A large manufacturing company had an increasing number of orders that were not getting fulfilled on the day they were placed despite being made prior to the shipping cutoff time. The ordering application had been internally developed using a custom application protocol based on XML over TCP for managing inventory and shipments among its warehouses. The unknown delay in their system kept orders from being routed to the appropriate warehouse terminal for fulfillment and shipping. Because of the company's custom application and network protocol, administrators had no third-party monitoring platforms with which to view the health of their application or understand it at a discrete transaction level.
A shipping delay of one day meant they had to pay rush-shipment fees to meet their customers' delivery timeline, increasing the company's freight charges by over 20% affecting their operating margin and profits.
- Track orders from the time of customer placement to the time it left the warehouse dock, following each step without having to instrument or modify the application
- Determine where, when, and why order fulfillment delays were happening
- Eliminate rush shipment costs caused by delayed orders
Because of the company's custom application and network protocol, administrators had no outside tools or platforms with which to view the health of their application.
Using ExtraHop's Universal Payload Analysis (UPA) functionality, the company quickly gained visibility into all transaction-level activity within the application and order-to-fulfillment process understanding the cause of fulfillment slowdowns. They created an ExtraHop Application Inspection Trigger (AI Trigger) to timestamp, get access to header information in the TCP session and the XML payload data associated with each step in the process and visualize the results. AI Triggers enable event-driven data extraction and visualization of nearly anything transacting on the wire.
Every message from the Point-of-Sale (PoS) system to handheld scanners using the application was able to be parsed, allowing the company to noninvasively extract, measure and analyze every discrete transaction in the ordering process. This included activity from a runner being dispatched to the item being picked and then loaded on the truck. For day-to-day monitoring, the company also created custom dashboards to show application transaction activity, defined alerts for when certain thresholds increased above a defined SLA, and had full insight into their business operations.
By tracking order IDs and correlating each unique ID with every application transaction they could visualize which devices were in use, the volume at which they were used, how long each device took to process an individual order ID request and response, and whether the transmission was successful.
It soon became clear that the intermittent problem was actually an under-staffing issue and had nothing to do with performance of the devices, the application, or the network because each transaction in the process was completing within the SLA time.
Before ExtraHop, the customer had no insight into the performance of their custom application and the life cycle of a given order after the customer placed it. Now the customer not only understands the full life cycle of their critical business transactions but also understands the impact of staffing levels in support of order fulfillment. They also have a historical trend of ordering activity so they can better plan staffing levels at peak times and reduce staffing at nonpeak times.
For the CIO, this scenario was a perfect example of how IT could assist with improving business processes. Not only did they cut their rushed shipping costs in half, they were able to use ExtraHop's transactional analysis to remove bottlenecks in the ordering process overall. They estimated that the time from order to product on the truck has been reduced by over 20% and plan to continue leveraging ExtraHop for IT and business operational analysis for further process improvement.