A prediction about e-commerce architecture

Let me start with a few musings on the evolution of e-commerce architectures. 

By that, I mean mostly the systems underlying e-commerce engines and how that stack has changed over the past 10-15 years. I'm also, well aware that the following, like everything else on SoMe, is a massive simplification of reality, so read it with that in mind. 

About 10 years ago we had large monolithic systems that encompassed both Content Management and E-Commerce and were primarily focused on the in-browser experience. Systems like Sitecore, Adobe EM etc. These did everything needed to run a website and/or a storefront in the browser. They generally didn't do everything equally well, but that was a worthwhile trade-off with the requirements that most companies had back then.

The first chunk of data we needed to free up was the Product Information. Once we realized how many different places throughout the supply chain, retail and digital properties we needed consistent and accurate product information, most companies started building or implementing PIM systems to sit between ERP, E-Commerce and Retail systems. Free up the Product Information, edit it in one single place and use it in many systems (either through syncing or through APIs).

3-5 years ago we started realizing that our customers wanted to be recognized across channels, eg in retail, on the website, in the app and customer support. We also needed deeper and better handling of customer data through our marketing channels. So we went ahead and freed up customer data by implementing CDPs (Customer Data Platforms). This trend has accelerated in the past year as doomsday talk of the dreaded cookiepocalypse has gotten louder and louder. 

In parallel with the above trends, we have seen the need to separate the front-end from the backend, in what we refer to as "headless" architectures or websites. Headless generally means that we develop the frontend separately from the backend, using APIs to fetch data from the backend directly into the frontend. The term is often used together with Jamstack (which refers to the technologies most often used to build headless frontends).

In short, over the past 10 years or so, we've tended to move more and more functionality out of the monolithic systems (be it CMS or E-com Engine) and that is a trend that we are seeing accelerate with the advent of "Composable Commerce" and/or "Packaged Business Capabilities". Over the past 3-4 years we've seen the rise of the MACH Alliance (https://machalliance.org/), and the advent of tools such as commercetools, contentful etc. which are cloud-native systems that are API-first in their approach to the world. In other words, they were built specifically for headless architectures. 

Basically, over the past 10 years, we've gradually moved 2 of the 3 core datasets out of our ERP systems, and out of our E-Commerce engines.

In the coming 3-5 years I think we'll see an increased focus on the 3rd dataset. As customer expectations for service, speed of delivery and transactional choice increase, we see more and more focus on Order data or transactional data. (I'm not saying that Order Management is, in any way, a new thing, just that it is beginning to be more critical to consumers from an outside-in perspective). The advent of MACH-compliant, API-first order management systems is an indication of this.

Take a look at Fluent Commerce (https://fluentcommerce.com/) and Omnium OMS (https://www.omnium.no/en/) for a few examples. Perhaps the most interesting take on this trend is CommerceLayer (https://commercelayer.io/). CommerceLayer is an up-and-coming e-commerce engine that has entirely done away with the "content" side of e-commerce, pushing product information and customer information into dedicated systems. It focuses entirely on the transactional side of e-commerce, trying to create an e-commerce engine that natively supports complex order mgmt scenarios without trying to do everything else well. 

I fully expect that in the next 3-5 years we will see e-commerce implementations without anything we would recognize as an e-commerce platform. Headless front-ends that are built composable, using something like Uniform (https://uniform.dev/) as a frontend orchestration layer, and otherwise integrating the front-end directly into PIM, CDP and OMS, with nothing but the basket left for the commerce engine to manage. In some cases, we'll likely see the basket integrated into the OMS (CommerceLayer), or perhaps the basket will be handled by something like Bolt (https://www.bolt.com/), a purpose-built basket and checkout experience. 

There are a million more things to say about #omnichannel #ecommerce #headlesscommerce #mach but we'll save that for another rainy day. And since I've been challenged to talk a little about leadership, perhaps I'll share some thoughts on what this type of transformation means from a cultural and leadership perspective. (Which is way harder to get right than the systems architecture).