Future of Commerce Blog

How to Optimize Your Supply


Supply = Demand. Sounds simple, right? However, selling products quickly becomes complex.

Product options, such as size and color, matter. Knowing exactly how many products to produce isn’t enough. You should know how many of each product option to produce.

To optimize supply, first properly design your product catalog, then forecast demand, and finally determine your optimal supply.

Properly Design Your Product Catalog

By properly designing and setting up your product catalog, matching supply and demand for every product option is easy.

  • Include options and option types. Build your product catalog properly to include options (e.g. color, size, scent, material) and option types (e.g. small, medium, large). Doing so will help format products so that each product family can be broken down into a hierarchy by option type.

Anytime you need to reorganize your product catalog, just use the move product tool in Stitch.

  • Use consistent option types. Always represent the same option (eg. “small”) in the same way instead of using “small” sometimes and “s” or “sm” other times.

Use the search or export function to quickly see what differences exist or use the product option sales report to identify any inconsistencies in option type in Stitch.

Forecast Demand

Accurately and consistently keeping track of inventory and sales make it easy to forecast demand.

  • Assume that the ratio of options sold (demanded) in the past is a good guideline. It is not recommend to assume ratios will vary without additional data. In this case, incorporate uncertainty into your result. For example, if expanding into a geographic region known for its love of vanilla scented candles, it may be justifiable to make or buy more vanilla scented products than your data analysis suggests.
  • Choose a representative time frame. The time frame should be relevant for your analysis and be long enough to be representative. When data is representative, even if there were a few atypical sales, they will have only a small effect on your demand forecast.

Every business is different and effects such as seasonality affect each differently. For many, Q4 2012 data is ideal to determine what to make for Q4 2013 because the two time periods are so similar. However, you may also look at your entire sales history or only this year’s history.

  • Calculate the ratio of product options. Divide the number of each option type sold by the total number of units sold.

In Stitch, use the Product Option Drill-Down Sales Report to easily determine the ratio of each option type sold.

1) Select the time period for comparison.

2) Select the Product Option Drill-Down Sales Report


3) Drill-down by clicking on the bar for the selected option-type.


4) Determine the ratio for each option type using the bar chart or the sales report table.


Consider incorporating revenue per unit or profitability into your decision, especially if unit cost data is available. Product options with very few sales may fall below minimum production levels or may not make sense economically to continue stocking.

Determine supply

For each option type (e.g. size) or option type combination (e.g. size and color), determine how many of each product option to produce or purchase.

  • First, determine the total number of units you want to produce or purchase. For example, suppose your cut-and-sew shop has minimums of 1,500 pieces or your supplier requires at least 1,500 pieces per order. This may be the number of units you want to add to your inventory.
  • Next, multiply the ratio for each option within that group by the amount you are going to produce/purchase. For example, if 0.08 (8 percent) of all shirts were size small during the previous period and you want to produce 1,500 units, then the number of smalls to produce is 0.08 X 1,500 = 120. Produce 120 smalls for an order of 1,500 units.
  • Repeat for each product option.
The sum of all product options should equal the total number of units you want to produce. In this case, 1,500 units. Suppose your options also include x-small, medium, large, and x-large, making up 5%, 40%, 35%, and 12% respectively. Your production run would consist of 75 x-smalls, 120 smalls, 600 mediums, 525 larges, and 180 x-larges.

Introducing a New Option

Suppose you want to introduce a new product in red, but have previously only offered it in white, black, and blue. You have no data for how popular red will be.

  • Be cautious and use the best data available whenever you are looking to introduce a new option. It is typically better to under manufacture than over manufacture.

The costs of under-manufacturing are often less than those of over-manufacturing. This is particularly true when the volume necessary to meet demand is low. Almost always, paying a premium to produce/purchase a hot item quickly is less expensive than trying to get rid of unsellable products.

  • Don’t fall prey to high quantity discounts. Often people are lured to higher quantities because there is a price decrease; however people often forget that no matter how big the price cut is, if you cannot sell the products, the decision is unprofitable. Often A 10% price cut on 20% products that cannot be sold due to over-producing, is often worse than paying full price, but being able to sell all of your products.

Tell us: How do you make purchasing decisions? What part of optimizing supply do you find most challenging?

Bridge Mellichamp

Bridge Mellichamp is the Director of Data Science and Special Projects. Numbers excite her more than you can imagine; at the core, she’s driven by helping Stitch and its customers make sense of their data so they can make incredibly smart business decisions.In her free time, she enjoys winning at booking flights, sand in her toes at Ocean Beach, and escapes to Tahoe.

More Blog Posts

Learn how today's fastest-growing brands manage their inventory.