IT Lessons Learned From Prime Day

Millions of shoppers flooded Amazon.com on Prime Day in July — but the online retail frenzy didn't just benefit Amazon. In fact, shoppers spent $14.2 billion across various e-commerce sites, an 11 percent increase from last year.

Amid persistent inflation, shoppers are searching for discounts, and sales events are proven to attract consumer attention. However, any serious influx in site traffic may also cause application performance disruptions for retailers (e.g., glitches or crashes) that result in inventory management issues, operational delays, and significant hits to revenue.

Looking back on past outages, major retailers such as H&M, Best Buy, and Target faced website performance issues during Black Friday and Cyber Monday sales in 2020. Additionally, in 2019, Costco's e-commerce site was down for over 16 hours on Thanksgiving Day, and site outages during 2018 resulted in over $100 million in lost revenue for brands.

Learning from these (expensive) lessons, many retailers have increased investment in their IT infrastructure before launching major online sales events. Yet incidents like the recent CrowdStrike outage show just how widespread the effects of one or two critical system disruptions can be.

To prepare for disruptions that are bound to occur during high-traffic e-commerce sale events, retailers must continuously monitor their IT infrastructure to assure network and application performance. Expanding observability end-through-end includes monitoring credit card authorizations, CRM systems, and even tracking inventory levels to ensure product availability. Continuous packet-level performance monitoring of critical retail applications is crucial for retailers to quickly triage and troubleshoot issues to ensure high-quality, positive shopping experiences from browsing to delivery.

Planning Ahead to Manage Peak Traffic

A good first step for retailers seeking to test and prepare their IT infrastructure for peak traffic events is to analyze historical patterns. Measuring anticipated and actual bandwidth requirements and usage volumes for busy periods involves analyzing historical packet data to predict traffic loads and comparing these forecasts with real-time usage during peak times. By understanding these metrics, IT teams are able to identify and mitigate potential bottlenecks, ensuring the application is equipped to handle the influx of user traffic without disruptions.

Retailers can manage the influx of traffic by integrating real-time analytics with proactive testing strategies. Testing network and application performance before, during, and after peak sale times allows IT teams to assess and maintain application quality continuously. Additionally, simulating high-traffic scenarios can help identify and address potential network and application performance issues, such as high latency or glitches. This type of testing may reveal the need to increase bandwidth, improve the availability of cloud-based resources, or realign back-end settings that do not scale to support peak-traffic events. It may also indicate the need to extend observability out to include additional systems.

Boosting Site and Application Performance During Shopping Holidays With Real-Time Monitoring 

Online retailers rely heavily on the performance and availability of their mission-critical applications to conduct daily operations and ensure business continuity, especially when they experience high traffic volumes. In fact, 70 percent of consumers say that the time it takes for a page to load influences their desire to make a purchase.

By analyzing network packet data at scale across web-facing applications and connected systems, IT teams are able to monitor the performance of their IT infrastructure in real time. This gives them visibility into the entire retail application environment, from end users to backend systems, including proxy servers, load balancers, databases, and web services. It also covers different parts of the retail ecosystem, such as contact centers, distribution centers, retail stores, and IoT devices.

These systems are designed to work together, so if one link in the chain erodes or breaks down, it's vital that retailers can quickly troubleshoot performance issues, reducing mean time to repair (MTTR) before they turn into more severe issues or outrages that can lead to customer attrition or reputational damage.

By monitoring shopper traffic, conducting host analysis, and obtaining packets during specific online sale events, IT teams can gather packet-level insights to analyze and better understand the root cause of potential slowdowns. This knowledge enables them to make the necessary changes to prevent future performance issues.

Preventing E-Commerce Performance Issues With an Enhanced IT Infrastructure

While Prime Day was a significant test for retailers' e-commerce infrastructures, if historical traffic patterns continue to hold true, the real challenge will take place during Black Friday and Cyber Monday. Past holiday shopping cycles have proven to overwhelm retailers that have failed to sufficiently prepare their infrastructure.

While many leading retailers seem to have successfully learned from past errors and invested in strategies to prevent outright outages, even minor performance errors — e.g., credit card processing or inventory management — can negatively affect customer satisfaction and hinder a retailer's success. With the holiday shopping season nearly upon us, now is the time to invest in greater monitoring and testing of the broader IT infrastructure that powers e-commerce sites. With network and application performance observability, retailers can quickly resolve potential issues to help keep customers on their sites and offer a seamless shopping experience.

Brooke Jameson is senior product and solutions marketing manager, NETSCOUT, a user experience monitoring solution.