How to leverage Amazon’s forecast data to reduce out-of-stock issues

BLOG: Retail

6 Min Read | February, 2024 | BY Jordy Rinaldi

Understanding Amazon Forecast Types: P70, P90, and Mean

Amazon provides different forecast types to cater to various business needs:

P70 forecasts are conservative, minimizing overstock by providing a 70% confidence level that demand will not exceed the forecasted amount.
P90 forecasts are designed to avoid stockouts with a 90% confidence level, although they carry a higher risk of overstock.
Mean forecasts offer a balanced view, ideal for businesses with flexible logistics and storage capabilities.

The choice of forecast depends on your risk tolerance and business strategy, such as prioritizing minimal overstock or aiming to maximize sales.

What forecasts do they provide, and which one should be selected?

These forecasts gauge future product demand by analyzing historical sales data, trends, seasonality, and other variables. They assist in projecting the expected number of units to be sold in each period, facilitating inventory management.

The decision on which forecast to utilize hinges on your business objectives and may vary across product segments:

For risk-averse businesses:

Opting for P70 could be preferable to mitigate overstock costs

If maximizing sales is the goal:

Employing P90 might be advantageous to minimize stockouts, particularly for high-demand or seasonal items.

For businesses adopting a more balanced approach:

The mean forecast could be chosen, particularly if flexible storage solutions and efficient supply chains are in place.

Aggregating Forecast Data Across Countries

For enterprises operating across various markets, consolidating forecast data from different countries can offer a holistic perspective on global demand, enabling streamlined inventory distribution and management.

To aggregate data across countries, follow these steps:

Downloading forecast data for each target country.

Summing up the forecasted demand and existing inventory across all countries.

Calculating weeks of coverage to ensure efficient inventory distribution.

Effective Purchase Ordering Using Forecast Data

Comprehending forecast data empowers you to enhance decision-making regarding order volumes. By synchronizing order quantities with predicted demand, you can refine your inventory levels, striking a balance to avoid overcommitting to orders while meeting the demand for sought-after items.

To align purchase orders with forecasted demand:

Compare forecast data with current inventory levels.

Determine the gap between expected demand and available stock.

Adjust order quantities to either match or slightly exceed the forecasted demand, based on your strategic goals.

Reject or postpone orders that surpass the discrepancy significantly, unless you are aware of imminent sales events necessitating extra stock.

(Click on the image to expand)

Doing this calculation manually each time is time-consuming and increases the risk of errors. To alleviate this, we've created Catapult, a tool designed to fully automate the data preparation process. With Catapult, you can bypass the data preparation stage and concentrate on implementing actions based on your results.

Managing Out-of-Stock Situations

Generating a forecasting report also aids in detecting and preventing instances of out-of-stock scenarios. Often, Amazon anticipates high demand but lacks sufficient stock. This can occur due to insufficient ordering or no ordering at all, resulting in out-of-stock situations. So, what steps can you take upon identifying such cases?

amazon vendor playbook

From Stockouts TO SUCCESS

Explore 4 hidden costs of stockouts, 15 proven tactics to stay in stock, real case studies and more.

DOWNLOAD FOR FREE

Your initial action is to gauge the impact by assessing the business implications of potential stockouts. This involves computing the anticipated stock shortage and projected lost cost of goods sold (COGS). Armed with this data, you can make informed decisions based on the potential damage this out-of-stock situation might inflict on your business and the measures you're willing to take to mitigate it.

If the estimated lost COGS is minimal, it may be worthwhile to accept the situation and allocate your resources to other, more impactful matters.

However, if the projected lost COGS is significant, you can:

Utilize the "Born-To-Run" system to prompt additional purchase orders.

Verify that there are no promotional activities scheduled or underway for these products, such as retail promotions.

(Click on the image to expand)

tip

Product Tagging for Better Inventory Management

Tagging products with key information such as stock status or upcoming promotions can enhance inventory strategy, helping prioritize actions based on the product's value and upcoming needs.

Conclusion

Effectively using Amazon's forecast data enables e-commerce managers to adopt a nuanced approach to inventory management. By choosing the right forecast model, aggregating data across multiple markets, and implementing strategic purchasing and stock management, businesses can reduce both overstock and stockouts, leading to more efficient and profitable operations. This strategic approach keeps businesses agile and competitive in the dynamic Amazon marketplace.

About the author

Jordy Rinaldi

Since 2021, Jordy has been active in product management and business growth across ecommerce and technology. As Team Lead Product Management at CATAPULT, he oversees product strategy and development initiatives.