Background decoration Background decoration

The Safety Stock Optimizer: Ensuring accurate safety stock levels set dynamically

Over the past few months, supply chain resilience took the spotlight as companies struggled to satisfy customer demand. The sanitary crisis amplified conventional problems such as inaccurate forecasts, transportation delays, and quality control problems, while companies’ safety stock levels were stress-tested, making it difficult for some to guarantee their service levels

Safety stocks exist as we cannot forecast what will happen in the future with 100% accuracy. Would this be possible, we would only need to carry the required amount of material needed for production until the next delivery. As the uncertainty grew with the latest global crisis, major industrial companies were tempted to raise their safety stock levels to increase their chances of reaching the targeted service level. 

However, before making the change, key considerations remain:

  • By how much should I adjust my safety stock levels? 
  • How will these adjustments increase or decrease my costs?
  • What will the effect on the service level be?

This is why one of our clients asked us a few months ago if we could develop with them a recommendation tool to dynamically readjust the safety stock level according to the expected lead time and desired target service level. As we were already importing their MRP related data for the Order Planner and Opportunity Dashboard, we applied the Total Cost of Ownership approach to develop the Safety Stock Optimizer

Here is why and how we did it.

Cost of stock-outs vs inventory costs

Safety stock is defined as the level of additional stock that is maintained to mitigate the risk of having a stockout caused by uncertainties in supply and demand. Adequate safety stock levels enable business operations to proceed according to their plans, increasing the chances of meeting the target service level (Monk 2009). 

In the case of GenLots, we operate on the production side of the supply chain and so, the service level corresponds to the probability of being able to fulfill the input material needs for production. For example, a 98% service level means that 98 out of 100 times your stock will satisfy your production needs. 

Manufacturers will try to satisfy production needs as much as possible so they can avoid stopping the production. At the same time, maintaining a high inventory level is very costly (cost of capital, storage costs, scrap costs…). This tension means that there is a trade-off between lost sales and the inventory costs when determining the target service level. A business that fails to understand the logic of correctly setting their safety stock level is at risk of being exposed to unnecessary costs and low service levels.

Why is a safety stock optimizer needed? 

To determine the safety stock levels, some operation managers resort to subjective methods including intuition, experience, or knowledge of the market. These methods are simple and time-efficient, but are prone to error and don’t take into account all the cost optimization that comes with the implications of increasing or decreasing those levels. 

On the other hand, companies may rely on internally validated formulas. However, in the face of constant changes to the global environment, these processes often become outdated and are not reviewed frequently enough. Major industrial companies often strive to be the most efficient, but when an unfortunate or unpredictable event happens it becomes difficult to optimize costs thoughtfully and rapidly.

Safety stocks need continuous, dynamic, and resilient adjustments in order to:

  • Minimize inventory costs
  • Maximize customer service levels

How is GenLots’ Safety Stock Optimizer different 

To build our own safety stock optimizer, we implemented King’s Formula (2011) upon the request of our client and adapted it to provide a degree of independence when it comes to the lead time and demand variability.

The way we built the tool has two major advantages: 

  • We allow the planner to simulate different service levels for each material to compare the impact on the company’s operations, financials, and safety stocks. The tool compares the order plan of the material with the current safety stock and the new one proposed by GenLots. As GenLots is already optimizing the order plan with the newly computed safety stock, we ensure that the resulting cost changes are taken into account when comparing the impact on the changes in safety stock. 
  • We can determine what the current service level is based on the current safety stock level. This allows material planners to efficiently reassess the safety stock levels on products at risk of being out of stock as well as on those products that may be generating greater inventory holding costs than necessary.


To provide a concrete example, the image below shows the effect of setting the target service level to 99% versus 95%. The graph shows the current safety stock of 13,000 units.

  • With a 99.0% service level, we would recommend lowering the safety stock to 7’563 units allowing up to €53’590 of potential savings if the order planner follows GenLots’ recommendations. 
  • With a 95.0% service level, the Safety Stock Optimizer recommends lowering even more the safety stock to 5’348 units, more than half of the current safety stock, and yields a 39% average inventory reduction. This inventory reduction will allow up to €77’260 in savings.

Safety Stock Optimizer demo product

With the data being refreshed daily on GenLots, the Safety Stock Optimizer will recommend something different every day, providing the planner with an updated safety stock recommendation in the case of an unforeseen event. 


Future developments and improvements

The tool is already available on the GenLots platform for our current clients who might need it and as it is the first version, here is how we plan to improve it: 

  1. The Safety Stock Optimizer recommends safety stocks based on King’s formula (2011), but we will add the capability to directly apply our client’s desired formula as other formulas might be better suited for certain types of materials. For example, we know that King’s formula assumes a normal distribution of the demand, but some materials are Binomial and some are Poisson distributed. 
  2. We are using the service level to determine the safety stock level; we thought it could be interesting for our clients to alternatively observe the actual service level of each material based on the current safety stock level directly on the Opportunity Dashboard. This could provide useful information on the materials that have too high a safety stock or on the safety stocks that are not covering enough demand. We could even imagine providing alerts on materials for which the service level is below a determined threshold in the Opportunity Dashboard.
  3. We are already working on integrating forecast variability into our formula for one of our clients. This will allow for increased safety stock level precision as well as increased supply security. 


If you are interested in learning more about the Safety Stock Optimizer and evaluating how our software can meet your needs don’t hesitate to get in touch with us at:

Product releases