When your brand hits $25 million in sales: do you go monolith, modular, or microservices?
Digitally native vertical brands are playing a huge role in the internet-driven global marketplace. With over 2 billion people shopping online already, the eCommerce market will surpass $6.5 trillion in 2023. That’s bigger than the GDP of every nation on the planet except for the US and China respectively. Online, the boundary between individuals and products has shrunk dramatically. This market environment offers the potential for huge profits for companies that connect best with their consumers and tap into the right growth models.
Thanks to eCommerce platforms like Shopify and Magento, it’s easier than ever to enter this thriving marketplace. However, any digitally native brand that experiences exceptional growth—we’re talking $25 million and upwards in annual revenue—will face a series of challenges to overcome before further scaling to satisfy growth targets and investor expectations. These challenges aren’t about product or supply chains or economics, but rather the very platforms D2C brands have so successfully relied upon for eCommerce in the first place.
Almost every digitally native brand got off the ground relying on a monolithic eCommerce platform. While Magento’s open-source codebase streamlined eCommerce back-end infrastructure for developers, Shopify has taken that standardization process much further in creating a universal eCommerce infrastructure for businesses and consumers—a ubiquitous proprietary, one-stop shop where one size fits most. The playing field has leveled out, creating huge opportunities for brands to launch and grow. However, those same brands found a hard limit to the efficacy of monolithic models upon reaching a higher threshold of success.
Once digital-only brands reach sales of $25 million annually, the challenges inhibiting further growth are ready in wait—even more so at $50 or $100 million. Businesses of this profile want to go truly omni-channel, expand to domestic brick-and-mortar retail, and leverage online international streams to keep growing and satisfy investor needs. This roadmap leads directly to the litany of shortcomings with which monolithic eCommerce inhibits leading growth merchants. Feature scarcity, crawling speeds, data inefficiency, and a lack of interoperability in particular are high-impact pain points that are now driving the next phase of eCommerce development.
Many of the solutions being developed involve breaking off pieces of the monolith structure to optimize them via a modular approach. And it started with the head. The incredible speed advantage created by detaching the front-end and back-end of sites to implement a headless architecture is a sign of things to come. While a complete about-face towards the trend of microservices is an over-correction riddled with inefficiencies, the modular approach allows digital brands to build the eCommerce solution that works best by maximizing the benefits of monolith and microservice without inadvertently spending more time and resources on building tech than on perfecting the core business.
Features and Tech
The first problem with the state of monolithic eCommerce today is feature development. It’s no secret that development of new features has significantly slowed on major eCommerce monoliths. New features and core elements alike are not available. Additionally, the quality of the features has diminished, as has the pace of the release cycle. Just last year, Shopify added capability for video for products. That technology has been around for five years!
Big platforms are built to accommodate the lowest common denominator, but what do you do if you want to be best in class? The obvious solution is to follow the innovation and buy off-the-shelf solutions or just build it yourself. In such cases, brands have to take the tech and integrate it, a process that often adds inefficiencies and stifles speed. It’s a high stakes conundrum that has been increasing with frequency across merchants.
There do exist solutions to many problems on existing platforms like Shopify and Magento, but they’re few and far between. When you’re cobbling together your infrastructure from disparate sources and plugging them into a system that does not prioritize interoperability, invariably nothing connects well enough, there’s a lack of control over the process. It’s like hammering parts together and hoping the car will run. But if one system goes down, it could bring down a host of others—and jeopardize your whole business in the process.
Even if an integration has been built, it often functions like a hack. All the data doesn’t fit together, or you end up with additional costs incurred by patching or working around elements to get you back to square one. So often I’ve seen teams scratching their heads asking: “There must be a better way to do this?” When we go to diagram their infrastructure, you can clearly see how the convoluted backend architecture is barely holding together.
