May 17, 2017
Lesson 3: Safety Stock
Why holding the right safety stock is so important
If your suppliers always delivered the products you need in full at the exact moment you needed them AND you were able to perfectly predict when the stock is going to be needed by the customer down to the minute every time, then you wouldn’t need any buffer or safety stock at all. You would plan to have the product come in through the back door at the same time as the customer walked through the front door, meaning you could work in a true Just In Time (JIT) fashion. But that’s not the real world, is it?
The real inventory risks
The purpose of safety stock is to protect you from the two major inventory risks in your business:
- Forecast risk – Your inability to always forecast 100% accurately
- Supply risk – Your suppliers’ inability to deliver on-time and in full
Let’s start with Forecast Risk: your biggest risk in forecasting is that you sell more than you thought you would. That leads to stock-outs, lost sales, and unhappy customers. If you continually sell more than you forecast for an item, it’s a good idea to keep more safety stock for that item so that you can still sell the item, even if the sales are more than you thought they would be.
Similarly, Supply Risk is when your supplier habitually delivers late. If you keep some safety stock on hand, you can keep selling the product in that time between when the supplier said they would deliver and when they actually deliver.
Suppliers have another risk: under delivering. If your supplier is prone to delivering short quantities, you should keep some safety stock to make up for the number of units they typically under deliver.
Measure and react to risks
In order to make sure you cover these risks, you cannot rely on a gut feel or your best guess — you have to measure them. Forecast accuracy of the system and people involved must be tracked and measured and factored into the calculation of the safety stock. Your suppliers’ reliability for delivering the correct quantities at the right time should also be measured and tracked. This needs to be done per product per warehouse to get to the correct level when you can start planning the correct buffers of stock to carry.
Inventory policy can be defined as balancing your investment in inventory with your desired service level to your customers.
Typically, if you want better service levels, then you’ll have to invest more money in inventory. Or if you want to save money in inventory, your service level will suffer.
Sometimes, you can have it both ways: by investing your money in the correct items (like your A items) and taking investment away from the less important items (like your C items), you can save money and provide better service levels to your customers.
The way to flex your inventory policy is also through safety stock. Your service level will be better if you carry more safety stock on an item, but that requires you to have more cash tied up in your warehouse.
The less money you invest in safety stock the better your cash flow, but your service levels may suffer as a result.
Your Target Service Level is a big component of your inventory policy. The other two important inputs to the safety stock calculation are your Replenishment Cycle (when I buy inventory and how much do I buy), and your Lead Time (how many days does it take for the inventory to arrive in the warehouse once I place an order).
Here’s a summary of the inputs to a good Safety Stock number:
- Forecast Risk: the more I under-forecast, the more Safety Stock I need to have
- Supply Risk: the more a supplier under-supplies, or delivers late, the more Safety Stock I need to have
- Lead Time: the longer the lead time, the more Safety Stock I need to have
- Replenishment Cycle: the shorter the cycle, the more Safety Stock I need to have (frequent, small deliveries are more risky)
- Target Service Level: the better service I wish to provide my customers, the more Safety Stock I need
Keep it dynamic
The problem with how most companies set Safety Stock is that they adopt a set-and-forget mentality. They set Safety Stock to a blanket four weeks for all items in a group in January. By September, no one has reviewed that decision and they are either sitting on mountains of excess inventory or their customers are buying elsewhere.
You have to review your Safety Stock levels regularly because your risks change, your forecasts change, or your business objectives change. It would be ideal to review them on daily basis, but you’ll probably need a good automated system to do that for you reliably.
What can you do to improve this today?
You may be a little stuck on simple things to do to help improve your buffer stock. All this information makes sense, but what should you do today to make a difference?
Let’s start with the basics you need to consider if you do not have an engine that can do all the above math at a product-per-warehouse level:
- Calculate the forecast accuracy per group of products. You cannot get to a product level with this approach, but start with a view of total sales against what was expected per product group. This provides some indication of whether this is predictable or not, and it can be categorized as a Very High, High, Medium or Low error per group.
- Estimate the supplier reliability per preferred supplier. Go through your suppliers and talk to the team to categorize the supplier reliability (both in quantity and date delivery) into Very Bad, Bad, OK, Good and Excellent.
- Create a schedule of all products per warehouse that you want to carry safety stock for. Make sure to include the lead time, preferred vendor and the product group for each. Add to that your ABC classification done earlier.
- Create extra sheets for the forecasts accuracy estimate per group as well as the supplier list and reliability so that we can create a VLOOKUP in Excel later for this.
- In these additional sheets, allocate a percentage to each of the categories created for the supplier risk and forecast accuracy. These will be applied later to the safety stock calculation, so make sure to enter as a fraction. For high supplier risk, you would use 0.9, for example, as a factor. For low risk, you would use something like 0.5. For forecast error, a high error could be 0.9 and a low error 0.5.
- Allocate a fraction against the ABC category you created and set say the A category as the one where you want the highest protection at 0.9, for example, and the C category at something like 0.7.
- Update the main schedule per product per warehouse to lookup the fractions for ABC, Supply risk and forecast risk, and then add in a calculation that takes the lead time in days or months and multiplies this by the fractions you just looked up. This is your safety stock in days or months that can now be applied to the forecasts to calculate the buffer needed.
This will give you a safety stock per product that is variable based on a risk assessment of supplier performance AND forecast accuracy. This is not the ultimate formula and there are many textbook examples of more advanced calculations you could apply if you have the resources and tools to do so. This will, however, give you a starting point that you can use to improve this thinking as your tools and experience improve.