Hell hath no fury like a customer scorned! All OEMs know that any failure to support their customers during critical equipment breakdowns can have an enormous impact on the OEM’s reputation and future sales. OEMs that sell industrial equipment to corporate customers possibly know this better than most, but the service firms of these OEMs experience an enormous challenge called the “Long tail effect” due to low volumes of commissioned equipment in the market. For them, a large proportion of their spare parts sell infrequently and managing this long tail is key to managing the spares distribution environment. The challenge is that:<ul style="list-style-type:disc;"><li>A commitment to ensure uninterrupted service implies that any OEM service firm is duty-bound to stock even items that fail infrequently. On the other hand, stocking large numbers of slow-moving SKUs increases cash tied up in inventory.</li><li>With new models being launched at an increasing frequency, the number of spare part SKUs that require servicing increases. Consequently, an increasingly large range of spares needs to be stocked.</li></ul> To resolve this conflict, many organisations focus on lead time reduction at all nodes in the service distribution network. The assumption is that if lead times are reduced, the norms (minimum inventory) to be held at various nodes for every SKU will reduce, enabling stocking of an increased range of spares for the same working capital, leading to improved service. At first glance, this assumption seems correct, and so the procurement and distribution professionals of such service organisations are always trying to reduce lead times or redesign the distribution network to minimise transit times from one node to another. However, it has been seen that even if they succeed in reducing lead times, significant inventory reduction is not always achieved. <h5>Norm and its relationship with lead time</h5>The norm for an item stocked in a Make to Availability (MTA) environment is determined by the peak demand within the replenishment lead time.  Comparing demand within multiple lead time buckets in the past and identifying the peak sales figure attained within any one of the buckets gives us the norm. For example, if the reliable lead time of an item is 15 days, then sales figures within successive lead time buckets of 15 days in the past year are compared. Thereafter, the maximum demand in a 15-day bucket among them is designated as the norm for the item. Therefore: Norm = peak demand during a reliable replenishment period, factored for variability x reliable replenishment time.  Based on the above formula, norm and replenishment lead time are directly related, as depicted by this straight line graph depicted in Figure 1:<figure class="image"><img style="aspect-ratio:4000/3042;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-1_1757846234.jpg" width="4000" height="3042"></figure>But while this may be the case often, this is not always so! Let’s examine whether the reality conforms to our intuition. For this analysis, we at Vector Consulting Group conducted a detailed study of both slow-moving and fast-moving parts of a large material handling equipment company in India. Depicted below is a graph (Figure 2) showing the annual trend of norm computation with increasing lead times for a slow-moving spare part of this service firm. This slow-moving spare part has been sold only 9 times in the last year, translating to a sale once every 40 days on average (365/9).<figure class="image"><img style="aspect-ratio:4000/3694;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-2_1757846245.jpg" width="4000" height="3694"></figure>This graph (Figure 2) is quite different from expected and indicates that the norm, when computed for such slow-moving parts, is a stepped function of lead times. In other words, the norm remains constant within a certain interval of lead time and changes in value only when moving from one interval of the lead time to the next. In the case of a fast mover which has sold daily for the past one year, the graph of norm vs the lead time for such an item looks like Figure 3:<figure class="image"><img style="aspect-ratio:4000/3623;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-3_1757846254.jpg" width="4000" height="3623"></figure>This graph conforms exactly to the initial intuition about the relationship between norm and lead time. Therefore, for fast movers, any reduction in lead time results in a reduction in norm, and this can be the right action to be taken for firms selling fast-moving goods like soaps and chocolate. But an initiative to broadly reduce lead times of items in a spare parts scenario irrespective, of their demand profile, may not be as successful in reducing inventory levels! <h5>Norm behaviour in a slow-moving parts environment</h5>It is clear that we need to delve deeper into the reasons for the norm graph appearing and behaving the way it does for environments with slow-moving parts. For this, let’s examine an item with a simplified demand pattern as depicted in the Table 1. Suppose the item has sold 4 times in the past month, the first sale being on 10th, the second sale after 5 days (on the 15th), the third sale after an interval of another 10 days (on 25th) and the fourth sale after another 5 days, on the 30th of the month. The minimum interval between two consecutive sales, as seen from the demand table is 5 days. Let’s assume for simplicity’s sake that the demand pattern of the item is repetitive, month on month..