Math behind optimal supply chain

In theory, an online retailer should not keep inventory in its warehouse. A customer always sees the product in stock, available for immediate purchase in infinite quantities. He makes an order, and you forward it to the customer, keeping your margin.

This is known as drop-shipping. It is possible, but can only be employed in some cases. More often than not, if the supplier is located on another side of the planet or the manufacturing process takes time, the customer won’t wait for a few weeks for new shoes. He wants them by this weekend.

Introducing warehouse

You can keep a pair of shoes for quick delivery on your site! When a customer orders, you can immediately ship this pair and purchase a new one for the next sale.

Yes, it requires freezing some money to keep the pair of shoes until someone orders it, but now your delivery time is a day, not weeks.

Warehouse to mitigate lead time

You just sold one pair and ordered a new pair. It will arrive in ten days. But what if another customer wants to order the same SKU before it arrives? Oh, you’re out of stock now, and you’ve likely lost the sale.

Can you keep more than one pair in stock? Sure, but how many? If there are too few, you’ll lose sales. If there are too many, you freeze more money than necessary in your stock.

To answer the question, we need to know the demand—that is, how many pairs it is expected to sell during that ten-day arrival window. If you sell, on average, one such pair in two days, the wise stock level quantity to have will be five pairs.

The level described here is known as “optimal quantity minimum” or “optimum min” for short.

Considering quantities incoming

In the example described, what if you have two pairs left in your warehouse and another five in transit on their way to you? Should you purchase more? Probably not. At the moment, when the subsequent sales will happen, the incoming shoes will likely land in your warehouse, and the stock will go up.

The process of leaving and entering your warehouse is continuous. So, for computations, it makes sense to operate with the “absolute stock level quantity,” which includes the stock available for sale and the total incoming stock.

In the example above, items forming the absolute stock are shown in brown color.

Taking the replenishment cycle into account

It’s unlikely that you have the time, capacity, and logistic means to purchase a replacement unit right when a sale is made. More realistically, you review the stock level of shoes once per week, for example.

In this case, you must keep more than five pairs of shoes to cover the purchase review period. In our example, there will be another four pairs of shoes, for a total of nine.

This level is known as “optimal quantity maximum” or “optimum max.” Ideally, your warehouse stock will swing between optimum min and max.

Meet recommended replenishment

Given all the base numbers, it is possible to recommend the number of shoe boxes to purchase. It is as simple as the difference between the optimum max and absolute stock level quantity.

That’s what you see in the “Repl” column inside the Stock Roll application.

Gotchas

An attentive reader might ask how the demand is calculated. That’s a difficult question without a single correct answer. Demand is an unknown value because, in general, it is hard to predict the outcome of marketing efforts, spikes created by a single big customer, a hype wave, and so on.

Stock Roll application takes a pretty straightforward, adequate approach in many cases. It looks back to the recent sale history and takes the average to determine the daily/weekly/monthly demand. The lookback window is adjustable within the application.

It’s not perfect. To handle spikes during upcoming holidays, for example, you might additionally keep some safety stock aside. This will further increase your investment in stock, but it will also save you from out-of-stock and missing sales during the hype.

There are other methods for demand computations. For example, so-called trusted intervals or manual sales plans. If you would like alternatives to be introduced in Stock Roll, please tell us about your use case by writing to hello@vierpunkte.com