What’s most interesting about this year’s AWS re:Invent is the sheer vastness of the platform and capabilities.
In-person events rule.
The fatigue factor had definitely set in doing remote conference events during the pandemic. It was delightful to see the 11th annual AWS re:Invent back in full force with over 50,000 registered attendees (that felt like over 100,000 people) buzzing around over 10 different venues to experience over 2,000 session across 50 content tracks. The conference closed out earlier this month.
AWS re:Invent has always been a tech-forward conference riddled with new product announcements, rich keynotes and a comprehensive catalog of sessions led by practitioners and customers. What’s most interesting about this year’s event is the sheer vastness of the platform and capabilities. Additionally, the technical ecosystem that has been created surrounding AWS platforms is very extensive from AI/ML, DevOps, Observability, Security, FinOps and many other areas that support AWS services. Fortunately, the floor was organized into neighborhoods to make it easier to find companies of similar capability.
Applications are rapidly becoming more and more loosely coupled with the ability to assemble and integrate new capabilities independently at speed like never before.
The AWS platform is a perfect example of this as it innovates at such a rapid rate and continues to expand into higher value managed offerings such as managed services (AWS Managed Services), integrated offerings (Amazon Redshift integration for Apache Spark ), SaaS offerings (AWS Supply Chain, AWS Connect) and many others.
The future of application development, specifically as it intersects data and customer experience, may never be the same again. Here are some key considerations for building a strong data and architectural foundation to solidify the future to deliver next generation experience for your customers.
Harnessing Data Proliferation Is the New Foundation of All Businesses
The vastness of a system is often underestimated and sometimes inconceivable. To fully grasp the magnitude of complicated, interconnected systems, continuous exploration with complimentary and evolving tools is needed.
Data exploration is becoming like ongoing space exploration. New discoveries are ongoing an evolving, for example, in star systems like the Eagle Nebula whose Pillars of Creation was first photographed in 1995 by the Hubble Telescope and now has been enriched by the James Webb Telescope, where new insights are now being unleashed through new data and images on star creation and many other new patterns of the universe.
Just like space exploration, data is becoming so vast, it’s nearly impossible for any single set of tools to deal with all the complexities. While new statistics on data production are appearing all the time, the reality is that in the next five years, more data will be produced since the dawn of digital age. To deal with this complexity, you need a variety of integrated data tools, real-time integration, proper governance and actionable insights.
AWS has assembled one of the most comprehensive set of cloud-based database platforms. There are eight purpose-built, non-relational databases and five relational databases including Aurora, which has become the fastest growing service in the history of AWS. It combines the performance and availability of traditional database with the simplicity and cost effectiveness of open source databases.
There is also a full set of serverless analytics services from large scale data processing with EMR, real time streaming data with AWS managed streaming for Apache Kafka or RedShift for a fully managed petabyte scale data warehouse.
Machine learning and AI services such as SageMaker is being used to train models with billions of parameters to make more than a trillion predictions every month.
Many companies are already turning these complexities into opportunities. Expedia is processing 600 billion AI predictions per year powered by over 70 petabytes of data. Pinterest stores over 1 exabytes of data on Amazon s3. Samsung’s 1 billion users make 80,000 requests per second. Netflix processes billions of traffic flows and processes terabytes of log data each day. Dow Jones has modernized their ML models to determine the best time of day to reach customers which has improved subscriber engagement rates by a factor of 2.
Data is now at the center of every business, powering critical decisions and enabling rich customer experiences.
Related Article: A Decade of Dramatic Change in Digital Customer Experience
Integrating Data Rapidly Provides Faster Insights
Integration has long been the bane of any data solution with ETL (extract transform load) being referred to as a “thankless, unsustainable blackhole”. AWS has spent time to make integration easier with federated query capabilities in Redshift and Athena to enable running queries across databases and clouds without moving any data.
AWS data exchange allows seamless integration of 3rd party datasets with your own data in Redshift with no ETL required. SageMaker has been integration with Aurora and Redshift to enable anyone to access machine learning models. The vision is to ultimately have a zero ETL future. Aurora and Redshift now have a fully managed, zero-ETL integration that allows consolidated data from multiple databases to be seamlessly accessed. This unifies the world of transactional data and analytics capabilities eliminating the need to write custom pipelines between Aurora and Redshift.
Similarly, AWS added Redshift integration for Apache Spark to easily run Spark queries on Redshift data from EMR, Glue and SageMaker within seconds. All of this without the need to move any data to S3 or manage any connectors.
These are significant steps forward to a zero ETL future.
Enabling Real-Time, AI Powered Customer Experiences
The immense amount of data and the future of data integration is creating a perfect storm for AI/ML models to become smarter to make better decisioning and predictions to further personalize customer experiences.
Expedia is the business of creating memorable experiences for its customers. Their platform is providing a lot more than coordinating travel transactions and has become a way for its customers to experience and learn about the world.
Expedia’s reach is vast with over 168 million members, 50,000 B2B partners. 3 million properties, 500 airlines powering travel in over 70 countries. They have amassed over two decades of travel behaviors, booking patterns, partner activity and much more. It’s fair to day data is their number one competitive weapon.
Rathi Murthy, Expedia Group CTO, provided several insights to their approach in using AWS to power these customer-first experiences. To leverage the data they collect, Expedia has made significant investments in AI/ML and now process over 600 billion predictions each year powered by over 70 Petabytes of data. They also are enabling extreme personalization by creating over 360,000 permutations of a brand’s page so travelers can see what’s most relevant for them.
According to Murthy, to accomplish this they had to modernize their data and application infrastructure by first transforming to a containerized architecture using Amazon EKS and Karpenter. Next, they had to figure out a way to process millions of images and reviews with sub millisecond latency by leveraging Amazon DB and SageMaker. Finally, they needed to create a new self-service platform that hosts over 29 million virtual conversations saving over 8 million agent hours while improving customer response times.