<figure class="image"><img style="aspect-ratio:4000/3016;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-img-4-1736130245_1776751904.jpg" width="4000" height="3016"></figure><figure class="image"><img style="aspect-ratio:4000/2280;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-4_1757846269.jpg" width="4000" height="2280"></figure>The tabular representation of the norm vs lead time graph is depicted in Table 2. Therefore, if this item were to be kept under MTA and its norm was computed based on past month’s sales data, it can be seen that the minimum achievable norm remains at seven units if the lead time is 5 days or less and thereafter increases as a stepped function (Figure 4) as seen earlier.<figure class="image"><img style="aspect-ratio:4000/2632;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-5_1757846282.jpg" width="4000" height="2632"></figure>This pattern is because the norm for any item under MTA is decided by comparing sales within multiple lead time buckets in the past and adopting the peak sale as the norm. For slow-moving items, a typical lead time bucket comprises of multiple days of zero sales and a few days where sales would have happened. As the lead time considered increases, the bucket size considered for determining the norm also increases. However, for slow-moving items, an increase in bucket size does not always result in an increase in consumption within the bucket since additional days in a bucket may have generated no additional sales. An illustration is provided in Figure 5 for a slow-moving item during a demand duration of 15 days. As can be seen, an increase in lead time bucket size from 10 days to 14 days only adds multiple days of zero sales and does not result in an increase in total sales within the bucket, which remains fixed at seven units. Only on the 15th day does another sale event get added inside the bucket, and so the sale within the bucket increases. Therefore, it can be said that for a slow mover, the sale within a lead time bucket depends on the number of sales events that get covered within a bucket. The greater is the interval between any two sales events, the larger the lead time bucket size required to count both of them.<figure class="image"><img style="aspect-ratio:4000/2570;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-6_1757846292.jpg" width="4000" height="2570"></figure>According to the demand profile in Table 2, the minimum duration between any consecutive sale events in the entire duration of 30 days considered is 5 days. Therefore, a lead time bucket of 5 days or less can accommodate only one sale event or none (each sale event involves seven units) throughout the month. For lead time bucket sizes between 1 and 5 days, the peak consumption within the lead time remains equivalent to a single sale event, which is 7 units. The minimum norm possible for this item is 7 units and is achieved when the replenishment lead time is equal to or less than the minimum duration between consecutive sales (5 days in this case). This minimum time between two consecutive sales of an item as seen historically, can be defined as Minimum Time between Sales (MTBS).  And such items, which have reached the minimum norm values achievable, based on their demand profile (MTBS) and current lead times, can be called base norm items. The only way in which the norms for such items can be reduced further is through the reduction of their MTBS below the current lead times for example, by employing demand generation activities.<figure class="image"><img style="aspect-ratio:4000/2348;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-7_1757846302.jpg" width="4000" height="2348"></figure>But, since demand profiles (which determine the MTBS) and replenishment lead times of stocked items vary based on their location in the distribution network, an item which is a base norm item in one location of the distribution network need not be a base norm at another location. Therefore, a base norm item stocked at a given location in the distribution network should not be replenished from the immediately previous location/node as long as the increased replenishment lead time due to it being replenished from a more aggregated point, lying farther away, does not exceed its Minimum Time Between Sales (MTBS) at the location.  For example: <ul style="list-style-type:disc;"><li>If a Base Norm item with an MTBS of 30 days is stocked at a dealer location, with a transit lead time of 10 days.</li><li>Then any further reduction in transit lead times to less than 10 days will not result in any norm reduction for the item at the dealers.</li><li>Conversely, any increase in transit times from 10 days up to the upper limit of 30 days will not result in norm and inventory increase at the dealer locations for such items.</li><li>Therefore, such an item can be supplied directly from the central warehouse, which is farther away, provided the transit lead time for the item from the CWH is below 30 days.</li><li>If the base norm items were to be replenished directly from aggregated points, then no stock for such items would be required in the intermediate nodes in the distribution network, thereby reducing the overall inventory in the supply chain with no impact on serviceability. </li></ul> The cash released through inventory reduction at the intermediate nodes can be used to add more spare part SKUs under MTA purview to enhance customer serviceability. <h5>How to treat base norm items</h5>In any distribution network, as items are stocked farther away from the central warehouse, they tend to become progressively slower in movement from their stocking points. For example, a part moving every day from the central stocking location may move once every 10 days from the branches and once every 30 days from a dealer stocking point. This deterioration in movement is more pronounced in OEM service organisations, particularly those that are affected by the long tail. Such items are likely to become Base Norm items for their locations.  