New tech in eCommerce
The goals for successful e-Commerce have changed little since online trade became an exciting area to tap into it. What did change during all this past time of trial-and-error? Was it the discovery and development of new technology? If so, how to best adapt to what is technically possible to the existing strategies and course of action. We hope that with this article, we can shed some light on the matter.
Although technological advancements give a lot of advantages for growth, in any business sector, all the massive platforms and frameworks are doomed to lack the flexibility that inevitably comes with a bigger size. And that’s somewhat true – massive project usually suffers delay and imperfection simply because of the sheer number of interdependent variables that must be put to work together. The idea that new tech must be flexible to adapt to any business needs does not seem to hold much water in eCommerce.
Big-commerce projects are cumbersome and slowly adapt to oncoming demands, and usually, ones that use a solution that used to work correctly, but generally around the time of initiation, and not for long after. It seems that only a chosen few can accurately foresee how the solutions of tomorrow will meet the problems of today. Even DevOps can be a let-down, with overestimated capabilities of team capabilities and responsiveness. Some answers sound good on paper, but when the bullets come flying – the spectacular projected results quickly cave in when facing unpredictability.
With eCommerce projects of considerable size, a unique approach is needed to cope with complexity. We think that platform flexibility depends on its innate ability to remain functional even when fragmented down to essential components.
When it comes to web APIs, GraphQL seems like an optimized version of the trusty REST APIs. For example, with GraphQL, data fetching queries are simplified to the extent that makes all previous ways seem like detuned methods that are always prone to over- or under-deliver. This situation creates an unnecessary burden on information flow, and available technical resources necessary to provide a smooth process execution seem to choke. GraphQL can fine-tune requests by extracting and presenting the exact data needed by the client, without needlessly flooding transfer volume with unnecessary data, as REST APIs often does. We addressed the future potential of new techs like GraphQL and others further down this article.
With the help of infrastructures like IaaS and a serverless backend for eCommerce, company teams can focus on business logic instead of indulging in stressful and time-consuming manual maintenance. Serverless tech also allows for applications that are part of managed services. With performance-oriented new techs like Nuxt.JS and Vue.JS, online commerce solutions gain unique advantages in scope and depth of influence. Without a limitation presented by a dedicated server, companies can manage their projects based on the actual resources they need, instead of paying for pricey solutions that never get utilized optimally. Because of the smooth resource management of a serverless approach, infrastructure costs can be cut down, even though the initial investment may seem steeper at first glance.
Additionally, with the power of managed services, developers can use less code to connect their apps and services, apart from the smooth up-time management and high return of investment that comes with it. Serverless is also better than the Cloud, because, with it, the price depends on the actual time of code invocation. In contrast, Could Services are known to run an idle time and potentially wasted resources. Scaling up is also incredibly easy with Serverless. With solutions like Lambda, when your e-store gains increased traction, it will automatically add up all the necessary additional functions and firepower to support them.
The fragmented nature of platforms managed services comes as a result of the notion that a feature-driven approach does not hold the flexibility necessary to run a smooth eCommerce. Choosing the right store platform becomes a matter of flawless connectivity between each framework component.
A few years ago, SaaS solutions were based on conventional platforms run on good old PHP or .NET, with low ownership and investment cost. However, when business owners went to try a different strategy, the online tools could only allow changes possible by the abilities and limitations of the admin panel. Enterprise solutions of the past were rigid to fundamental change required by business needs, and APIs were not even standard terms in developers’ vocabulary.
Feature-based solutions — as opposed to architecture-based — began to appear as an inferior approach by the time smartphones became popular and replaced feature phones. After the phone tech switch, developers had the freedom to provide improved customization that was not possible before. With the emergence of cloud tech, the API-first method became the right way to go, with their ability to handle interchangeable interfaces. This is especially evident in present time enterprise solutions for eCommerce. Many platforms like Magento, Shopware, Big Commerce rely on high-performance components. Still, the crucial part for all of them is how these components can best stick together in ways that also maximize the efficiency of team skills and effort.
The Emergence of Shopware
The eCommerce of today brings so many requirements that developers sometimes focus more on R&D than to provide the IT services they build and support. New demands in software development focus a lot more on the ‘how’ than on the ‘what.’ This radical change in approach prompted the creators of Shopware to present businesses with a completely redesigned platform, where common eCommerce denominators are defined, and code architecture has adopted the highly beneficial API-first method.
