Modern Sourcing with Microservices: An Implementation Guide

Modern sourcing is more complex than it has ever been – and we shouldn’t expect to see that change anytime soon. A host of factors – an ever-expanding universe of fulfillment methods; last mile innovations; new methods for inventory utilization; node priority; labor capacity; and more – will only continue to add complexity to the process. 

SaaS Order Management System vendors are attempting to solve the problem for retailers using AI and Machine Learning solutions that optimize on multiple attributes like speed and cost.  Ultimately we agree that leveraging AI will be part of a modern sourcing approach.   However, how and when you implement this technology is critical.  We have seen a number of AI sourcing implementations leaving retailers scratching their heads because of limited visibility into why shipping nodes were chosen during sourcing decisions. In addition, these solutions don't provide a pragmatic solution to help retailers test and learn through different sourcing configurations.    

At Nextuple we think there is a better way.  

It’s All About the Data

Sourcing decisions are only as good as the data informing those decisions. We’ve already agreed that modern sourcing is complex. Imagine how much more complex – and how quickly things can go off the rails – if you have problems with inventory accuracy, or don’t have visibility into labor capacity at nodes. How do you determine node eligibility? Which stores should be enabled for Ship-From-Store?  

You’ve got to get the data right to make smart decisions.  To ensure you have visibility into all of the important variables, and can make the best decisions for the customer and your bottom line, we encourage you to slow down and address the following five areas before you start thinking about how to leverage AI. 

1. Conduct an Organizational Assessment - Do you have the staff and the right skill sets to analyze and tweak your sourcing dials? Retailers often jump into adding new features, then realize that managing all the configuration data to support the features takes more time or different skills than they realized.
2. Analyze Fulfillment Capacity - Do you have a good view into what volumes can be driven off of your stores and fulfillment centers? What is their normal to peak capacity ratio? This information is critical for making sourcing decisions based on capacity utilization.
3. Understand Labor Costs - Labor costs vary by market. For national retailers, there can be significant swings in cost around the country. Knowing your true labor costs is important for creating cost profiles for sourcing.
4. Ensure Calendars and Schedules are Accurate - The difference between keeping a promise and delivering a day later could come down to a node’s working calendar or the carrier pickup schedule.
5. Understand Boxing by Node - Items shipping from the same location may not always ship in the same box. Knowing the box sizes carried by nodes is critical in making decisions about split shipments. 

Test and Learn

To truly make intelligent sourcing decisions in the modern age, you’ve got to be able to use real-time data. And you need to test rule sets to see what works. Moving the dial in one area can have unintended consequences elsewhere.  

We recommend a crawl-walk-run approach to implementing new sourcing capabilities.  AI and Machine Learning solutions are useful and helpful to augment capabilities, but cannot fully replace a rules-based engine. While having lots of sourcing dials is important, the ability to audit their performance, test new configurations and iterate your way to modern sourcing is key to long-term success.

Once you have good visibility into what’s happening along the supply chain, you can work through rule sets to ensure you understand their impact down the line.  We’ve built a guide to help you.  Check out our Roadmap for Sourcing.