In many cases, owing to extremely low customer tolerance times, especially in the spare’s environment, it may be required to hold a considerable number of such base norm items at the POS (point-of-sale) in the dealer locations. In such cases, most firms, as a matter of long-standing policy, tend to replenish all SKUs from the immediately preceding nodes (like regional warehouses) in the distribution network. Such items lie unsold for long periods of time before being consumed and replenished in a comparatively short time after consumption, so the inventory turns for such items are very poor. Therefore, ideally, an item identified as base norm at any node in the distribution network should not be stocked as an MTA at the node. It should be maintained at a more aggregate location/node where it does not qualify as base norm - this aggregate node should be the location from where the movement of the item is faster than the lead time to deliver the item to that node (in other words, the MTBS at the aggregate location should be lower than the replenishment lead time to the node). As has been explained, replenishing such base norm items from such aggregated locations, which are farther away, will not result in any increase in the norm for such items at the POS locations. On the contrary, this will help eliminate the unnecessary inventory held at the regional warehouse for such items.  <h5>Conclusion</h5>As can be seen clearly, our inclination to build sophisticated distribution networks with a large number of intermediate nodes often only adds to the complexity and takes away from our ability to provide superior service to our customers. Organisations have to evaluate their stock on an ongoing basis, and any past decision to stock Base Norm items at multiple stocking points in the distribution network should be reevaluated and opportunities for removing the intermediate stocking nodes should be identified. The cash released thus can be used to stock other slow-moving SKUs that were hitherto not stocked.

Key Takeaways

OEMs stock a vast array of seldom-used but critical parts to maintain efficient customer service, thereby blocking capital. Can there be a balance of stocking enough to prevent service disruptions without blocking capital?

Spares distribution: When the long tail wags the dog

Hell hath no fury like a customer scorned! All OEMs know that any failure to support their customers during critical equipment breakdowns can have an enormous impact on the OEM’s reputation and future sales. OEMs that sell industrial equipment to corporate customers possibly know this better than most, but the service firms of these OEMs experience an enormous challenge called the “Long tail effect” due to low volumes of commissioned equipment in the market. For them, a large proportion of their spare parts sell infrequently and managing this long tail is key to managing the spares distribution environment. The challenge is that:<ul style="list-style-type:disc;"><li>A commitment to ensure uninterrupted service implies that any OEM service firm is duty-bound to stock even items that fail infrequently. On the other hand, stocking large numbers of slow-moving SKUs increases cash tied up in inventory.</li><li>With new models being launched at an increasing frequency, the number of spare part SKUs that require servicing increases. Consequently, an increasingly large range of spares needs to be stocked.</li></ul> To resolve this conflict, many organisations focus on lead time reduction at all nodes in the service distribution network. The assumption is that if lead times are reduced, the norms (minimum inventory) to be held at various nodes for every SKU will reduce, enabling stocking of an increased range of spares for the same working capital, leading to improved service. At first glance, this assumption seems correct, and so the procurement and distribution professionals of such service organisations are always trying to reduce lead times or redesign the distribution network to minimise transit times from one node to another. However, it has been seen that even if they succeed in reducing lead times, significant inventory reduction is not always achieved. <h5>Norm and its relationship with lead time</h5>The norm for an item stocked in a Make to Availability (MTA) environment is determined by the peak demand within the replenishment lead time.  Comparing demand within multiple lead time buckets in the past and identifying the peak sales figure attained within any one of the buckets gives us the norm. For example, if the reliable lead time of an item is 15 days, then sales figures within successive lead time buckets of 15 days in the past year are compared. Thereafter, the maximum demand in a 15-day bucket among them is designated as the norm for the item. Therefore: Norm = peak demand during a reliable replenishment period, factored for variability x reliable replenishment time.  Based on the above formula, norm and replenishment lead time are directly related, as depicted by this straight line graph depicted in Figure 1:<figure class="image"><img style="aspect-ratio:4000/3042;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-1_1757846234.jpg" alt="Vector Consulting Group | Article | Spare Distribution | Norm v/s Lead Time - Expected Lead Time" width="4000" height="3042"></figure>But while this may be the case often, this is not always so! Let’s examine whether the reality conforms to our intuition. For this analysis, we at Vector Consulting Group conducted a detailed study of both slow-moving and fast-moving parts of a large material handling equipment company in India. Depicted below is a graph (Figure 2) showing the annual trend of norm computation with increasing lead times for a slow-moving spare part of this service firm. This slow-moving spare part has been sold only 9 times in the last year, translating to a sale once every 40 days on average (365/9).