Taking Everything Apart: Composable Commerce

By Yelena Shkolnik

In our recent article on the evolving shopping experience, we envisioned a future where shopping happens at the point of inspiration, where Cyber Mondays become as delightful as Black Fridays, and where experiences themselves are merchandised. Something substantial gets in the way of achieving this dream – the site architecture that houses these ecommerce experiences. In this post, we will focus on the software layer that can enable merchants to weave this “commerce everywhere” magic, what’s been holding back the revolution and where we think the investing focus should be.

Hitting the Jargon Head-on

Currently, most ecommerce platforms are monolithic. This means that the site’s frontend is tightly coupled with its backend. To make a change to a site’s frontend, the change also needs to be reflected in the backend, which makes it time-consuming and costly for developers to create new consumer experiences across new channels.

Monolithic architectures are fraught with challenges. For one thing, it is hard to customize them without ending up with a mountain of technical debt. New releases are a challenge given that code burden and the challenge of developing in a waterfall approach. With a monolith a retailer is likely to only deploy new code once a quarter. That infrastructure can also be hard to quickly scale up if, say, a pandemic drives ecommerce up 10x overnight.

As retailers try to deploy omnichannel experiences across voice, chat, in-store, online, etc., a monolithic architecture makes it hard to flow all of those experiences into a consistent checkout. Most retailers will end up trying to masking-tape together their ERP systems and a variety of ecommerce solutions, challenging their ability to see real ROI out of those efforts.

But a harmonized omnichannel experience is more important than ever. Buy Online Pickup In Store services have surged, with 40% of retailers now offering the service, compared to just a quarter last year, according to omnichannel platform NewStore. More than half (55%) offer in-store returns of online purchases.[1] More important than this general trend, COVID has certainly taught retailers the value of being nimble. The ability to seamlessly transition a majority of revenue to a new channel overnight, to rapidly adapt messaging, to quickly innovate promotions or generally respond to world events, has never been more obvious.

And beyond all of that, monoliths generally mean slow sites. And in a world where apps such as Amazon and Instagram have conditioned customers to virtually instant experiences, site latency issues are a big deal. 79% of customers who experience poor performance will not buy from an ecommerce site again, and 53% of mobile users leave a site that takes more than 3 seconds to load.[2]

Enter composable commerce.

The shift to composable commerce is a part of the broader JAMstack movement in web architecture. Unlike traditional architecture where every page request sends the server on a new database query mission, JAMstack websites serve static files immediately from a Content Delivery Network when a request is made. When this JAMstack technology is combined with Progressive Web Apps (PWAs), the result is a mobile store experience designed to look and feel like a native mobile app, right in the browser. The frontend rendering of the view moves to the PWA itself while the data describing the content and products is called through APIs connecting the frontend and the backend.

While the concept of a loosely coupled frontend has been around for some time, a microservices based architecture takes the headless approach a step further by allowing brands to select multiple best-of-breed services and stitch them together to create a completely custom experience. Composable encapsulates headless, or the type of architecture that separates out the frontend from the backend business logic and data functions, but also describes the broader re-platform to leverage microservices.

Composable commerce gives brands the flexibility to innovate with their frontend and create differentiated experiences for their customer. Retailers that adopt a composable structure stand to come out far ahead. According to Gartner, organizations that have adopted a composable approach will outpace competition by 80% in the speed of new feature implementation.[3]

The Magic of a Composable Stack

Why should brands go headless instead of sticking with the status quo? Let us look at a few scenarios where the headless charm might come in handy.

Scenario 1: The technical lead of a company has been hearing rave reviews about a new CMS service in the market but the company’s contract with a monolithic vendor prohibits him from trying out the service for himself.

Instead of relying on one system, the tech lead can cherry pick the microservices that will solve the exact needs of the business, creating a more efficient technology stack.

Scenario 2: The Sales head of a fashion retailer is super excited about this Black Friday sale as it would allow her to report some face-saving numbers to her CEO after the year-long slump. When the D Day comes, shoppers go berserk and the order volume grows 10x. But three hours into the sale, the site crashes.

In a monolithic application, all the resources such as datacenter and network are one stack and there are limits to scaling a single stack. On the other hand, when you break up an application into multiple vertical stacks, each piece (the website in this case) can be scaled independently without slowing down the entire system during times of peak traffic.