API-first, although a term that is slightly overused today, is, in fact, a fundamental approach in eCommerce that is the result of years of research and development. When an online platform intermediates existing services and processes using the application interface, it’s not always possible to go about doing that without re-thinking the full stack of existing tech, and the efficiency of how components communicate with each other.
Shopware is designed to accommodate existing business logic by automatically providing the corresponding endpoints for APIs. The default endpoints can be further modified to reflect specific intentions in terms of data display and management. There are two main APIs that Shopware provides – one for the presentational part accessed by clients and one for the admin section. The first one – so-called Sales channel API – is meant to handle data related mainly to orders, user registration, and the checkout process. The second API is intended to manage data exchange in the back end and is used in processes like data synchronization between the admin and the front-end channels.
Shopware’s ‘blood type’: Symfony & Vue
As an open-source platform, Shopware relies on the help of authentic and robust frameworks to tie things together. Symfony, one of the most popular frameworks for reusable PHP components, works as the backbone for essential operations like caching and session handling. Shopware creators chose Symfony, not only because of its undeniable popularity but based on their long years of personal experience with it.
Looking at what makes JS framework explicitly a popular tool is a bit useless because trends change rapidly, and are an unreliable factor for accurate predictability. jQuery, Angular, or React.JS are widely recognized and used now, but they might soon give way to intricate newcomers. Like Svelte, for example.
Svelte is a compiler, and because of that, the dev community is reluctant to label it as a “framework.” Still, its ease of use is similar to Vue.JS, both of which are noticeably more convenient to use than React.JS or Angular. Svelte does not hold framework references and compiles JS, HTML, and CSS into tiny modules. It doesn’t have the popularity it deserves yet, but it has a few benefits to consider - it uses less code and offloads a lot of work that otherwise would put a strain on the browser. We believe tools like Svelte could reach the notoriety of React, Vue, or Angular soon, even overtake them and become the next big thing in eCommerce.
While Svelte is heralded at unifying CSS, HTML, and JS, Vue.JS helps developers separate these three components, utilizing the benefits of each much better during development. Rookie programmers can quickly get the hang of it because Vue does have a gradual learning curve. What’s more, the community has created a friendly environment where newcomers are welcome and to quickly tap into a lot of circulating knowledge and experience.
Vue has the compound reach of Angular, but the stability and SEO-friendliness of React. Most developers with a background in Magento, WooCommerce, Shopify – even Shopware – have begun to realize the incredible productivity potential for practical eCommerce that comes along when Vue.JS.
When compared to React, Vue is marginally easier to master, and fun too, even for junior developers. Due to its simplicity, detailed documentation, and minified syntax, Vue.JS is perfect for building single-page apps. But more importantly, it also has proven potential with more significant projects, without the danger of sacrificing the performance of efficiency.
With supporting tools that pair fantastically with Vue.JS like Nuxt or Vuex, customizing solutions is only a matter of master-minding the correct approach. With improved meta tags management, Nuxt-based projects have an SEO advantage. What’s more, Nuxt supports standard SPAs with smooth and clean navigation and manages the incredibly useful options for social sharing. Promoting pages or products is comfortable with Nuxt. It creates an environment where social networks and app’s content communicate clearly with the search engines, for maximum reach and exposure. Competitive advantages like these are what help marketers gain more clients and increase conversion. The effect of the SSR method further amplifies quick and smooth navigation. Server-side rendering allows for processing logic to be executed asynchronously so that the end-user can enjoy clean interactivity on the storefront.
If Vue.JS shines in more modest projects, Angular is meant to supports enterprise solutions by design. With the potential to render complex and big projects, this framework is naturally a bit more challenging to master too. However, its architecture is made to complement code quality. Angular is excellent when employed by whole teams since individual coding mistakes have less impact on the finalized output. This type of workflow optimization is possible with JS supersets like TypeScript. Angular fits well in solutions meant to satisfy the requirements for big and demanding projects. Since it already has rich and reliable libraries on speed dial, it rarely calls for any third-party sources for additional support. It does the job without much external aid. In the world of application frameworks, Angular is the big wolf that does the job by himself.