<figure class="image"><img style="aspect-ratio:4000/3694;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-2_1757846245.jpg" alt="Vector Consulting Group | Article | Spare Distribution | Real-time data of a slow-moving spare parts where inventory norms function as a stepped function of replenishment lead times." width="4000" height="3694"></figure>This graph (Figure 2) is quite different from expected and indicates that the norm, when computed for such slow-moving parts, is a stepped function of lead times. In other words, the norm remains constant within a certain interval of lead time and changes in value only when moving from one interval of the lead time to the next. In the case of a fast mover which has sold daily for the past one year, the graph of norm vs the lead time for such an item looks like Figure 3:<figure class="image"><img style="aspect-ratio:4000/3623;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-3_1757846254.jpg" alt="Vector Consulting Group | Article | Spare Distribution | Real-time data of a fast-moving spare parts where inventory norms scale linearly with replenishment lead times." width="4000" height="3623"></figure>This graph conforms exactly to the initial intuition about the relationship between norm and lead time. Therefore, for fast movers, any reduction in lead time results in a reduction in norm, and this can be the right action to be taken for firms selling fast-moving goods like soaps and chocolate. But an initiative to broadly reduce lead times of items in a spare parts scenario irrespective, of their demand profile, may not be as successful in reducing inventory levels! <h5>Norm behaviour in a slow-moving parts environment</h5>It is clear that we need to delve deeper into the reasons for the norm graph appearing and behaving the way it does for environments with slow-moving parts. For this, let’s examine an item with a simplified demand pattern as depicted in the Table 1. Suppose the item has sold 4 times in the past month, the first sale being on 10th, the second sale after 5 days (on the 15th), the third sale after an interval of another 10 days (on 25th) and the fourth sale after another 5 days, on the 30th of the month. The minimum interval between two consecutive sales, as seen from the demand table is 5 days. Let’s assume for simplicity’s sake that the demand pattern of the item is repetitive, month on month..<figure class="image"><img style="aspect-ratio:4000/3016;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-img-4-1736130245_1776751888.jpg" width="4000" height="3016"></figure><figure class="image"><img style="aspect-ratio:4000/2280;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-4_1757846269.jpg" alt="Vector Consulting Group | Article | Spare Distribution | Norm versus lead time graph with flat norm at 7 units till 5 days and stepwise rise at higher lead times." width="4000" height="2280"></figure>The tabular representation of the norm vs lead time graph is depicted in Table 2. Therefore, if this item were to be kept under MTA and its norm was computed based on past month’s sales data, it can be seen that the minimum achievable norm remains at seven units if the lead time is 5 days or less and thereafter increases as a stepped function (Figure 4) as seen earlier.<figure class="image"><img style="aspect-ratio:4000/2632;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-5_1757846282.jpg" alt="Vector Consulting Group | Article | Spare Distribution | Comparison of replenishment lead time against norms" width="4000" height="2632"></figure>This pattern is because the norm for any item under MTA is decided by comparing sales within multiple lead time buckets in the past and adopting the peak sale as the norm. For slow-moving items, a typical lead time bucket comprises of multiple days of zero sales and a few days where sales would have happened. As the lead time considered increases, the bucket size considered for determining the norm also increases. However, for slow-moving items, an increase in bucket size does not always result in an increase in consumption within the bucket since additional days in a bucket may have generated no additional sales. An illustration is provided in Figure 5 for a slow-moving item during a demand duration of 15 days. As can be seen, an increase in lead time bucket size from 10 days to 14 days only adds multiple days of zero sales and does not result in an increase in total sales within the bucket, which remains fixed at seven units. Only on the 15th day does another sale event get added inside the bucket, and so the sale within the bucket increases. Therefore, it can be said that for a slow mover, the sale within a lead time bucket depends on the number of sales events that get covered within a bucket. The greater is the interval between any two sales events, the larger the lead time bucket size required to count both of them.<figure class="image"><img style="aspect-ratio:4000/2570;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-6_1757846292.jpg" alt="Vector Consulting Group | Article | Spare Distribution | 15-day lead time bucket analysis for slow-moving item with zero sales days and additional sale counted only on 15th day." width="4000" height="2570"></figure>According to the demand profile in Table 2, the minimum duration between any consecutive sale events in the entire duration of 30 days considered is 5 days. Therefore, a lead time bucket of 5 days or less can accommodate only one sale event or none (each sale event involves seven units) throughout the month. For lead time bucket sizes between 1 and 5 days, the peak consumption within the lead time remains equivalent to a single sale event, which is 7 units. The minimum norm possible for this item is 7 units and is achieved when the replenishment lead time is equal to or less than the minimum duration between consecutive sales (5 days in this case). This minimum time between two consecutive sales of an item as seen historically, can be defined as Minimum Time between Sales (MTBS).  