Emma Sleep's business-led blueprint for composable commerce

Connected digital technologies permeate every aspect of business operations today. There was a time when IT was a function that only touched certain aspects of the business, and it was often seen as just another cost to be controlled. But for a growing proportion of enterprises, technology is at the core of what the business does, and this has changed perceptions of its role. As Andreas Westendorpf, Chief Technology Officer at fast-growing German mattress firm Emma Sleep, puts it:

I think there was one change in wording a couple of years back that can be the moment where the switch flipped. It was when 'IT' became 'tech' ...

If you don't treat IT as just an enabler for business, but IT *is* the business — call this tech — then this is not about cost anymore. This is about revenue. This is about value streaming.

Adapting to this new role requires the IT function to change its approach. That was certainly the case at Emma Sleep when Westendorpf joined two years ago. He says:

The software development team was holding back the overall business development of the company ... [It] created a bottleneck in the organization because everything needed to go through the same funnel, through the same 'meat grinder' of software development. That made development of business-related features really slow.

Empowered product teams

This, he adds, was a demonstration of Conway's Law, which states that software design inevitably reflects an organization's communication structure. The solution was to change the structure by adopting author Marty Cagan's concept of product management, creating small, empowered product teams in which software developers, product managers and designers work together on specific goals. Westendorpf explains:

Teams can move more autonomously, and they actually discover the problems to solve, the 'What?', and then develop the solutions that will solve those problems, the 'How?', themselves.

The empowered product manager has a broader remit than the traditional role of 'product owner' in agile development, more focused on business value rather than simply completing a software project. He adds:

Product owners ... only look at the metrics and the success factors of software development, not of the business. And that is the fundamental difference to a product manager. A product manager takes ownership of the impact on the organization and the business. So the product owner role is part of the larger role of product manager. But the product manager looks at revenue, looks at conversion rates in e-commerce, looks at customer NPS, these kinds of business-relevant metrics, instead of only availability and deployment speed.

Going MACH for faster innovation

This enables a faster, more responsive innovation model that in Emma's case matches the needs of a business operating across 30 countries and doubling in size every 12-24 months. He explains:

Our organization is doing not one or two major things at once, but five to ten. We're approaching this year the US and China — the world's largest economies — both at the same time. We're introducing new products, getting to a multi-product company beyond just mattresses and beds. We are developing in the sleep technology area — not only the smart mattress, but also apps. We're getting stronger in the B2B area, wholesale retail, because we came from B2C and we have the digital capabilities to do retail better than conventional retailers.

We want to do all these different things in parallel concurrently. Right now, the old platform is holding us back because we only can focus on a couple of things at the same time. With this new approach, with more teams being able to work independently but concurrently, we can drive multiple things.

A technology upgrade to a more adaptable platform is an essential part of this effort. Emma is replatforming from its existing Magento e-commerce system to a composable commerce architecture centered on Fluent Commerce and Commercetools and built on MACH principles, where MACH stands for Microservices, API-first, Cloud-native SaaS, Headless. Westendorpf was talking to me at the recent MACH One conference, where he was a speaker in his role as a volunteer MACH Alliance Ambassador, advocating for education and adoption of MACH. The architecture lets Emma's teams choose the right solution for each specific business domain, such as the storefront, order management, warehouse management, the integration layer, and so on. He explains:

For each of these business domains, we wanted to select the best solution for us to work with. Fortunately, MACH now — and the SaaS companies who are members of MACH — supply solutions that can be well integrated with each other due to the API-first approach. That was not the case five or ten years ago, although the technology principles were there already, they were not applied in e-commerce yet. Now this is the case. Now we can build these ecosystems with multiple solutions, that we might even change solutions over time without being locked in.

It also provides the flexibility to try out different solutions for different use cases, he adds:

In such a composable commerce architecture, you can have even, let's say, redundant functionalities for a reason. For example, most of the countries with system A and then China with system B. That's feasible in such an environment.

Composable is 'not for everybody'

This flexibility requires skillful management and is not an architecture he recommends for a start-up. He elaborates:

Composable is not for everybody. A basic and the core requirement is that you are able — not outsourced but within your own company — to do sophisticated software product development. That means that you need to build up solid software engineering practices and culture in your company ...

In the beginning, if you're starting off a business, then you're in survival mode. You need to take decisions for today and tomorrow and not for in two years. The decisions taken five years back when we went with Magento were good decisions, wise decisions. We simply outgrew it. If we would have gone with composable back then, it might have killed the company, to be honest.

Now that Emma has reached a much larger scale, one avenue it's exploring is to extend the use of its data warehouse management system to the 100 or so local companies in its supply chain, from which it sources manufacturing and distribution for its various markets. The goal is to ensure it can access reliable data about stock and lead times, rather than relying on potentially fragile API data connections to whatever systems the local partner has in place. This allows it to have more insight into operations while still subcontracting the physical operations to an external partner. He explains:

We give them just the software and we work together with the data in it, but we don't have this fragile setup any more. This way we can optimize processes ... but not get into the heavy assets.

The roll-out of the new composable platform has already begun in some of Emma's smaller markets — two 'pioneer' countries that are still small in volume and one 'emerging' country where volume is growing. The next step is to roll out in one of the more complex 'trailblazer' countries before rolling out in scale within the next year or so across the rest of the business. The company will then be better placed to achieve its business goals, he concludes:

[It will] remove the brakes from all these ambitious people and what they want to do, and give them the right tools. That's the overall goal.