React was born out of the need for Facebook to deliver modern web experience for their users, and Facebook is also using it for eCommerce. React has initially been shaped as a library, but the complexity and abundance of auxiliary tools have upgraded itself to a status of ‘framework.’ With such rapidly developing tools like React, labels and nomenclature are obsolete – it could easily be the preferred method for eCommerce in a few years.
One of the noticeable advantages of Reactor.JS is its ability to selectively modify what’s necessary for web content, using the tree-like HTML DOM model to display data. React is very popular, and like Angular, it is the preferred tool used for massive projects such as some current and near-future examples by companies like Facebook, Netflix, or Microsoft.
Among the exciting options for the JS framework is the creative-force driven Ember. As a front-end dedicated model, Ember is excellent for scalable web projects suitable for both desktop and mobile. Its creators tried to provoke the creative potential of developers by allowing them to create unique app features, with code that is not only app-specific but less in volume.
What’s next for eCommerce
One of the most substantial benefits of traditional shopping, and, simultaneously, one of the most apparent weaknesses by eCommerce, is the opportunity for customers to feel and see the products. Some shoppers will always prefer to touch the items they intend to buy. One of the potential deal-breaker in that regards is Virtual Reality. Its unique try-before-buy setting presents opportunities for customers to enhance their item browsing experience and bring it as close – if not more – to the IRL brick and mortar shops. Nike, for example, has an interactive app that makes a 3D imprint of your foot, making it easier for you to check available selection that matches your exact size and shape. This is an example of how eCommerce should also work: as an auxiliary tool that helps the clients use online tools for later offline shopping.
The touch part is almost unachievable in eCommerce. Or, at least, it is a bit outside of the technically possible right now. However, eCommerce has some aces in the sleeve to battle the negative sides of digitalization. The usual suspects are, of course, the standard offers like targeted coupons, timely discounts, free shipping, and fast delivery. Sometimes, online stores must list products at lower prices than their outlet counterparts, just to stay competitive. But there’s more going on beneath all the layers of enticing features that make eCommerce a preferred shopping method. Some shoppers who habitually visit malls to check the things they will buy personally, also use their smartphones to compare and review products before getting them, usually from the comfort of their home or on the run. Such examples are hinting that the future of eCommerce will undoubtedly preserve and work as a promo tool for traditional shopping.
When it comes to new tech in eCommerce, few things seem to pose influence with undeniable certainty. For one, marketers must pay attention to trends like customization, mobile commerce, and a multichannel approach to sales. These factors are all parts of business models with proven success. For example, algorithms for personalization are the preferred method for Netflix users when they choose what to see next.
Successful online business owners are those who pay attention to supreme customization and their ability to extend their influence and reach without having to re-think their platform or strategy. Product recommendation is part of such plans. We have recognized two models that companies use to present their clients with a natural continuation of their shopping journey. The first, called collaborative filtering, focuses on similarities between clients’ preferences and habits. This method can be seen on eCommerce sites under sections titled ‘what others also viewed.’ The second method—content-based filtering— examines the personal viewing history and uses it to create suggestions for similar products, categories, or articles. We think what works best is usually a combination of the two, where personal preferences are cross-referenced with what’s trending.
Chatbots and voice assistants are gaining notoriety, as AI gets better and better each day. In eCommerce projects with rich inventory, a non-human interaction can be quite useful to help clients reach what they seek quickly, everyone included, always – an impossible task for a human being. Some of the simple chatbot functions resemble nothing more but interactive FAQs. The more advanced ones support contextual help based on a massive database of possible scenarios. Contextual help is applicable in projects where there is a lot of information that needs to be analyzed, like Amazon (Alexa) or Google Analytics (Insights).
The generation gap is always relevant for considering the buying potential of customers, because it includes everyone, but separates them in simple groups with contrasting differences. For example, knowing the shopping patterns of baby boomers compared to the up-and-coming millennials is relevant to how eCommerce changes now and in the future. While people below thirty don’t have the buying capacity of their parents, they are the ones who value personalization and use social media when shopping. Additionally, millennials already account for approximately one-third of all shoppers worldwide and do most of their shopping online. On the other side, boomers are known to be less picky, and they can leave satisfied with off-the-shelf products. Gen X, being the liaison between the old and the new generation, naturally characterizes with features native to both, although some sources claim they lean more on the conservative side of the scales.
