How to Determine Service Levels

Service levels are a critical component of supply chain management, representing the balance between inventory holding costs and stockout costs. At their core, service levels define how often a company can fulfill customer demand without running out of stock. Striking the right balance is essential—carrying too much inventory leads to high holding costs, while insufficient stock can result in lost sales and dissatisfied customers.

The Challenges of Setting Service Levels

Supply chain managers should evaluate service levels on a periodic basis to ensure they aren't leaving money on the table. However, service levels are not always straightforward to set and interpret because they involve complex statistical considerations:

1. Non-linear inventory relationship – The amount of inventory required to maintain a given service level does not increase linearly. As service level targets increase, inventory requirements grow exponentially.

   

   
   

2. Assumptions about demand distribution – Many off-the-shelf inventory management software solutions assume demand follows a normal distribution, but in reality, demand is usually not normally distributed.

3. Ambiguity in service level definitions – Different stakeholders may interpret service levels differently, leading to misalignment in expectations and execution.

A Data-Driven Approach to Setting Service Levels

Rather than arbitrarily setting a service level, companies should take a data-driven approach by defining the costs of under stocking and over stocking. By doing so, they can establish an optimal service level that minimizes total cost rather than relying on guesswork.

Defining and measuring the factors that drive inventory costs is the first step to optimizing service levels. Each company will have a different set of costs, but here are some common ones:

Holding Costs:

• Storage and warehousing costs – Expenses related to storing inventory, including rent and utilities.

• Capital costs – The opportunity cost of tying up capital in inventory rather than other investments.

• Insurance and depreciation – Costs incurred from insuring goods and accounting for their potential loss in value over time.

• Obsolescence risk – The likelihood of products becoming outdated and unsellable.

Stockout Costs:

• Lost sales and revenue – The immediate opportunity cost of not having products available for customers.

• Customer dissatisfaction and churn – The long-term effects of failing to meet customer expectations.

• Expedited shipping costs – Additional expenses incurred when rushing stock replenishment to meet demand.

• Reputation damage – The risk of losing trust and brand equity due to frequent stockouts.

Stockout costs are typically more abstract to define than holding costs. This is the step where it might be necessary to make some assumptions. However, it is still more intuitive to define the cost of a lost sale than to set a service level off the top of your head.

Example: Determining the Optimal Service Level

Let's consider a retailer that sells two different products. The first is a high-value hair dryer that is purchased for $50 and sold for $100. The second is a low-cost comb that is purchased for $0.50 and sold for $1.00. Both of these items are important to keep in stock, but they obviously have different demand patterns and purchasing priorities.

Let's suppose we have the following costs associated with these items:

Cost Factor	Hair Dryer	Comb
Storage Cost	$2/unit/month	$0.10/unit/month
Capital Cost	$0.21/unit/month	$0.002/unit/month
Lost Sales Opportunity Cost	$50/unit	$0.50/unit
Reputation Damage Cost	$20/unit	$20/unit

We also make the assumption that the purchasing cycle is one month, and that the holding costs are realized as the cost of over stocking per unit, per month. These assumptions could be changed for your specific business context, and the following analysis would remain the same.

The hair dryer has a total holding cost of $2.21/unit/month and a stockout cost of $70/unit. The comb has a total holding cost of $0.102/unit/month and a stockout cost of $20.50/unit. Relative to the stockout cost, the comb has a much lower holding cost. This indicates that it is optimal to maintain a high service level on the comb and to maintain a lower, tightly controlled service level on the hairdryer. The ratio of holding cost to stockout cost, along with a probabilistic demand forecast, directly translates to an optimal service level.

Since the derivation of the optimal service level can become quite complex, we will make the assumption that the demand is normally distributed. For some organizations, this assumption is appropriate. However, it is often suboptimal and can leave efficiency gains and cost savings on the table.

When we assume that the demand is normally distributed, the optimal service level (SL*) can be calculated as:

where

This corresponds to an optimal service level of 96.64% for the hair dryer and 99.50% for the comb.

Since the comb is cheap to purchase and store, we can simply procure a large quantity so that they are always available. Since the hairdryer is more expensive and has a higher holding cost, we must be more deliberate in how much we order to avoid a costly overstock situation.

How Glassoff Consulting Optimizes Service Levels

At Glassoff Consulting, we optimize the service level at a micro-level for every SKU in a client's product line. We do this by constructing probabilistic demand forecasts and accounting for various business constraints, including:

• Minimum order quantities (MOQs)

• Bulk purchasing discounts

• Case pack sizes

• Lead time variability

• Etc.

All of these factors have an important influence on the optimal service level and must be considered to remain operationally efficient.

Take the Next Step

If you’re ready to optimize your inventory strategy, Glassoff Consulting can help. Schedule a free 30-minute consultation with our experts to learn how our data-driven approach can maximize your service levels while minimizing costs.