Scenario 3: A luxury clothing brand known for its in-store experience decides to open an online store. A customer looks at the collection online and adds a suit to the cart. He then visits the store to try it on before buying it. However, the store associate struggles to find the suit and disappoints the customer who leaves without making a purchase.

Headless commerce ensures that every known detail about a customer’s past purchases and preferences is available both online and in-store at store associate’s fingertips leading to a unified experience. Next time when the customer visits the store, not only is the associate able to see the cart but is also able to curate a virtual look that the customer might like on her tablet.

Scenario 4: CEO of a billion-dollar company announces his intention to double the ecommerce revenues in two years. Five teams raise their hands to rewrite the code of the underlying platform but no actual change sees the light of the day.

With a microservices based architecture, every team takes complete ownership of their microservice leading to more autonomy, better clarity and deeper business expertise. Rapid deployment enables the organization to be as flexible and quick as it needs to be to compete.

Delivering a Practical Version of Headless to those Driving Demand Today

The business case for the adoption of headless is relatively straightforward for large retailers, who are already likely struggling with their infrastructure. Still, daunting costs of re-platforming have kept many from making the switch.

There are significant considerations when it comes to the adoption of a fully composable commerce solution by smaller players. First, every microservice has to be deployed and managed separately, adding to the overall complexity. All endpoints need to be managed, and updates across the various services used need to be regularly incorporated. The total cost of ownership can become more than a monolithic application depending on the number of individual platforms used. For small teams who do not have the technical resources in-house, getting external help plus investing in additional infrastructure such as DevOps tools might be overkill.

As enterprise adoption lags and small players are content with monoliths, we see near-term demand from two groups:

  • Mid-market merchants could present a large market. For this category, going the open source route and fully managing their infrastructure in unlikely; their needs are simply not that custom and the resources required are enormous. Just a custom checkout experience can involve thousands of lines of code that needs to maintained.
  • A similar story exists for another target buyer, which actually lives across organizations of all sizes – a non-technical team looking to innovate with ecommerce experiences that funnel back to their checkout, OMS and 3rd party logistics providers. A PM or merchandising team won’t have adequate engineering resources to deploy a composable site for a pop-up site experience, but they increasingly seek the flexibility to build these around their core site.

The interesting thing is that neither group needs all the flexibility a fully composable solution offers, but they do need more than what a monolith has. In most cases a pure headless solution will do the trick. Though it can't offer what a full composable stack can, it can reduce latency and deliver unique site experiences. We are excited about frontend store builders that can be easily coupled with major commerce platform backends to serve this demographic.

Generally speaking, these customers need a mid-range offering, a low-code or no-code, templatized but extensible layer. If they do want to lean into composable beyond headless, they need solutions that can offer flexibility on endpoints and an extensible framework, but that are basically a templatized middle ground and come pre-built for ease and short time to value.

What’s Taking so Long?

We are still in the very early innings of this shift. While it’s exciting to hear “digital commerce is rapidly evolving and the future landscape will include API orientation at its core,”[4] it’s good to remember that Gartner wrote that in 2016. While the tremendous value of a composable stack is clear to many, it is a rare event that lands a tech leader with the budget to re-platform. Transitions are largely motivated by business leaders clamoring for faster and better omnichannel experiences, with many in the c-suite likely unaware of the backend changes that happen to enable those. Early innovators will continue to drive the composable trend: brands trying to compete on experiences and uniquely complex retailers who require very different checkout or storefronts than Shopify can provide.

But there is hope of accelerating adoption – storefronts, backends and infrastructure layers are entering the market with solutions that offer quick time to value, and meet demand from mid-market and business teams trying to create unique retail sites. And as customers’ digital experience becomes only more critical in the battle to win market share, it’s hard to imagine adoption won’t climb.

Disagree with Us?

Are you a brand evaluating the decision to transition from a monolith to the flexibility of composable commerce? An investor who has different perspectives? We’d love to hear from you. Drop us a line.

Huge thank you to Palak Papneja who led our research in this category and drafted much of this thesis.

[1] Howland, Daphne. “Retailers' best e-commerce asset may be their brick-and-mortar stores.” RetailDive, Dec 2020

[2] Salesforce survey of 12,000 consumers and 3,600 business buyers, July 16 through Aug. 18, 2020

[3] Gartner Composable Commerce Report June 2020

[4] Gartner IT Market Clock for Digital Commerce 2016, Mike Lowndes, 29 September 2016

By Yelena Shkolnik

Back to Insights