Voice assistants are powered by cutting-edge tech, and they may require quite extensive prior work and research, but their scope and convenience leverage a noticeable impact on connected experiences for shoppers.
Headless: a path to infinite flexibility
Headless Commerce has gotten a lot of attention in media and tech blogs. This tech aims to separate the dependency between the backend services and the storefront user environment. With Headless CMS, backend developers can work independently from the UI-experts. The latter can enjoy the liberty of re-shaping the storefront in any way, without having to mess with the admin section. One sure benefit of a headless backend is that you can have different APIs and display the front-end using a single channel. On the flip side, this sometimes is only doable if you happen to have the right team for the job. Your front-end designers need to have the knowledge and understanding of how Headless integrates with necessary services.
With Headless Commerce, you don’t need to worry about platform updates because the APIs who manage data flow is always guaranteed to run their latest version, and update automatically.
APIs have the flexibility to use a code that works independently from the code used to integrate them.
This relation means APIs rarely break down compared to components run on the traditional eComm platform. With headless, you may need to do some hardware upgrades and build the new tools, with the help of a skillful team. But once everything is fit into in place, it works with the flawless determination of clockwork click. The headless approach provides much better control for upcoming integrations and enhancements and is one of the most influential techs for future eCommerce shop houses.
What’s more, Headless projects are not dependent on existing environments or service provider. If you decide, you can resell your project to customers or lend it to a business partner. The new owner does not need to worry about losing any of the existing functionalities. Whatever further changes and upgrades are due, they can be done using a plethora of open-source complementary techs and frameworks. Here are some of them we think are legit candidates to shake things up soon in the E-commerce arena.
Nuxt.JS is a Vue-based framework designed to create and support contemporary web apps. It’s versatile and has a developer-friendly syntax inherited from Vue.JS. Nuxt has a robust architecture that is incrementally adaptable – single-page apps, as well as enterprise-level projects, are both within its scope of influence. Versatility extends far beyond scope and complexity – with Nuxt, you can build apps based on server or serverless environment, and even have the luxury to switch between the two without trouble. Connections to endpoints are possible through REST and GraphQL, and even PWA integrations are only one integrated module away.
With the static generated application backed by Vue, Nuxt can also be used as the backbone for the increasingly popular microservices models for eCommerce. The performance side of Nuxt for even big online stores shines through the HTML-generated content, easy to pick up by search engines, and fast to display with tools like CDN hosting.
This technology represents a new paradigm for eCommerce and is powered by an API-first approach. Its optimized data fetching and smart syntax make it an increasingly preferred tool than alternatives like REST. GraphQL-based API returns only the data requested, unlike REST, where a lot of unnecessary data may float around and deplete resources. GraphQL is designed to not under- or overdeliver, which is not only an intelligent data delivery practice but saves precious resources that can always be utilized elsewhere. It connects data from multiple endpoints in a single go, improving performance by a considerable margin.
GraphCMS is the debut platform to include GraphQL framework, and as such, it offers minimized payload and elastic interaction with APIs. This backend system can support high-traffic web stores and covers everything above a mid-size business. The role of GraphQL in GraphCMS is indispensable. It facilitates headless content management and promotes team collaboration. With GraphQL APIs, you can scale-up effortlessly, and aggregate data input coming from multiple sources, including user- and AI-generated. It has never been easier to handle and distribute content with many origin points.
When eCommerce entrepreneurs want to adapt their content to match the requirements of a specific device, they consider changing their CMS altogether. And sometimes it is the only option. Not only that, but they need to patiently wait and apply the updates that will come when these devices roll out their next firmware upgrade. This harsh re-course is avoidable with GraphCMS. You can build the application content, meant to be customized, modified, and distributed through several channels without having to change the tech or process flow from the ground up.
When a project is initiated, a lot of factors involved must be cataloged and universalized between all team members. This must be done to make sure that when the initial building blocks are put down, it will be with the idea they must become solid steppingstones for later improvements. As a developer or a software company, repeatedly doing the same project foundations is a waste of time. Therefore, developers have the habit of using established methods and techs that already hold a proven record for intended needs. Fortunately, the same principle of reusing what works also applies to UI designers and their working habits and environment.
