Every founder knows this type of pivot story. A struggling check-in app called Burbn had many designed features, but users loved one specific feature. After stagnating at low user base for a while, the founders cut everything else, and the one thing that was left became Instagram.
It is the classic pivot narrative in startup history, and it often gets cited as proof that you should watch what users do rather than what you built. That advice is fine from a design thinking, user experience, and user-led innovation perspectives. The problem is what the story gets used for after that, which is to argue that Instagram was a case of a company discovering a hidden use for something it had built.
I screened it for ExAI based on the EGS Process Model. It does not hold up as a clean exaptation case, and the reason it does not hold up is more instructive than a “yes, this is an exaptation” would have been.
What Burbn actually was
Kevin Systrom named it after bourbon, his favorite drink. It launched as an HTML5 mobile web app in 2010, inspired by Foursquare, and it did a lot of things. Users could check into venues, make plans, earn points for hanging out with friends, message each other, and post pictures. Comments and likes sat on top of all of it similar to how other social media platforms like Facebook operated at the time.
But a few months after launch, the numbers were bad. Reports put Burbn at roughly a hundred users at its peak. Foursquare and Gowalla had the check-in category sewn up, and Burbn had no obvious reason to exist inside it.
The founders had built a complete, native iPhone version of Burbn, looked at it, and threw it out. Systrom later said it felt cluttered and overrun with features, that they decided to start from scratch, and that they cut everything except the photo, comment and like capabilities.
Quite obviously, those last two sentences contradict each other: Starting from scratch and cutting everything except are two different claims about the same decision. That contradiction is at the center of this case, and I will come back to it.
The redesigned Instagram shipped on October 6, 2010. The name combines instant camera and telegram. It hit a million users inside six months and sold to Facebook about eighteen months after launch, for a billion dollars, with fewer than 20 employees and no revenue.
Reading it as modules instead of as a product
The most useful way to look at this is not as a product that changed. It is as a set of modules (aka components), some of which were kept and most of which were thrown away.
Burbn held roughly nine of them: check-ins, plans, points, messaging, meet-up tools, photos, comments, likes, and the social graph underneath. The operation Systrom and Krieger performed was exclusion: Five modules out, four retained. Then a new architecture built around what was left with a native iOS app instead of an HTML5 one, filters, the square format, a public feed. Krieger has said they carried some of Burbn’s interface ideas and its testers across too.
This framing comes out of the modularity literature, and Pierpaolo Andriani and Andrea Carignani have done the most useful work connecting it to exaptation. Their argument is that modular systems make functional change easier because modules can be excluded, substituted and recombined without rebuilding everything. Modularity creates the opportunity, but it does not create the shift.
That distinction is doing all the work here.
Where the screen pushes back
EGS asks a small number of blunt questions before it will call something an exaptation. Three of them matter in this case.
1) Did the resource acquire a function it did not previously have? Look at the photo module. Before the pivot, you posted a picture and other people commented on it and liked it. After the pivot, you posted a picture and other people commented on it and liked it. More people, better pictures, faster upload, nicer interface. The activities remained the same. What changed was the module’s position in the product, from something attached to a check-in to the thing the whole app was about.
2) Was the new use the reason the change was made? Yes, explicitly. They cut features in order to deliver photo sharing. The function was selected first and the rebuild followed. That is deliberate design, which is what adaptation looks like. An exaptation runs the other way: the use shows up and then someone notices.
This test is also where Systrom’s personal history matters. He built a photo sharing site with his fraternity brothers at Stanford. He studied in Florence, where a professor took away his camera and handed him a Holga. Mark Zuckerberg tried to recruit him in 2004 to work on a photo product, and he turned it down. He wrote the first filter himself. None of this reads like a man who was astonished to learn that people like posting pictures.
3) Did the same thing persist across the change? Contested, and it is the contradiction I flagged earlier. If the resource is the whole app, it did not persist. Instagram was iOS native, Burbn was HTML5, and the founders described starting over. If the resource is the photo, comment and like subsystem, it did persist, because that is precisely what they said they kept.
From a case study methodology perspective, two of the three questions come back negative. On a strict read, this is a pivot: a reallocation of effort toward a function that already worked.
Then the level changes, and so does the answer
Here is where it gets interesting, and where I want to be explicit rather than tidy.
Everything above is a component-level reading. Point the same questions at the whole artifact and the answers move.
Burbn’s function (aka job), as a system, was to help you coordinate where you are and what you plan to do. Instagram’s function was to help you share your life visually. Systrom framed it that way himself at launch, and his sharpest version of it is that the product went from a check-in with an optional photo to a photo with an optional check-in. At the system level, the selection-defining function changed. The components did not.
It would be easy to pick whichever level gives the verdict you want. I think that is the single most common failure in how pivot stories get classified, and it is worth naming as a rule: the unit of analysis is allowed to shift as a case develops, but when it shifts you report both readings and you say that it shifted. Suppressing the shift is how an ordinary product decision gets written up as a discovery, and how a framework quietly becomes unfalsifiable.
Where this lands
The Burbn-to-Instagram is a boundary case, with real emergent behavior inside it driven by initial users.
The emergent part is specific and it is not about the photo feature. It is about who decided. The founders designed Burbn around check-ins. The users selected photos. That selection was not designed, not forecast, and not under anyone’s control, and it is the reason the product exists. There is a detail that makes the point sharply: when the pivot was announced, a good part of the Burbn community pushed back. They did not want another photo app. They wanted their check-ins.
What keeps it from being a clean case is everything that happened afterward. The function already worked. The rebuild was aimed at a target chosen in advance. The artifact was substantially reconstructed. That is a company executing well, not a company finding something.
The better exaptation story is two degrees away
If you want a genuine discovery from this story, it is sitting in Systrom’s biography.
The Holga is a cheap plastic camera built in Hong Kong in the early 1980s to give ordinary families an affordable way to take pictures. Its light leaks, vignetting and soft focus were not features. They were the consequences of building a camera as cheaply as possible. Then the domestic market for it collapsed, the cameras reached photographers abroad, and those flaws became the entire reason to own one. Instagram’s early filters are, in large part, a digital reconstruction of defects that a cost-cutting manufacturer never intended anyone to want.
That is a function nobody designed, attached to properties that were pure byproduct (i.e., spandrel). This case has not been screened yet as an exaptation. My expectation is that it clears where Burbn does not.