Safety stock is additional inventory held to reduce the risk that an item might run out of stock and acts as a buffer between forecasted and actual demand levels during the order period. It is included in order proposals to cover uncertainty in the forecast.
Safety stock will be included in order proposals whether future demand comes from the statistical forecast or the user-entered forecast override.
Calculating safety stock
There are three main factors that influence the calculation of safety stock:
The statistical forecast in AGR is calculated solely on the sales history of an item in a location. The more robust the sales history is, the better the forecast will be. AGR’s forecasting module picks up seasonal patterns and trends if the item has enough historical data. If the sales history is either very volatile or very short, AGR’s forecasting module will not be able to produce an accurate forecast and there is a higher risk of forecast uncertainty so safety stock will be calculated to allow for overachievement against the forecast.
The confidence factor (CF) influences the service level for the item. The higher the confidence factor for the item, the higher the safety stock will be. The confidence factor assigned to an item is determined by the ABC value calculated for that item.
AGR will use the best fit statistical forecast using the item's sales history but due to the nature of forecasts, there will always be some uncertainty.
Overriding confidence factors
Confidence factors can be overridden in the Item Details section of the item card.
There are four general levels for confidence factors.
A confidence factor of 99.9999 is the highest CF possible, resulting in almost no chance of stock out.
A confidence factor of 90 should be high enough for most products that require high safety stock. A stockout would occur only in extreme cases.
A confidence factor between 51 and 89 would be suitable for products with a lower importance than A items if you want some safety stock but you are comfortable with occasional stockouts when sales volume is higher than forecasted.
A confidence factor of 50 means that no safety stock is calculated.
Limits on confidence factors
Confidence factors cannot be set to 100 since that would yield an infinite amount of safety stock.
Confidence factors between 0-49.9 are invalid, since AGR is not concerned about items which may sell fewer than the forecasted units.
Safety stock is calculated for the order period. The first example below shows how the order period affects safety stock.
The examples below go through how the different factors impact safety stock calculations.
Changing the order period
This item that has a long and robust sales history, a high confidence factor (95%), and an order period of 8 days (lead time days = 1 and order frequency = 7). Forecasted sales quantities are generally around 2,000 units per month and safety stock for the order period is 81 units
If the lead time was changed to 6 months (lead time days = 180) but nothing else changed then safety stock would go up to 1025 units for the order period.
Safety stock seems wrong
Monthly sales on an item have never exceeded 1100 units per month, but AGR calculates a safety stock of 1500 units. It looks like AGR is trying to maintain nearly 6 weeks of safety stock which feels like too much.
Looking closely at the item reveals that it
is classified as an A product, so the confidence factor is 90
has a lead time of 130 days and order frequency of 30 days, so the order period is 60 days (5.3 months)
has a short and volatile sales history with quick growth in the beginning and then decreasing just as quickly, thus creating uncertainty in the forecast.
This item fits all the criteria for a high safety stock, since safety stock is a factor of forecast uncertainty, order period, and confidence factor.
Once we see why the safety stock is so high, we can then decided whether the amount is justified. The safety stock shown on the chart is calculated per order period, not per day, week, or month. The calculations below can give us the safety stock per day and then per month.
Safety stock for order period = 1500 units.
Order period = 160 days.
Safety stock per day = 1500 / 160 = 9.4 units per day
Safety stock per month = 9.4 * 30 = 280 units a month.
The sales in July were around 1100 units and around 800 units in August. When forecasting the sales for the following month, it wouldn’t be unrealistic to expect to sell 1100 units again. By setting a confidence factor of 90, you are saying that you want to be more prepared for a sale of 1100 units.
AGR is not trying to maintain 6 weeks of safety stock. Rather, it is taking into account the fact that the forecast might be wrong by up to 1.5 months of stock when ordering for roughly 5 months. In other words, AGR forecasts sales of around 800 units a month but adds 280 units of safety stock to make up for the uncertainty in the forecast.
High forecast uncertainty
Safety stock can be adjusted to take into account large fluctuations in the sales history that make the resulting statistical forecast a less reliable indication of future demand.
In this example, the item has no sales in recent months, leading to high uncertainty in the forecast. There is more robust data further back, but will the future match this or more recent activity? AGR set the safety stock to a much higher level than the forecast.
The higher a product’s confidence factor is, the higher the safety stock will be. AGR treats items with high confidence factors as items that should go out of stock very rarely. In the two items shown below, both have forecasted sales of 200 units per month but with different uncertainty in the forecasting.
In the first item, there is low uncertainty because sales vary between 190 and 210 units per month, so the safety stock is 10.
In the second item, there is high uncertainty with sales varying between 100 and 300 units per month, so the safety stock is around 40.
Although somewhat simplified, the graph below is a recreation of the two AGR screenshots from above. The forecasted sale for both items is 200 units, but the safety stock is not the same.
The blue line represents the first item with a low uncertainty. It is highly concentrated around 200, meaning that the likely sales will be near 200. Sales could be between 125 and 275.
On the other hand, the red line represents the second item with a comparatively high uncertainty. The sales will most likely be 200 units, but they might also fall within a wider range than the first item, between 75 and 325 units.
The confidence factor for each item is set to 90. The shaded area under each of the red and blue lines is 90% of the total area underneath the line. The right hand side of the blue shaded area is at the value 210 and of the red shaded area at 240. Revisiting the items above, that’s exactly what the safety stock is for each of those items.
If the confidence factor were to be changed to 99, the blue shaded area would move far to around 275 units so that with 99% certainty, the expected sales will be up to a maximum of 275. In that case, the forecasted sales would be 200 units and safety stock would be 75 units.
If the confidence factor were to be set at 50, the right side of the blue shaded area would fall exactly on the forecasted sales of 200, so there would be 0 safety stock. This is ideal for products that have no uncertainty in their demand, for example, made to order items.
Min stock and safety stock
If minimum stock (Min Stk) is set, this is included in the order proposal calculation in addition to safety stock.