And such items, which have reached the minimum norm values achievable, based on their demand profile (MTBS) and current lead times, can be called base norm items. The only way in which the norms for such items can be reduced further is through the reduction of their MTBS below the current lead times for example, by employing demand generation activities.<figure class="image"><img style="aspect-ratio:4000/2348;" src="https://web-assets.vectorconsulting.in/storage/editor_uploads//vcg-article-3-spares-distribution-when-the-long-tail-wags-the-dog-7_1757846302.jpg" alt="Vector Consulting Group | Article | Spares Distribution | Comparison of replenishment lead time against Minimum Time Between Sales (MTBS) to determine the minimum achievable base norm for inventory at a distribution node." width="4000" height="2348"></figure>But, since demand profiles (which determine the MTBS) and replenishment lead times of stocked items vary based on their location in the distribution network, an item which is a base norm item in one location of the distribution network need not be a base norm at another location. Therefore, a base norm item stocked at a given location in the distribution network should not be replenished from the immediately previous location/node as long as the increased replenishment lead time due to it being replenished from a more aggregated point, lying farther away, does not exceed its Minimum Time Between Sales (MTBS) at the location.  For example: <ul style="list-style-type:disc;"><li>If a Base Norm item with an MTBS of 30 days is stocked at a dealer location, with a transit lead time of 10 days.</li><li>Then any further reduction in transit lead times to less than 10 days will not result in any norm reduction for the item at the dealers.</li><li>Conversely, any increase in transit times from 10 days up to the upper limit of 30 days will not result in norm and inventory increase at the dealer locations for such items.</li><li>Therefore, such an item can be supplied directly from the central warehouse, which is farther away, provided the transit lead time for the item from the CWH is below 30 days.</li><li>If the base norm items were to be replenished directly from aggregated points, then no stock for such items would be required in the intermediate nodes in the distribution network, thereby reducing the overall inventory in the supply chain with no impact on serviceability. </li></ul> The cash released through inventory reduction at the intermediate nodes can be used to add more spare part SKUs under MTA purview to enhance customer serviceability. <h5>How to treat base norm items</h5>In any distribution network, as items are stocked farther away from the central warehouse, they tend to become progressively slower in movement from their stocking points. For example, a part moving every day from the central stocking location may move once every 10 days from the branches and once every 30 days from a dealer stocking point. This deterioration in movement is more pronounced in OEM service organisations, particularly those that are affected by the long tail. Such items are likely to become Base Norm items for their locations.  In many cases, owing to extremely low customer tolerance times, especially in the spare’s environment, it may be required to hold a considerable number of such base norm items at the POS (point-of-sale) in the dealer locations. In such cases, most firms, as a matter of long-standing policy, tend to replenish all SKUs from the immediately preceding nodes (like regional warehouses) in the distribution network. Such items lie unsold for long periods of time before being consumed and replenished in a comparatively short time after consumption, so the inventory turns for such items are very poor. Therefore, ideally, an item identified as base norm at any node in the distribution network should not be stocked as an MTA at the node. It should be maintained at a more aggregate location/node where it does not qualify as base norm - this aggregate node should be the location from where the movement of the item is faster than the lead time to deliver the item to that node (in other words, the MTBS at the aggregate location should be lower than the replenishment lead time to the node). As has been explained, replenishing such base norm items from such aggregated locations, which are farther away, will not result in any increase in the norm for such items at the POS locations. On the contrary, this will help eliminate the unnecessary inventory held at the regional warehouse for such items.  <h5>Conclusion</h5>As can be seen clearly, our inclination to build sophisticated distribution networks with a large number of intermediate nodes often only adds to the complexity and takes away from our ability to provide superior service to our customers. Organisations have to evaluate their stock on an ongoing basis, and any past decision to stock Base Norm items at multiple stocking points in the distribution network should be reevaluated and opportunities for removing the intermediate stocking nodes should be identified. The cash released thus can be used to stock other slow-moving SKUs that were hitherto not stocked.

<h5>Spares distribution: When the long tail wags the dog</h5>

Download PDF

Applying TOC thinking can help better manage distribution of spare parts and ensure uninterrupted service to the customers. Read to know more.