Dirk Hoerig is co-founder and CEO of Commercetools, and has worked in ecommerce for more than twenty years.
In this interview, he describes “headless commerce” and explains why this API-first model, often built on cloud infrastructure and microservices, matters so much in online retail and digital transactions as a whole. He also discusses where the next round of commerce innovation may appear, including on AR devices, in vehicles, and machine-to-machine in B2B transactions.
What is headless commerce?
Let’s assume everybody is familiar with ecommerce, but what is headless commerce?
DIRK HOERIG: Headless commerce is essentially a technical approach in which you split every experience layer in a shopping app — product details, images, videos, the add-to-cart button, everything a consumer sees and uses — from the core functionality underneath. From a product standpoint, that means offering all of the technologies, or capabilities, as a service, similar to how you may use compute functions from Google Cloud or AWS, and then allowing users to build any kind of experience on top.
This gives you two big advantages. First, you can run the same underlying technology stack across any device or touchpoint. That matters because today you don’t have only one website. You also have a mobile app, and perhaps something tailored for tablets. Store point-of-sale systems are linked in, and commerce is appearing in cars too. Pulling the logic or functionality apart from the user interface makes it simpler not just to connect these channels, but also to support each one.
Second, you gain much more room to create exceptional shopping experiences. Think of it like building a house: when you can pick the building blocks and arrange and move the house however you want, you have far more freedom to create what you envision. But if the walls and doors are already locked into place, the only real change you can make is the color of the paint.
That latter experience, where you’re boxed in, would be like a legacy ecommerce platform?
“Legacy” is probably the best term, but I wouldn’t say this is just a small set of old, obsolete systems. In terms of install base, they still account for more than 90% of the market, because systems have not been replaced very often. And whether they are 20 years old or 5 years old, they are still built on the same old paradigm, or the same old legacy technology. Some products from the late ‘90s — when we saw the first commerce platforms emerge — are still being sold, only with newer version numbers. But the basic idea remains: everything we built in the 2000s was meant to copy the webshop behavior of amazon.com.
Until 2010, mobile didn’t matter all that much from a commerce-platform point of view because, in the western world, internet penetration on mobile devices was much lower, under 5%. It was really only in the past five to six years that brands and retailers had to begin seriously planning for new channels and adding flexibility to their online operations.
The market is expanding, consumer behavior is shifting rapidly, new devices are appearing quickly, and loyalty to brands is very low. … So you need to elevate the experience and generate a few ideas.
What’s the benefit to consumers from a headless architecture? Put another way: We all know a bad online shopping experience when we see one, so how does headless commerce help solve that?
As a consumer, you don’t know whether you’re shopping on a headless commerce platform because that sits in the underlying architecture. But what you do notice are two things. First, the shopping experiences are more customized and feel considerate and inventive, rather than seeming like somebody simply ticked some boxes. Second, the brand you’re dealing with can react to consumer demands, or to your shopping demands, much more quickly.
Our customers are shipping new updates to their features and consumer experiences several times a day. As a result, as a consumer, you always receive a current experience and faster responses to feedback than a site running on a more monolithic platform could deliver.
Why cloud and microservices?
Speaking of monoliths, the industry is moving toward microservices, as well as cloud infrastructure, as the base for headless platforms. How important is this architecture to what you’re trying to do?
It’s the de facto standard for consumer-experience enablement technologies because these days, to succeed online, you have to be highly scalable and flexible. I’d say that across the entire IR500, you won’t find a single exception. They have to update the consumer experience, ideally several times a day, with no downtime. That is the No. 1 priority for reaching their business goals, and it’s very hard to do that without a cloud native platform.
What’s going on here is that the technology team at a large company has a massive queue of requests from different departments aimed at improving customer retention, conversion, or engagement — you have to keep improving everything all the time — and they are trying to optimize it in many small but iterative steps. Microservices let you separate functions from one another so that when you alter something in the product logic, for instance, it does not affect what the customer sees while shopping and adding items to the cart.