Design systems are meant to make it easy for front-end experts to work using standardized elements like layout colors, shapes, page elements or typography. The use of UI-libraries makes it easier for interface creators, especially in a team environment, where standard practices or specific approach can quickly be followed and applied, and common problems - eradicated with a single blow.
If you solve a problem from the developer’s perspective, you must think about saving the repetitive work for one whole department. Reusable components must be part of a unified design system where the collective effort utilizes the power of each unit. To build a design system that accommodates for everyone involved, you can use solutions delivered by the help of frameworks like Storefront UI, which relies on the collective force of a whole community. And with component libraries like Storybook, you can let developers test them without having to adapt them via code modifications. The high usability of UI is meant to benefit the end customer and their shopping experience. Still, with design systems, the benefits of usability are also available for the actual creators themselves. Clients expect great customer experience; why shouldn’t the makers of this experience also be a part of it?
When new dev tools hit the market, it’s the programmers who are the ones who not only give a first evaluation but the most relevant one too. That is true not only because they are the ones who understand it best. The greatest tools are the ones that benefit everyone, not just an arbitrary part of the whole system. For example, WYSIWYG was not developers’ favorite since day one, but it tried to be. Editors—who are also non-developers— could not only edit the text part of the content but use the powerful inbuilt editors and make a mess of the code. WYSIWYG was the industry’s attempt to unite talented teams and help them work together better. But it somehow failed to deliver, despite continually being improved. One day, maybe.
The problem with content editors is that they usually come with too much power that content creators often unintentionally misuse and create a mess (which developers later must fix). We use a term to describe such circumstantial incompatibility: separation of concerns. Following this concept, content modifications can be done separately from what gets rendered on the page. This separation is achieved through a mediator like JSON or WikiText.
Data and Content
Clean and SEO-friendly content comes through static HTML pages. This is also how editors and developers can best modify and deploy code or content changes without conflicting each other’s work. Generated HTML is static by nature, but the data sources that put it on display are quite dynamic. Because data services are intense, almost every app relies on server-side action, usually, thanks to the intelligent design behind new techs like GraphQL or REST.
Intricate new community-based projects like Jamstack is what highlights the benefits of prerendered files. This concept is about the flawless interaction between the static and dynamic content that most websites have, most eCommerce websites. Jamstack is not concerned with a specific technology but is an attempt to modern build tools and practices and make a better sense of the seemingly different stack components like JS, API, and HTML.
Finally, static content is much safer than its dynamic counterpart. Flexible APIs dynamically update the endpoint and effectively remove the chances of hacks.
Event-Driven Architecture (EDA)
The core difference between Service Driven Architecture and EDA can be exemplified between these two words: request and response. While SDA uses tech like REST API to get the job done, EDAs break down all processes to single events. The only way to reverse an event is by launching another event. This chronologically bound structure also serves as a memory ‘journal’ where a list of things to happen depends on what already took place. EDA reduces the question/answer time between the client and services and has a processing model able to run multiple services using events as a trigger for a controlled chain reaction.
Here’s one example of how Event-Driven Architecture relates to eCommerce. When a customer purchases an item, a few consecutive events are set in motion. Still, for the shipping and logistics department to receive the green light for shipping the item, they need a confirmation that the customer already paid for it. Applications that use this architecture run on components whose relationship is very linear in nature, and hence easy to understand and follow. For the asynchronous part of internal communication, EDA uses something called event broker: the middle man who makes the necessary adjustments to keep up with the code logic.
One of the glaring differences between SOA and EDA is that when the processing logic needs change, you need to adjust the API service endpoint (SOA). On the opposite side, EDA can add additional events to the mix without an explicit definition. Also, the subscription-based model of EDA sends events towards the clients, compared to SOA, where the clients make a request and wait for feedback. EDA can complement the dynamic nature of eCommerce content, and we will see more of it soon.
The new technology in eCommerce does bring a recognized pattern in to view – new tech is always more complicated.
Complexity is the worst enemy of security.
Fortunately, technologies like PWA prove that some new and complex tech already comes furnished with substantial security measures and protocols.
When it comes to future tech in online shopping, what separates the winners from the losers have a lot to do with the knowledge to recognize a suitable tech, and the guts to trust it unconditionally.