How to calculate the correct amount of inventory to order

In our previous Blog post, “The Hidden Costs of Carrying Too Much Inventory”, we discussed some often overlooked costs relating to inventory.

Ordering too little means you run into an out of stock situation, leading to reduced revenues and a poor customer experience. On the other hand, ordering too much inventory impacts your business’ margins via hard costs like storage fees and obsolete stock, and also growth via opportunity costs like reduced sales and marketing budgets.

Since ordering too much and too little inventory can have major impacts on businesses we thought it would be useful to discuss frameworks on how to calculate the correct amount of inventory to order.

These frameworks can get into more advanced mathematical modeling (which we may discuss in a later post) but here we are going to focus on simpler frameworks that smaller companies (<$10M in annual revenue) will find most applicable.

Below we discuss some of the most important aspects to consider when calculating the amount of inventory to order from your manufacturer.

Ordering the Correct SKUs

We touched on this point in our previous post, “8 Tips to Prepare for 10x Sales in Q4”, but it is worth restating again: not all SKUs are created equal, and they should not be treated as equal.

The key thing to focus on is making sure the main selling products are always in stock.

Figure 1 shows that the top 10% of products by sales make up over 60% of sales volume, while the top 20% make up roughly 80%.

Focus on keeping your top selling 20% of products in stock. The remaining 80% are second class citizens.

Understanding Your Order Constraints:

Your manufacturer will place constraints that will impact the order quantity.

Minimum Order Quantity (MOQ)

This varies from manufacturer to manufacturer and what they class as an individual SKU. For example, a manufacturer may state the MOQ is 100 units per SKU where they define a SKU as all variants (e.g. different colors all contribute to the MO). While other manufacturers may state that each color variant has its own MOQ of 100 units. 

MOQs are only an issue when your ideal order quantity is less than the MOQ. In our experience that is only relevant for smaller companies (<$100k in yearly revenue) or companies with extremely large SKU counts (in the thousands), like apparel products.

Lead time

By knowing the lead time of inventory orders you can understand a core data point that enables you to order the correct amount of inventory. The individual aspects of lead time are:

• Sub-suppliers – You may use multiple sub-suppliers that need to be taken into consideration. Many contract manufacturers will hold inventory from sub-suppliers for components that are used frequently, but if it is an unusual or expensive component, the contract manufacturer may order the item only when you place the purchase order with them.
• Contract Manufacturer – Once you have placed a PO, the manufacturer will have a timeline to complete your order based on their backlog and processes. 
• Transport – Most contract manufacturers are based in China so the options of getting inventory to the US are either air or ocean shipment. Including time to get through customs, air shipment can take 3 to 4 weeks, while ocean can take 6 to 8 weeks. 
• 3PL Receiving – Once the inventory arrives at your 3PL (Fulfillment Center) it will need to be received into their systems before it can start to be shipped to customers. We have seen this take anywhere from 2 to 7 days. 
• Overlap – This refers to how much older inventory you want in stock after the new inventory has been received. Having overlap (similar to but not the same as Safety Buffer Stock below) gives some extra inventory to ensure you never go out of stock in case some of the lead times above take longer than expected.

Let’s look at an example of these lead time constraints with some real data we have seen.

Predicting Future Sales

Now you have calculated the amount of time it will take to get new inventory in – let’s say it is 16 weeks. The next step is to establish how many units will be sold in this timeframe. Let’s look at some data points to work this out.

Purchase Orders (POs) – Not applicable for Direct to Consumer only e-commerce companies, but for companies that sell to retailers or distributors, having POs in place for the following quarter will give all the guidance required on how many units are needed to be ordered.

For Direct to Consumer e-commerce companies it is a bit more difficult than summing up POs. You’ll need to look at historical figures and use trends to estimate future sales.

Predicting Growth – Using Figure 2 we can see sales growing over time for a Kickpay client’s product. It shows the moving average for the total number of units shipped over the previous 16 weeks.

The most recent 16 week period had 3,184 units shipped. During the previous 16 weeks the figure was 1,973 units, a ~1,200 (60%) unit growth. If you were taking an optimistic view you could say sales would grow another 60% to 5,100 in the next 16 weeks but that seems excessive since the sales rate would be accelerating. 

A more conservative view is that sales would grow at the same rate by 1,200 units, to 4,400 units shipped over the next 16 weeks. That would still be optimistic, so I’d proceed with caution, but it’s not outside the realms of possibility.

One point to note: most sales graphs for products don’t look as positive as this graph shown in Figure 2. Most will show either steady growth, with lots of ups and downs, or are relatively flat, so bear that in mind when comparing your data. 

When predicting growth it is worthwhile to note if the historic data takes into consideration periods of time that have seasonal spikes in sales, like Q4. In the example above we didn’t take into account the spike in sales that comes with Black Friday / Cyber Monday. Let’s see how we could incorporate this. 

Sales Spike Events – Let’s have a look at how sales spiked last year over Black Friday / Cyber Monday (BFCM).

We can see in Figure 3 over the course of BFCM (week 48) and the week before Christmas (week 51) sales jumped by 2x to 60 units shipped per week. 

If you are ordering inventory to be sold over BFCM we can build this sales spike into our estimates.

If we use the data from Figure 2, it would mean ~200 units (3,184 units / 16 weeks) will be shipped over the next few weeks. If shipments doubled to 400 units / week for weeks 48 and 51, it would mean an extra 400 units should be ordered to cover the sales spike.

This would bring a total order volume up to 4,800 (4,400 units from standard growth + 400 extra units from BFCM) to be sold over 16 weeks. 

Now we have used historical data to get an estimate on sales growth for your next order. The important thing to remember is that these frameworks aren’t a crystal ball into the future and won’t be 100% accurate. If you want to have further peace of mind, we can use some safety buffer.

Safety Buffer Stock – This is extra inventory in case sales increase at a faster rate than initially estimated. 

We normally see 10%-20% being a good line in the sand for extra inventory. Wanting to do anything more than that suggests there is a big change on the horizon in relation to sales channels (i.e. a large retailer PO, a feature on Amazon, a significant jump in marketing budget, etc.) which you’ll need to incorporate into your initial estimates. 

Using 15% buffer would take our order volume up to 5,500 (4,800 units + 15%).

Overview

After making sure you are ordering the correct SKUs and you are over your Manufacturers Minimum Order Quantity, you need to focus on the lead time of your inventory. 

Inventory lead time includes:

• Manufacturing time (Contract Manufacturer and Sub-suppliers)
• Transport time 
• Receiving
• Overlap

In our example, we calculated this to be 16 weeks.

Once you know how much inventory you need to hold before another order is required, we need to estimate the number of units that can be sold in this timeline.

For companies selling to retailers we can use POs to get a close to exact number on the required amount of inventory required.

For Direct to Consumer e-commerce businesses we need utilize historic data to make estimations on future sales. In our example we used:

• Future sales (over 16 weeks): 4,400 units
• Sales spike events (e.g. Cyber Monday): 400 units
• Safety Buffer Stock: 700 units (+15%)

Total: 5,500 units.

This would still be an optimistic scenario, since most products don’t have the sales growth of the example we used, but it illustrates a framework you can apply to your own products.

---

If you are looking to see how Kickpay can be helpful in financing your next inventory order, come and and learn more here.