysis has shown that making the choice is not as simple as going with the least

expensive supplier, as equivalent results may be obtained by purchasing at a significantly higher

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 18

cost from a supplier who allows replenishment based on re-estimations of demand throughout the course of the selling season. For Gross Margin, this cost may be as much as 50% more than what is charged by a traditional supplier and for GMROI, it may be as much as twice as great. Flexibility in inventory replenishment comes at a price - but it’s a price that retailers can more than afford to pay.

References [1] Gilreath, T.L., J.M. Reeve, and C.E. Whalen, Jr. (1995) “Time is Money: Understanding the Product Velocity Advantage,” Bobbin, March. [2] Hunter, N.A., R.E. King and H.L.W. Nuttle, (1992) “An Apparel Supply System for QR Retailing,” Journal of the Textiles Institute, 83, No. 3, pp.462-471. [3] Hunter, N.A., R.E. King and H.L.W. Nuttle (1996) “Evaluation of Traditional and Quick Response Retailing Procedures using a Stochastic Simulation Model,’Journal of the Textiles Institute, 87, No.1, Part 2, pp.42-55. [4] King, R.E. and N.A. Hunter, (1996) “The Impact of Size Distribution Forecast Error on Retail Performance,” NCSU-IE Technical Report 95-8, Industrial Engineering Department, N.C. State University, Raleigh, NC. [5] King, R.E. and N.A. Hunter, (1997) “Quick Response Beats Importing in Retail Sourcing Analysis,” Bobbin, March. [6] Nuttle, H.L.W., R.E. King and N.A. Hunter, (1991) “A Stochastic Model of the Apparel Retailing Process for Seasonal Apparel,” Journal of the Textiles Institute, 82, No. 2, pp. 247259. [7] Whalen, C.E., J. Friede, and S. Fitzpatrick, (1997) personal communication.

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 19

Figure 1: Seasonality

Mid-peak Seasonality Pattern 0.06

Fraction of Total Demand

0.05

0.04

0.03

0.02

0.01

0 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Week of Season

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 20

Figure 2: Impact of lead time, volume error, and retail markup percentage on Gross Margin Figure 2a Retail Markup = 80%

Percent Change

60 50 40 30 20 10 0 1 3 Lead Time (weeks)

5

-10 7

0

-20

10

20

30

Volume Error

-30

Figure 2b Retail Markup = 100%

Percent Change

60 50 40 30 20 10 0 1

3 Lead Time (weeks)

0 5 7

-20

10

20

30

-10 Volume Error

-30

Figure 2c Retail Markup = 120%

Percent Change

60 50 40 30 20 10 0 1

3 Lead Time (weeks)

Break-even Wholesale Cost Analysis

5

7

-20

-10

0

10

20

30

Volume Error

-30

www.techexchange.com

Page 21

Figure 3: Impact of lead time, volume error, and joboff price on Gross Margin Figure 3a Joboff Price = $0

Percent Change

60 50 40 30 20 10 0 1

3

Lead Time (weeks)

10

0 -10

5 7

20

30

Volume Error

-20 -30

Figure 3b Joboff Price = $7.50

Percent Change

60 50 40 30 20 10 0 1

3

Lead Time (weeks)

0

10

20

30

-10

5 7

Volume Error

-20 -30

Figure 3c Joboff Price = $12.50

Percent Change

60 50 40 30 20 10 0 1 Lead Time (weeks)

Break-even Wholesale Cost Analysis

0

3

-10

5 7

-20

10

20

30

Volume Error

-30

www.techexchange.com

Page 22

Figure 4: Impact of lead time, joboff price, and SKU mix error on Gross Margin Figure 4a SKU Mix Error 0/0/0

Percent Change

40 30 20 10 0 -10

12.50 1

3

Lead Time (weeks)

5

7

2.50

5.00

7.50

0

10.00

Joboff Price ($)

Figure 4b SKU Mix Error 20/20/10

Percent Change

40 30 20 10 0 -10 1

3

Lead Time (weeks)

