Modernizing FeedbackNow by Forrester Architecture, Boosting Scalability and Reliability with AWS

FeedbackNow by Forrester is a real-time, comprehensive customer experience (CX) solution that continuously captures feedback to instantly improve customer experiences. Consisting of feedback collectors, an intuitive analytics engine and a powerful platform, FeedbackNow helps clients extract insights from feedback — and act on that knowledge instantly. However, the initial solution was built on a fragile legacy architecture that struggled to collect data and load balance in high-traffic situations.

After Forrester acquired the solution in 2018, the leading research and advisory company tapped AWS design partner Cardinal Peak to rebuild the FeedbackNow architecture on different parts of the AWS platform — like EC2, RDS, ECS and EKS — improving its scalability and reliability.

Imagine a set break at a sold-out concert, with thousands of fans lining up for the restrooms. The crowds quickly overwhelm the facilities. If a displeased concertgoer finishes in the restroom and tries to report the horrible conditions only to receive an error that the real-time feedback system is down, the likelihood they vent on social media instead — potentially sparking public backlash — increases significantly.

Now, apply that scenario across thousands of airports, hospitals, corporate settings, hotels, convenience stores and other important venues, and the challenges posed by poor reliability and scalability become increasingly evident. From revenue impacts and customer dissatisfaction to operational inefficiencies, safety risks, regulatory noncompliance and a competitive disadvantage, the negative impact of unreliable and unscalable systems is multifaceted and can be severe.

With physical experiences increasing in importance as customers demand issues to be addressed in their moment of need, CX leaders need solutions that provide visibility into front-line service. FeedbackNow combines in-the-moment feedback with Forrester’s leading CX analytics, focusing on improving the experiences that drive higher customer satisfaction and loyalty.

The original architecture of FeedbackNow, before its Cardinal Peak-led transformation with AWS, presented a litany of challenges that severely impacted its reliability and operational efficiency and necessitated a comprehensive overhaul.

The FeedbackNow platform initially leaned heavily on virtual machines (VMs) to host critical components like databases. As traffic surged, the system struggled to handle the load efficiently, with the lack of robust load-balancing mechanisms exacerbating this challenge. One of the most glaring indicators of the system’s fragility was the need for manual server reboots at staggeringly frequent intervals — every one to two hours. This overreliance on VMs resulted in frequent downtime, causing disruptions in service and undermining user trust.

Additionally, the complex FeedbackNow infrastructure relied on 120 servers dedicated solely to reporting and 40 to 50 other VMs further burdened by the need for frequent patching and updates. FeedbackNow’s database systems were also outdated, further straining system reliability and posing significant maintenance challenges.

Recognizing the limitations of the existing infrastructure, Forrester selected AWS partner Cardinal Peak to evaluate specific migration needs and migrate the FeedbackNow platform’s workloads and databases to fully managed AWS databases.

The meticulous and well-executed migration process spanned two years, commencing with a gradual transition of applications from VMs to containers hosted on Amazon Elastic Container Service (ECS). During this time, extensive testing and validation helped ensure a seamless transition and that the FeedbackNow system could dynamically adapt to varying workloads.

The complexity of managing multiple databases running on VMs was streamlined by migrating to Amazon Elastic Compute Cloud (EC2) instances and then Amazon Aurora, while the app that the smile boxes communicated with was moved to ECS. With a single Aurora instance capable of capturing the same volume of reports that previously required a far more extensive infrastructure, this consolidation and migration ultimately reduces the number of databases from three to a single, more efficient PostgreSQL database running in Aurora.

The transformation also extended to the software stack, with a migration from Windows IIS running PHP to Amazon Elastic Kubernetes Service (EKS) running Node.js on EC2 instances, enhancing both stability and performance. Recognizing the need for further scalability and operational efficiency, EKS adoption helps manage the increasing volume of votes.

The collaborative efforts of Cardinal Peak’s cloud and QA engineers and Forrester’s DevOps team culminated in the successful transition from virtual machines to RDS instances, allowing FeedbackNow to capture the staggering volume of 10 million votes per month without costly disruptions. FeedbackNow’s AWS migration not only addressed the reliability and scalability issues but also positioned the platform for continued growth, operational efficiency and enhanced user experiences.

Ushering in a new era of reliability, scalability and operational efficiency, the migration of FeedbackNow by Forrester to AWS revealed several tangible outcomes and benefits.

By implementing auto-scaling mechanisms and migrating databases to managed services like Amazon Aurora, FeedbackNow essentially eliminated downtime from upgrades and enabled the system to handle usage spikes smoothly. The Forrester DevOps team now seamlessly performs updates and upgrades without disrupting service, ensuring uninterrupted operations.

The transformation also led to a significant reduction in operational overhead. Shrinking the sprawling server fleet comprising 10 ingestion servers to just two medium-sized EC2 instances simplified management and resulted in substantial cost savings.

With the legacy infrastructure replaced by AWS-managed services, the codebase was more efficient and easier to maintain. And less code means fewer servers. Enhancing the overall stability and reliability of the system while simplifying ongoing development and maintenance, this modernization freed developers from getting bogged down with operations and maintenance tasks, instead focusing on deploying back-end architectures and building new capabilities.

In AWS, Cardinal Peak replaced 10 extra-large instances running EC2 and C# for smile boxes with two medium instances running Java, unlocking a considerable expense reduction. By optimizing resource usage, reducing the server fleet and embracing fully managed services, FeedbackNow positively impacted the bottom line.

AWS’ auto-scaling capabilities allow applications to dynamically adjust their capacity to handle varying workloads, leading to improved performance during peak usage times. Amazon also offers high availability and fault-tolerant solutions and a range of security and compliance features to ensure applications remain accessible in the face of hardware or software failures and help Forrester maintain a secure and compliant environment.

By embracing the cloud and enlisting the expertise of Cardinal Peak, Forrester illustrates how strategic partnerships and cloud technologies can propel businesses toward new heights of performance and innovation.

Reach out today to discuss how Cardinal Peak can help migrate your product or platform to the cloud for boosted scalability and reliability, at reduced infrastructure costs.