Cloud

The COVID-19 pandemic changed the retail landscape for sellers and consumers alike. Lockdown accelerated the pace of digital transformation and customer experience-focused initiatives. Many non-retail businesses and predominantly brick-and-mortar retailers recognized the need for online presence to engage with consumers directly and enable them to easily purchase products and services on digital channels. As a result, retailers see ecommerce as a growth engine to increase their omnichannel revenue. According to Morgan Stanley, global ecommerce is expected to increase from $3.3 trillion today to $5.4 trillion in 2026. 

Today’s consumers and shoppers expect ‘always-on experiences’ from ecommerce and mobile retail applications. Achieving a highly available, performant, resilient and stable ecommerce platform is critical to attract and retain consumers in today’s hyper competitive business world as only 1 in 40 visitors completes a purchase.

The checkout stage is the final stage in the sales funnel. However, only 1 in 3 customers at the checkout stage completes their purchase. Therefore, the checkout experience is the most critical, and retailers need to optimize accordingly.

Shopping carts represent one of the most important components of an ecommerce platform. Maximizing resiliency and latency of the ecommerce sessions are the two topics we’ll cover in this blog.

But first, what is a session?

A session is a period of time wherein a user interacts with an application. While the user is interacting with an app, this interaction data is stored both on the client side and on the server side to provide a connected user experience. On the client side, we store this data as cookies; on the server side, we call them sessions. For a seamless user experience, it’s important that these two separate data types are in sync. 

Sessions are critical to provide personalized, consistent and seamless experience that helps move a customer through the sales funnel. If the server loses the data for a given session, it cannot provide a stable experience to the client. Therefore, it’s important to save and resurrect a session’s data for unhandled exceptions and server/zone/regional failures efficiently. 

In a service-driven economy, session stability, scores, along with latency metrics, are the top two KPIs to monitor. 

1- Stability/Resiliency

High Availability, Business Continuity and Disaster Recovery are and should be the critical terms for every ecommerce platform. 

99% of uptime sounds pretty good.  What about 99.9%?

Let’s detach ourselves from the illusion of the numbers and delve into the reality.

Per one year, 99% uptime equates to 87 hours, 40 minutes of downtime, and 99.9% uptime means 8 hours, 46 minutes downtime1. For Amazon, revenue loss for an hour is over $13M; for Walmart it is nearly $2.5M; for Wayfair it is $589k, and the list goes on. Practically, for every second the website is not 100% functional, it’s losing money. The picture is even more dramatic during the holiday season.

Every dollar spent on enhancing the session stability and resiliency is worthwhile, not only for the seamless customer experience but also for the bottom line.

99.999% uptime translates to only 5 minutes and 16 seconds of downtime in a year.

*Statistics listed here are theoretical numbers measuring a retail application’s uptime

2- Latency

If a single microservice becomes a bottleneck in the funnel, that delay defines the whole experience. The shopping cart, for example, is where the transaction happens – that’s the stage that deals with operational and transactional data. Any millisecond latency there, will translate into lost revenue and a poor customer experience. Since we’re in an era where “the fast eat the slow”, latency means losing to competitors. 

Distributed applications require a distributed architecture so that they can serve up to their users wherever and whenever. Like CDNs, application servers, and database servers, session cache servers cannot be considered differently. They need to be distributed and located close to the application servers. 

Why Redis Enterprise for your session store?

Redis is an in-memory database delivering multiple data models with best-in-class performance. It provides low latency and high throughput to support responsive applications and websites. Redis Enterprise provides performance that can increase the user experience and ensure retail applications and ecommerce websites are always fast and responsive. 

Redis Enterprise provides persistence to endure server failures and is the only active-active geo-distributed Redis database based on Conflict-Free Replicated Data Types (CRDT)  technology, which provides high availability and business continuity not only among zones but across regions. Redis Enterprise can manage conflicts from individual application servers, allowing customers to have a continuous and seamless multi-channel shopping experience. 

Redis Enterprise Active-Active provides 99.999% availability.

Redis Enterprise can scale up capacity and performance in response to demand from real-time applications through its Redis Enterprise proxy, with no need to change application code and without downtime or disruption. By providing automated failure detection, failover, and cluster recovery, Redis Enterprise helps ensure that retailers can continue operating even after experiencing bursts of traffic during seasonal peaks or unexpected surges in demand.

By storing sessions on Redis Enterprise, sessions become more stable with less latency.

Solution Architecture

The diagram2 above depicts a microservices architecture for a retail / ecommerce application on Google Cloud. The incoming ecommerce traffic comes through Google Cloud Load Balancer along with Cloud Armor to safeguard against DDoS and application attacks from intruders, and with Cloud CDN to accelerate content delivery to end users while reducing serving costs. Then, the incoming requests are processed by microservices hosted on containers in Google Cloud Run scalable serverless environments. Individual microservices can scale in or out based on real-time workload requirements without overprovisioning unused resources. The product catalog, inventory system, customer profile, etc. are stored in fully managed-database services such as CloudSQL and third-party Redis Enterprise fully-managed services to provide a persistent store and a real-time session store, respectively. Finally, back office custom or packaged applications such as order fulfillment and management systems are hosted on various Google computing environments, such as Google Compute Engine and Google App Engine. Google Cloud Looker is an important component of the Google Cloud visualization suite, helping retailers to improve productivity, decision-making and innovation by delivering more insights. In all, the illustrated retail solution architecture is deployed on the industry’s most secure, performant, and scalable cloud platform. Retailers can rest comfortably running their core business on the innovative and highly flexible Google Cloud.

