Wikipedia will tell you that “headless commerce” was coined in 2013, that a Forrester report lit the fuse, and that the rest is adoption curves. The entry is not wrong, exactly. It is a stub where a story should be. I spent the better part of a decade in or near the rooms where that story happened, at Adobe, at IBM, and eventually as Chief Marketing Officer of Elastic Path, and I have been quietly collecting the primary sources ever since. This is the fuller version, including the part nobody has been able to prove, which is my favorite part.
A cliff at the end of every beautiful page
In the early 2010s I was leading commerce strategy at Adobe, and we had a problem that sounds quaint now. We helped brands build gorgeous marketing experiences, and every one of them ended at a cliff: the moment a customer wanted to buy, they fell out of the brand moment and into a clunky e-commerce site that looked like it was built by a different company, because it usually was. It seemed crazy to me that brands would pour their souls into products and then present them as thumbnails in rows and columns. We started asking a strange question for the time: what if we did not need the store at all? What if the transaction could happen wherever the customer already was, inside the content, inside the app, inside the moment? We proved it on Adobe’s own business, making the leap from a free trial to a paid license a button inside Photoshop instead of a trip to a web page. I even tried to get Amazon to license us 1-click. They politely declined.
That work had a name inside Adobe: experience-driven commerce. I did not coin it; one of the leaders I worked for did, and handed it to me to run with, which I did for years. Chasing it meant evaluating every commerce engine on the market with an unusual requirement: we did not want their storefront, we wanted their APIs. In 2013 that was a short list. Forrester was circling the same collision from the analyst side, and that November, Stephen Powers and Peter Sheldon published the report that gave the industry permission to think this way: “Content And Commerce: The Odd Couple Or The Power Couple?” Adobe was one of the companies interviewed for it, and I was on the other end of some of those conversations. Here is the detail history forgets: the report never said “headless.” It recommended that content management be “loosely coupled” to the commerce engine. Accurate, architecturally precise, and absolutely never going to survive a sales meeting.
Meanwhile, in Vancouver
The company that kept showing up at the top of our API evaluation was Elastic Path, out of Vancouver. Their conversion story was not theory. They had run the official store of the 2010 Winter Olympics, thousands of SKUs and more than ten thousand orders a day, and came out the other side convinced that commerce was escaping the website. They had customers like Symantec who wanted to sell inside a software download flow, where there is no storefront to send anyone to. So in February 2012 they launched what they called the Digital Commerce Engine, built around the Digital Commerce API, an API designed from the start to let commerce live inside anything: an app, a game, a kiosk, a download. The acronym was DCAPI. Hold that thought.
By July 2013 they had evolved the idea into Cortex, an API layer designed to carry not just the data of a transaction but the logic of one, so the flow of a purchase could live behind the API instead of being rebuilt in every interface. Years later Forrester would write that this orchestration layer was what separated Elastic Path from every other API-first vendor. At Adobe, it had already changed how we thought about the whole category.
Meanwhile, in Munich
While Vancouver was shipping, a parallel story was unfolding in Germany. A company called commercetools had been working since 2010 on a platform deliberately built with no front end at all, and in April 2013 their founder, Dirk Hoerig, launched it as SPHERE.IO and gave the architecture a name: headless commerce. I want to be unambiguous about this, because I once spent an unreasonable amount of time digging through hard drives hoping to prove otherwise: the public coinage belongs to Dirk, and I tip my cap. It was the right name. “Loosely coupled” describes a diagram; “headless” makes you see a body with no head, and once you see it you understand the architecture without a single slide. Naming a thing that well is its own kind of talent, and the name did real work for all of us, his competitors very much included.
The name also did not appear out of nowhere. “Headless” had decades of history in computing, where a headless server is one running without a display. The content management world was converging on the same word at the same time: Contentful was founded in Berlin in 2013 and built its whole identity around being a headless CMS. And here is my favorite artifact in the folder: an SAP hybris product datasheet, version dated October 2013, that uses “headless commerce” three times, including an entire feature section titled “Headless commerce support.” Six months after the term was introduced, the biggest suite vendor in the market was printing it in collateral as if everyone had always talked this way. Good names spread like water finding cracks. Nobody owns water.
The DCAPItator
Now back to that acronym. Here is the story the way Elastic Path veterans told it to me, and the reason this piece has its title. The Digital Commerce API was DCAPI, and engineers being engineers, DCAPI became the DCAPItator, with the Terminator energy fully intended. The DCAPItator’s job, after all, was violence: it ripped the head off the storefront. Decapitating the platform, ripping the head off, headless. You can see the whole semantic pipeline, and if you have ever worked with engineers you can hear exactly how it happened, somewhere between a sprint demo and a pub.
I have to be honest with you about the evidentiary status of this story, because the honesty is the point of this piece. I have searched, and had research run, through archived press releases, whitepapers, developer forums, and every captured page that can be captured: there is no public document from 2011 to 2013 in which Elastic Path uses “decapitator” or “decapitating the storefront” in its marketing. The lore is structurally plausible, linguistically sound, and archivally invisible, which is exactly what you would expect of a joke that lived in an engineering bullpen. So I am doing what historians do when the record runs out: publishing the question. If you were at Elastic Path between 2010 and 2013 and you have an email, a deck, a Jira ticket, a whiteboard photo, or, please, a t-shirt with the DCAPItator on it, send it to me. I will update this piece, with credit, and happily revise history in public. That is how the record is supposed to work.
Naming it was not the same as selling it
There is a third chapter that tends to get collapsed into the first two. A category is not an architecture plus a name. A category exists when buyers believe it, analysts structure markets around it, and competitors are forced to argue on its terms. In 2018 I got the chance to work on that chapter personally when I joined Elastic Path as CMO, and we bet the company’s entire positioning on headless. We evangelized it relentlessly, to the press, to the analysts, to anyone with a replatforming decision and a conference badge. By late 2018, Forrester’s B2C commerce Wave formally described “headless commerce vendors” as a market group containing exactly two companies: commercetools and Elastic Path. The two teams that had spent five years building the same idea from opposite sides of the Atlantic were now, officially, a category. A year after that, a Forrester analyst told us the most common question from his clients had become “why is everyone saying they’re headless?”, and the competitive attacks against us had shifted to debating whose headless was the real headless. When your rivals are fighting inside vocabulary you spent years amplifying, the category exists.
What I actually think about credit
So who invented headless commerce? The honest answer is the least satisfying and the most common in the history of technology: several teams at once, because the conditions demanded it. Smartphones fractured the buying journey, monoliths could not keep up, and everyone close to the problem started pulling heads off platforms at roughly the same time. Elastic Path verifiably shipped the architecture first and has the 2012 press release to show for it. Dirk Hoerig named it, brilliantly, in 2013. Forrester gave enterprises permission to want it. Contentful taught half the developer world the vocabulary from the CMS side. And a decade of marketers, mine among the loudest, made buyers believe it. Simultaneous invention is the norm, not the exception; the myth of the lone coiner is just easier to fit in a Wikipedia box.
What I take from it is a lesson I now apply everywhere: in a new category, the architecture is necessary, the name is powerful, but the spoils go to whoever makes the market believe. And somewhere out there is a hard drive with a 2012 slide that says DCAPItator on it. Go find it. I will be here.