That sounds incredibly simple, but older software was not designed for it. Everything built 20 years ago was made not to be altered on a regular basis. Back then, once a system was in place, you left it alone for years. And if you did need to make these changes, you built your platform the way Amazon or Walmart did, so they could use microservices-based architectures.
Another major issue here is scalability and availability. Today, online shopping represents about 20% of retail share, and it will likely rise to 25% over the next four years. That is very different from 15 years ago, when it was 4% or 5%, and when major retailers might have done $200 million in online sales each year. Now we are talking about sales well into the billions of dollars for major retailers.
After this ecommerce boom, the market mostly answered with, “Let’s move existing software to the cloud.” But these platforms were never built to operate in the cloud. So they are not using the cloud in the most efficient way; they are not getting the most out of the resources. I think there needs to be a clear distinction between what is on the cloud and what is in the cloud. Or between cloud native development and not-cloud-native development.
Having a headless commerce architecture gives you the flexibility to work across any kind of touchpoint, device, and business model. Microservices make it possible to develop independently and introduce changes — your product catalog team can make changes separately from the cart team — and then the cloud makes sure you have high scalability.
Looking at the current business environment: the market is expanding, consumer behavior is shifting rapidly, new devices are appearing quickly, and loyalty to brands is very low. If something disappoints you and it is not your favorite brand, then you will probably move to someone else. Customer acquisition costs are extremely high, and it does not seem like they are going down. We are heading into a recession, so customer retention will matter more than anything else right now. So you need to elevate the experience and generate a few ideas. And if those are all your requirements, I believe you need something that runs natively on the cloud and gives you flexibility on the API side.
The subscription economy, or in-app purchases, is making its way into physical goods — how would you do that if not in a headless fashion?
New business models
It’s not only about scale, right? There are also new devices and interfaces, such as voice and smartwatches, plus newer consumer trends like sneaker drops.
Oh, yes. We have a customer, a large online fashion retailer that runs “drops,” which create huge spikes in demand that can’t always be foreseen. So, yes, I completely agree that this is about new business models and use cases, not simply swapping what you already have for something more modern.
All companies now have to consider new business models. For instance, many automotive companies — beyond reserving and purchasing a car online, and handling aftermarket parts and merchandising — are now also doing in-car commerce, including our customers Audi and BMW. You can buy upgrades in the car while you’re driving it, from this navigation display panel. You can say, “Okay, give me more maps, give me better entertainment functionality.” The subscription economy, or in-app purchases, is moving into physical goods — how would you do that if not in a headless fashion?
There is the sheer spread of devices online, changes in consumer shopping behavior, the pandemic … What do you think has been the biggest shock to ecommerce over the last several years?
I think the biggest challenge, really, was for all of these companies to realize that they are no longer fast enough to react to all consumer trends and requirements. And it’s often that your customers don’t tell you what they want — they simply go elsewhere. And, often, companies don’t see this happening — and what their customers are searching for — until it’s too late.
There’s a very good example from Best Buy, which a former chief architect wrote about a few years ago. They wanted to move their “Add to Cart” button from one side of the page to the other, because a test said it would raise the conversion rate a little bit. It took them about 6 months, involved many people, and cost more than a million dollars. But the worst part was that it made it hard to make other changes during that same period. Meanwhile, Amazon was pushing changes constantly and growing its market share.
This was the biggest shock, because then you realize there are so many things underneath that you have to change. If you think about it like a car, as a retailer you want to concentrate on upgrading your chassis — your design, your UI, all of that — but first you need to know what kind of car you’re building and what condition the engine, gears, and wheels are in.
With the pandemic, ecommerce adoption rose because more companies figured out, “Oh, wait, now we need ‘click and collect.’ But we don’t need it tomorrow, we needed it yesterday.” And then they realized, “But our point-of-sale system isn’t ready for that. We can make that change to our online platform, but it will take six months. And we’re in lockdown; we’re closed the next six months. So what do we do until then?” The pandemic put everything that wasn’t working into the spotlight.
I think most companies already knew it before, and if they weren’t able to react fast enough, then it became a problem.