Summary

The retail industry has become more competitive than ever. Consumers demand ‘always-on’ engagement and if one retailer does not deliver that experience, a consumer will not hesitate to switch to another that does. A responsive, seamless, and personalized experience across all channels is now the norm. There have been many changes in ecommerce customer behaviors. Shopping carts, for example: customers have begun using shopping carts to bookmark products. They start shopping on one channel and complete on another one. That’s why shopping carts need to be rock solid and very responsive. 

For low latency, high availability, business continuity and effective multi-channel experience, retail applications and ecommerce platforms should consider Redis Enterprise. In regards to maintaining session stability, Redis Enterprise, with its active-active feature, provides 99.999% (five 9s) availability, which equates to about 5 minutes of downtime in a year. 

All major ecommerce software or development frameworks already integrate with Redis.3 If serious about retail/ecommerce platforms, there must be a factor in the resiliency and latency elements at every level of your architecture: at application level, at data level and at session cache level. 

In summary, Redis Enterprise is a core piece of technology to provide real-time and highly connected retail customer experiences. Google Cloud provides the rest of what retailers need to build and run their core business 24×7, 365 days a year, including Black Friday and Cyber Monday.

Customer Reference

Ulta Beauty, Inc. is an American chain of beauty stores headquartered in Bolingbrook, Illinois. Learn more on how Redis helped Ulta Beauty dazzle customers during lockdown and boosted revenue by 40% using data and digital. 

Next Steps

To learn more about Redis Enterprise solutions in retail or try out Redis Enterprise and Google Cloud technologies, here are a few links to get you started:

Redis Enterprise for Retail

Start using Redis Enterprise Cloud on Google Cloud

Github repo showcasing retail solutions by running Redis Enterprise on Google Cloud

Get started on Google Cloud here

Learn more about Google Cloud retail solutions

1. Statistics listed here are theoretical numbers measuring a retail application’s uptime
2. Image from Redis, June 2022
3. In the absence of a single reference for all ecommerce software and development frameworks, please see the following examples: Spring, Flask, Adobe Commerce, Kong

Editor’s note: Google Cloud’s Technical Account Managers are a rare breed. People like Dipty Chander have business understanding, technical chops, great communications skills, and an extraordinary talent, which allows them to cut through complexity and deliver customer success. As if that’s not enough, Dipty is on a personal mission to bring 1 million people from underrepresented groups into the world of cloud software writing.

What is your main job at Google Cloud?

I’m a Technical Account Manager (TAM), which means I work with customers to make sure Google Cloud understands their needs and expectations. It’s wonderful because it’s so complex. For example, I’m a TAM for a large European retailer that currently has over 100 different projects in the works, which means I’m working with software engineers, of course, but also people across several different groups at Google Cloud—including Marketing and Product Management—to ensure customer needs and expectations are met.

Is this an unusual role within the industry?

Many cloud providers and enterprise tech companies have similar positions. However, Google is unique because you can blend highly technical knowledge with business acumen to help customers make a difference in their businesses. Google also has more experience building products and easy-to-use tools than its counterparts.

Another differentiator is Google’s culture. Internally, Googlers are always interested in learning and understanding things; we judge by what we learn, and we don’t blame someone for making a mistake. This mindset creates a better environment for learning because there is respect between colleagues and a common desire to share information.

Are you saying that the internal culture affects work with the customer?

Absolutely. First and foremost, our mindset is based on understanding the customer. We ask ourselves how we would react if we were in their situation, because we know it’s difficult to work well with customers who don’t feel understood.

I know there is always a colleague I can reach out to for help in finding solutions or answering questions. It really makes a difference to feel like you can collaborate, rather than working alone. This doesn’t mean there isn’t any difficulty or bad news. For example, I joined Google Cloud at the start of the Ukraine war, which created a lot of challenges in the cloud space. And I’ve seen challenges in migrations due to misunderstandings, which can happen anywhere, but are resolved quicker when there is a strong culture of communication.

You also contribute to an amazing online project.

I am the president of E-mma, a nonprofit that supports diversity in technology, primarily through teaching coding skills to underrepresented groups. I became interested in the subject while studying computer science in college. Out of the 500 students in my class, there were only about ten women. That was striking. 

As I continued in school, I really started to appreciate the power of computer code—you can create so many kinds of applications. I asked myself, why there aren’t more women doing this? Where are the people with disabilities? Where are the people from different backgrounds? 

E-mma had about 20 people when I joined in 2014, and today there are 400 volunteers in 16 offices across France, Spain, and Belgium. We’ve trained about 130,000 people, ages seven to 77. When we launched project Code At Home during the pandemic, I thought we’d reach 1,000 or 2,000 people, but we ended up getting 70,000 people! 

Wow. But – how can you help build something like that while you hold a job?

It’s not hard to do it with my job, because of my passion for it. During COVID-19, cloud infrastructure was important for everyone around the world, specifically for getting information out, working from home, staying in touch with loved ones, and working on a cure. It saved lives. With Code at Home, we had kids telling us, “You’re giving us hope in a dark time.” Nothing is better than that. Next year, we’re doing a second version of Code at Home, and this time our goal is to reach 1 million people.

So, I get up in the morning and have a cool job at Google, while also being able to help other people with E-mma. It’s great.