Customers that are backed by tools and a strong data foundational can innovate on their customer experiences more readily than their competition.
Related Article: 5 Digital Customer Experience Trends for 2023
Moving Toward a Multi-Dimensional, 3D World
The world is multi-dimensional. But today, our digital systems mostly live in a 2-dimensional world. The next wave of customer experience will come from evolving these digital systems to support multi-dimensional online experiences. 3D will soon be as pervasive as video is today.
Online shopping is just one obvious example where a 2-dimensional experience is subpar and creates inefficiencies. Without having a clear way to anticipate fit, online shoppers will frequently order multiple sizes and return what doesn’t fit. How much easier would it be if retailers presented 3-dimensional images to their customers that enables them to visualize how the product will actually fit?
The technology is now emerging to support this. WebGL, a 2D/3D web graphics library initially launched in 2011, has now matured to provide rich interactive web pages, games and applications directly in your web browser. The quality and complexity of visual elements that WebGL can render is far greater than what HTML and CSS can achieve.
Amazon has released a Virtual Try-On capability for shoes which uses augmented reality allowing you to visualize every angle of the shoe before you buy. Ikea, Pottery Barn, Warber Parker and many others have been early adopters of this technology.
Creating 3D objects is not easy as there is tremendous effort involved in generating a model and converting images. A 3D model is worth a 1,000 pictures. To assist, there is a technology called Photogrammetry that helps automate the process of using 2D images to create an approximation of a 3D image. An emerging technology called Neural Radiance Fields (NeRF) is a neural network that can generate complex 3D objects based on 2D images by using machine learning to create “radiance field” which is the model of how light is reflected by the scene. These tools and techniques are making it easier to realize 3D objects.
In addition to managing the 3D images, having a 3D engine that provides a real-time virtual environment to present objects is critical. O3DE is an open-source 3D development engine for game, simulation, and multimedia creators that works in the cloud. Others include Unity and UnrealEngine from Epic Games.
Platforms for creating immersive experiences are also emerging. Matterport allows you to use your own capture device (mobile, camera, etc.) to make a 3D representation of your room or environment. This is being used across industries such as real estate, retail, construction and facilities management.
Bringing a full 3D experience together requires the fusion of models, sensors and data. Systems need to develop spatial intelligence to successfully perceive and derive insight from visual data. This cognitive process is known as an aptitude for understanding visual information in the real and abstract word as well as an innate ability to envision information.
The future of online experiences will soon be multi-dimensional, fully immersive 3D experiences. AWS is at the forefront of providing the capabilities to delivery this.
Using Event-Driven Architecture to Deliver Scalable Experiences
Synchronicity is an illusion. The world does not operate in a synchronous fashion. Activities in our everyday life all happen in parallel around us, and as humans, we expect to operate asynchronously. But most of our systems and applications today are monolithic and function synchronously. This limits the real-time nature of the system, the ability scale and respond to peak demands. It’s impossible to meet your customer’s experience expectations with applications that only work synchronously.
The solution is asynchronous, event driven architectures. Asynchrony leads to loosely coupled systems. And loosely coupled systems create fewer dependencies, isolate failure points, and provide for an evolvable architecture. This allows complex systems to evolve and not worry about creating an end-state architecture at the start. Event driven architectures lead to loosely coupled systems that enable global scale.
Consider AWS S3 (simple storage service) which started as eight separate microservices when first launched in 2006 and is now over 255 microservices and has never stopped running. That’s evolvability. Delivering next generations experiences requires you to create event driven architectures.
Trustpilot is a digital platform that allows customers to review a business from which they’ve purchased a product or service or contacted customer service. They help consumers make trustworthy decision on who to buy from. They get over 1 million reviews each month on their platform and now have over 190 million review on the platform.
Angela Timofte, director of engineering at Trust Pilot, discussed how their journey moving from a monolithic system that limited the experience of their customers to an event driven architecture to build reliability, flexibility and integrity into the system. Using DynamoDB, Kinesis, SQS, and Lambda allowed them to evolve the platform to perform critical functions like enrollment, publishing, fraud detection, compliance. Most critically, they are now delivering better experiences to their customers and enabling them to make trustworthy decisions in real time.
Don’t Let Your Past Define Your Future
The US Standard railroad gauge (i.e. the distance between the rails) is 4 feet, 8 1/2 inches wide and is still in widespread use today. This was derived from the English system which was established because the people who built the railroads used the same tools that they used for building wagons. This spacing for the wagons was preserved a so that they would fit in the ruts of the very old, long distance roads.
Those first roads were built by the Roman Empire to provide needed transportation to expand the empire. They traveled by horse drawn chariot that were made to the same specification which were designed for a chariot to be drawn by two horses. The width of the chariots’ wheels needed to accommodate the widest part of the horses — their behinds — which turns out to be 4 feet 8 1/2 inches. The ruts in the road just happened over time creating a cascading effect on infrastructure design for years to follow.
What’s most impactful is these design decisions didn’t stop there. The correlation continues even further to the Solid Rocket Boosters (SRB’s) attached to the booster rockets attached to the sides of the Space Shuttle’s main fuel tank. The SRBs are made by Thiokol at a factory in Utah. The engineers who designed the SRBs might have preferred to make them a bit fatter, but the SRBs had to be shipped by train from the factory to the launch site. The railroad from the factory runs through a tunnel in the mountains. The SRBs had to fit through that tunnel. The tunnel is slightly wider than a railroad track, and the railroad track is about as wide as two horses’ behinds. So a major design feature of what is arguably the world’s most advanced transportation system was originally determined by the width of a horse’s behind.
It’s not too late to change your future.