On the B2B side, your customers are, in their free time, consumers — and they expect a similar kind of experience for work-related transactions. Why should shopping be dull between 9:00 and 5:00, and be exciting after 5:00?
The future of ecommerce
What do you think is the area most ready for innovation in ecommerce? Is it UX, backend technologies, or perhaps new business models to take advantage of those technological advances?
It’s probably about thinking through new business models. Of course, not everything that no longer worked well offline, or in an analog way, will automatically work well just because it’s digital. Some business models are simply outdated and need to be changed. I think every company, whether they’re retailers, brands, manufacturers, or software companies like us, should constantly ask themselves, “What is my purpose here? What is the value that I’m providing for my customers, and is that a big differentiator?”
When we launched Commercetools in 2014 and talked with our first customers, this technology was relatively new to some of them, and I had to explain ideas like API-first. I had customers asking me, “So, Dirk, that’s great. Now, we have all the flexibility that we can finally build whatever we want. But what should we build?” We were glad to help where we could, but only a business can truly understand its customers and create the strategy that will best serve them.
I think the other area ready for more improvement is B2B, especially around automation so there’s less human interaction, less filling out forms. Think about those analog processes where, too often, something still has to be printed out, signed, scanned, and then faxed or emailed back. It’s really about business innovation and driving improvements like predictive maintenance or something similar, where a machine could detect a faulty part and automatically order it. There’s a lot of stuff where you can mix machine learning and AI with other business logic systems, and then a headless platform, to enable the whole shopping process automatically.
For example, we have robotic manufacturing customers, where our APIs are directly embedded and built into the product. So you have a machine-to-machine B2B scenario, where the commerce solution is just working as a mitigator in between the systems, and is responsible for providing the products, calculating the price, creating the cards, making the order, and so on. Once you have a headless platform, and you implemented all of that logic and functionality, you can almost use it in every type of business use case.
Also, on the B2B side, your customers are, in their free time, consumers — and they expect a similar kind of experience for work-related transactions. Why should shopping be dull between 9:00 and 5:00, and be exciting after 5:00? I think there’s a lot of efficiency on the B2B side that’s going to happen over the next two years.
Technology is getting closer and closer to our body. We already have it in our pockets or on our wrists, but glasses or other devices, if they’re good enough, will make a huge difference on everything that we are doing and how we are doing it.
What is a new technology that’s going to influence how you have to build your product, or that customers are going to require you to support?
I think there are two things, one is back-end, and one is front-end. On the back end, once machine learning and artificial intelligence get to a more advanced stage, they will help brands and retailers improve a lot of decisions. Because a lot of stuff is human-based and offline. When you look at the processes of retailers, they work with huge Excel catalogs on updating price lists and product data. And often when it comes to finding the right price, providing discounts, or coming up with the best marketing campaign for Valentine’s Day, Halloween, or something like that, it’s a very human thing. But you’d like to be able to calculate what’s working best by collecting the right data and then making smart decisions based on customer behavior.
I believe that in the next five years, getting commerce technologies like ours into a kind of autopilot mode that helps businesses make faster and smarter decisions, and automate them, will make it easier for retailers and brands to scale and compete. Especially those that have complex business models, are selling in many countries, and have lots of shoppers with diverse backgrounds.
The second thing is connected to how the consumer market is evolving. I’m talking about augmented reality, which I believe is going to be really important because it’s a natural thing that technology is getting closer and closer to our body. We already have it in our pockets or on our wrists, but glasses or other devices, if they’re good enough, will make a huge difference on everything that we are doing and how we are doing it. It will have a huge impact on commerce.
Ecommerce completely changed with the introduction of smartphones. They’ve changed all of our behavior — how we interact with each other, how we look at and buy products, and how we inform ourselves about products, companies, and brands. This was a huge impetus for building Commercetools, because we saw that businesses needed a commerce solution that could expand beyond a simple webshop and onto new platforms. When a new device becomes as ubiquitous as the smartphone, there will be another inflection point.