The current monolithic platforms that are out there aren’t identifying these limitations as a problem. They seem to think their current solutions will suffice. In fact, it’s a very small niche of their clients that are dealing with these challenges. When you look at the bigger picture of eCommerce, 80% of businesses are earning far less than $25 million in sales annually. Smaller companies are their core business of monolith platforms, where they make most of their money.
So even as these mid-market digital native brands are very successful, they aren’t the priority on the platforms they rely upon. Making them as such would come as a detriment to the masses of small businesses that use monolithic eCommerce platforms and are happy with simplicity and uniformity. That’s not to say they haven’t tried: Magento experimented with bifurcating into Magento Enterprise and Magento Regular, Shopify similarly with Classes. But these developments clearly have not worked.
Speed and the Monolith
In eCommerce, the most constant UX pitfall is a lack of speed. Your website is your customer experience. Without an optimized architecture, the user experience is drastically slower. Beyond 2 seconds a page and sales are lost by the millisecond. Some tests have found the average Shopify page load clocks in at over 7 seconds. The reason for this is that monolith architecture runs every single shop—from $50-a-month side hustles to $25-million-a-year brands through the same backend architecture. While it tends to be extensible and robust, it is so at the cost of speed. It’s a huge problem for monolith merchants across the size spectrum, but for the fastest growing online retailers, it can cost millions of dollars a year in lost revenue. Amazon recently calculated that a one second slowdown in page load would cost them $1.6 billion annually.
If your site runs on Shopify or Magento, you’re likely well aware that their front-end themes are more and more sluggish. This problem is getting progressively worse over time. The monolith platforms keep building on their model, which gets slower as it gets bigger. Technically, that system adds its own debt and time, and the sites in turn get slower over time. What you thought was a snazzy speedboat when you launched is more like the Titanic by the time you achieve major growth, and there’s an iceberg or two waiting for you to navigate beyond when you get there.
That’s why many of the online big boys have switched to a headless architecture that disentangles the front end user experience from the back-end tech. This implementation provides a huge speed boost, and is the only way for online retailers to ensure a load time of under 2 seconds. It’s often the first major departure fast-growing retailers take from their monolith platform. It’s a modular solution that is a sign of big things to come.
There is a wealth of amazing marketing technology that merchants can utilize to target and connect with consumers. Brands selling more than $25 million annually want the most advanced platforms available to automate customer experience personalization, remarketing, retention, and loyalty mechanisms on a large scale. Those platforms need plentiful and available consumer data to understand customers and create the most effective marketing strategy in real time. When merchants attach marketing technology into a monolithic eCommerce platform like Shopify or Magento, the data integration does not flow with the speed or functionality required to succeed. The cutting edge martech platform is left sputtering, and sales are left on the table.
Unfortunately, whatever monolithic eCommerce platform you’re using can’t get your data to all the right places. It’s just not their field of specialization, and they are significantly behind the pace of development. Instead, retailers can look for off-the-shelf solutions to get the data tagging, but in practice they also struggle to integrate. Next-gen automated personalization is the frontier of digital marketing—a no brainer for growing DNVBs to harness—but using a monolithic eCommerce platform means many are missing out on its actual value. Getting the data in the right place at the right time with enough speed is proving a major hurdle.
Expanding to Domestic Retail and International Online
There are now approaching 2000 digitally native brick-and-mortar retail locations according to Forbes, a charge led by brands with products as divergent as Tesla and Warby Parker. Expanding to domestic retail is a decision with complicated problems: How do you collect and handle customer data in the physical realm? How do you merge a customer’s online and offline profile? There’s very little infrastructure to manage this kind of growth from digital to physical, so growing brands have the option of building it themselves or piecemealing together various apps and platforms into their existing eCommerce monolith. A ton of weight is added to the process.
Even though we live in a digital world, going international as an online merchant is a complicated process. You have to tackle things like currency conversion, duty fees, foreign taxes, shipping the product overseas, re-pricing, launching third party logistics, local customer service, translations—and that’s a lot to say even in one language. It’s far more complicated than plugging in to set up your first shop.