5

7

2.50

5.00

10.00 7.50

0

Joboff Price ($)

Figure 4c SKU Mix Error 40/40/20

Percent Change

40 30 20 10 0 -10 1 Lead Time (weeks)

Break-even Wholesale Cost Analysis

3

5

7

2.50

5.00

10.00 7.50

0

www.techexchange.com

Joboff Price ($)

Page 23

Figure 5: Impact of lead time, volume error, and retail markup percentage on GMROI Figure 5a Retail markup = 80%

110 Percent Change

100 90 80 70 60 50 40 1

3

Lead Time (weeks)

5 7

-20

0

-10

20

10

30 Volume Error

-30

Figure 5b Retail markup = 100%

Percent Change

110 100 90 80 70 60 50 40 1

10

0

3

Lead Time (weeks)

Volume Error

-10

5 7

30

20

-20 -30

Percent Change

Figure 5c Retail markup = 120%

110 100 90 80 70 60 50 40 1

Lead Time (weeks)

Break-even Wholesale Cost Analysis

3

5

7 -30

-20

-10

0

www.techexchange.com

10

20

30 Volume Error

Page 24

Figure 6: Impact of season length on Gross Margin Figure 6a Demand Underestimated by 30% 25

Percent Change

20 15 LT = 1 10

LT = 2

5

LT = 4

0 8

10

12

14

16

18

20

-5 Season Length (weeks)

Figure 6b No Error in Demand Estimation 25

Percent Change

20 15

LT = 1

10

LT = 2

5

LT = 4

0 -5

8

10

12

14

16

18

20

Season Length (weeks)

Figure 6c Demand Overestimated by 30% 25

Percent Change

20 15

LT = 1

10

LT = 2 LT = 4

5 0 -5

8

10

12

14

16

18

20

Season Length (weeks)

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 25

Figure 7: Impact of season length on GMROI

Percent Change

Figure 7a Demand Underestimated by 30% 90 80 70 60 50

LT = 1 LT = 2

40 30 20 10 0

LT = 4

8

10

12

14

16

18

20

Season Length (weeks)

Percent Change

Figure 7b No Error in Demand Estimation 90 80 70 60 50 40 30 20 10 0

LT = 1 LT = 2 LT = 4

8

10

12

14

16

18

20

Season Length (weeks)

Percent Change

Figure 7c Demand Overestimated by 30% 90 80 70 60 50 40 30 20 10 0

LT = 1 LT = 2 LT = 4

8

10

12

14

16

18

20

Season Length (weeks)

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 26

Figure 8: Impact of weekly volume/ SKU on Gross Margin Figure 8a Demand Underestimated by 30%

Percent Change

30 25 LT = 1

20

LT = 4

15

LT = 8

10 5 0 1

2

3

4

5

6

7

8

9

10

Vol/sku per week

Figure 8b No Error in Demand Estimation

Percent Change

30 25 LT = 1

20

LT = 4

15

LT = 8

10 5 0 1

2

3

4

5

6

7

8

9

10

Vol/sku per week

Figure 8c Demand Overestimated by 30%

Percent Change

30 25 LT = 1

20

LT = 4

15

LT = 8

10 5 0 1

2

3

4

5

6

7

8

9

10

Vol/sku per week

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 27

Figure 9: Impact of weekly volume/ SKU on GMROI Figure 9a Demand Underestimated by 30%

Percent Change

100 80

LT = 1

60

LT = 4

40

LT = 8

20 0 1

2

3

4

5

6

7

8

9

10

Vol/sku per week

Figure 9b No Error in Demand Estimation

Percent Change

100 80

LT = 1

60

LT = 4

40

LT = 8

20 0 1

2

3

4

5

6

7

8

9

10

Vol/sku per week

Figure 9c Demand Overestimated by 30%

Percent Change

100 80

LT = 1

60

LT = 4

40

LT = 8

20 0 1

2

3

4

5

6

7

8

9

10

Vol/sku per week

Break-even Wholesale Cost Analysis

www.techexchange.com

Page 28