A fast-growing DNVB only needs to maintain one online single shop with a few connected apps, and it’s up and running. Expanding beyond the monolith-approved box requires so many different systems that need to communicate and integrate, and data has to transmit back and forth seamlessly to run operations. If one piece doesn’t integrate perfectly, it creates something of an operational nightmare. Success and expansion have proven fissure points at which many wildly successful digitally native companies run into major problems in their eCommerce growth strategy—and the problem is with the platform, not the product.
The Modular Solution
To find answers, growing businesses have two main options: Hire in-house developers to build costly bespoke architecture, or engage with an agency to find a strategy. Internal devs are great at building functional tech, but the key to a post-monolith success strategy is all about perspective. Interacting with countless standout DNVB’s to iterate the most functional and sustainable architecture available has given the Anatta team a panoramic view on the problem. From that vantage point, it’s clear that solutions are emerging in one way more than others.
Current practices find monolithic solutions on one end of the spectrum and the microservices approach on the other. Microservices break down the monolith into a hundred individual mechanisms and build a solution for each component. That can work great. You can do amazing things. But companies find out very quickly that while they had one or two devs working on their monolith upkeep before, they can need upwards of fifty to put together and maintain their architecture with a full-on microservices approach. That requires huge cash flow. A 50 person dev team can cost $5-10 million dollars a year. The platform will work great, but suddenly you’re not selling amazing shoes anymore, you’re a tech company with no actual tech product to sell.
Let Andy Dunn’s words be a reminder…
The modular approach, one that we’re seeing emerge as the best of both worlds—with none of the worst of either—sits between the two ends of the spectrum. It moves you from the monolith software towards a microservice approach, with a focus on the development you need most while maintaining interoperability. You can even stay with a monolith, while only using its tech sparingly for certain core services. The intention is to break down elements of the architecture into large chunks: taxes, discounts, product information, all with their own modules. Instead of 100 microservices, you’re working on 20 modules. With this approach, you can choose Avalara for tax processing, Klaviyo for email marketing, whichever CRM works best for you, and connect them at junctures that make sense and work. A modular approach is the logical step for companies who have moved beyond the monolith approach, but don’t require the extreme detail and complexity of a microservice approach. It’s like legos. You can swap out individual blocks, change the colors, but the integrity of the structure will remain.
The move towards headless architecture is the lowest hanging fruit towards a modular perspective. It’s the first step in the process. You start seeing results immediately: internationalization, translations, faster speed. But it’s just the starting point. We’re breaking away the pieces of the puzzle at Anatta. Rothy’s, Athletic Greens, Four Sigmatic, Molekule, for example, are all moving components away from the core monolith towards more streamlined tools and utilities. A Product inventory manager, for example: Now they can manage inventory in one place rather than ten different Shopify sites, Amazon, and Google shopping.
Similarly, inventory management and order management across platforms are big opportunities to streamline in a modular fashion for companies of the profile we’ve discussed. We see review platforms as a module, fulfillment and data collection, email service providers, and others. Those are the first wave that we encounter most often becoming integrated in a way that creates a true module, interoperating with speed and efficiency. There’s more and more examples of this process. Companies like Glossier and Casper have done a great job of taking on this kind of approach, and they’re bringing in over $100 million in revenue annually.
In democratizing access to an online marketplace, eCommerce monoliths like Shopify have changed the world for the better, and helped create immense value for people all over the world. But today’s most outstanding digital brands require outstanding infrastructure, a need that is not urgently aligned with the hundreds of thousands of online stores that drive the mission of monolith platforms. Modular solutions are the logical development to the resulting infrastructural disconnect, a solution that maximizes monolith functionality with the dynamic needs of individual brands and the cutting edge in tech. The result will be a more mature and functional eCommerce infrastructure and thus ecosystem—one where they always have your particular fit and size in stock.