The idea hiding inside Disco began at the end. A shopper had already clicked buy. The hard work of attraction, persuasion, and payment was finished. Then came the order-confirmation page, a strangely quiet patch of digital real estate where attention remained high and almost nothing happened. Conner Sherline looked at that blank space and saw a market.
His first product placed recommendations for complementary independent brands after checkout. A cookware buyer might meet a pantry brand. A customer who had just found one useful thing could discover another. The merchant showing the recommendation earned distribution elsewhere in the network. The shopper got a relevant suggestion. Disco, then called co-op commerce, measured the exchange and learned which pairings worked.
It was modest enough to describe as a widget and ambitious enough to contain a theory of the internet. Independent merchants did not lack traffic in aggregate. They lacked the shared infrastructure that let a marketplace turn many separate transactions into useful intelligence. Sherline’s wager was that brands could pool some of that signal and distribution while keeping their own names, storefronts, and customer relationships.
That distinction matters. A centralized marketplace asks merchants to gather under its roof. Disco puts connective tissue between the roofs. The company’s work has since expanded into auctions, multiple offer formats, measurement, enterprise partnerships, and API-first products. But the animating question has barely changed: what can independent commerce businesses do together that each one cannot do alone?
An education in attention
Sherline arrived at that question by moving through the machinery of digital discovery. He graduated from Southern Methodist University in 2013 with a marketing background. The move that followed was less polished than the résumé it eventually produced. He has recalled driving a U-Haul to New York with about $5,000 saved, leaving it parked under a bridge while he spent nearly two weeks finding a small apartment he could afford.
The available job was an old internship at Mashable paying $15 an hour. His account of the period includes a diet of dollar pizza slices and Mamoun’s, a recognizable New York calculation in which time, rent, and appetite meet on a narrow budget. It also makes the next line of his career feel less preordained. He was not stepping onto a manicured founder track. He was trying to make the city work.
That year he joined Facebook in New York. Over five years, his roles crossed chief-of-staff work, sales, business development, and product marketing. He worked around Facebook and Instagram as they became default places to discover young brands, including work connected to Dynamic Ads. Sherline watched the advertising auction mature and prices rise. In a 2020 interview, he recalled sub-$2 cost-per-thousand impressions in 2013 and figures above $75 during the election period seven years later.
At Affirm, where he moved in 2018, the view shifted deeper into the transaction. Sherline led product marketing for payment, lending, and marketing products and worked closely with ecommerce merchants. Facebook had taught him how discovery scaled. Affirm showed him the captive attention and first-party context that lived inside independent stores. The combination produced an apparent contradiction: the collective traffic of those shops was enormous, yet each merchant kept bidding against the others for the next customer on a few outside platforms.
“Why are all these brands competing against each other in the auction? Why don’t they unify in some way?”
Conner Sherline, on the road to Disco
The answer became a company. Sherline raised initial capital in early 2020 and started with the post-purchase surface because it already had the ingredients that broad brand partnerships often lacked: context, a measurable action, and an incentive for both sides. Cooperation could stop being a soft promise and become an accounting system.
From widget to commerce graph
Early proof came from willingness. By the end of 2020, the young network had signed roughly 200 brands. By the 2022 Series A announcement, Disco said its post-purchase network spanned more than 600 direct-to-consumer brands and had observed more than $1 billion in transactions and 40 million shoppers during its first year. The investors in the $20 million round included Felicis Ventures, Shopify, Sugar Capital, Bessemer Venture Partners, Indicator Ventures, and RiverPark Ventures. Disco said it had raised $26 million in total at that point.
The capital was attached to a larger product ambition. Every purchase carried clues: which products traveled together, which shoppers responded to which offers, and which pairings created incremental value rather than noise. A small merchant could not see those patterns beyond its own walls. A network could.
Sherline has framed the data argument as democratization. Amazon can use a vast field of transactions to understand relationships between products and people. An independent merchant sees a narrow slice. Disco’s network offers a way to learn from interaction patterns across stores, then return useful ranking, merchandising, discount, audience, and partnership insights to participants.
There is an elegant business mechanic inside that civic-sounding language. Shared intelligence only becomes durable when the system can reward contribution, protect quality, and show economic results. Disco’s early network screened brands rather than accepting an unbounded catalog. Its later products added closed-loop attribution and an auction in which advertisers could bid for placement across checkout partners. Cooperation acquired prices, rankings, and performance metrics.
Sherline reported a notable operating milestone at the end of 2024: Disco’s first month above $1 million in revenue. He credited a sequence rather than a single breakthrough. The company repaired its attribution system, introduced new ranking, launched a format called Nurture that he said produced more than four times the engagement of the original DiscoFeed, and added the Disco Auction. Average revenue per user had more than doubled that year, he wrote.
The next expansion moved beyond the company’s DTC beginnings. A 2025 partnership with Gopuff put Disco’s personalized offers into instant commerce. A later Mindbody partnership carried the system into booking and service transactions. Sherline described distribution partners as co-creators who own the experience and customer relationship. Disco’s job was to improve monetization without treating that relationship as disposable inventory.
By May 2026, Disco described itself as infrastructure for commerce media wherever transactions happen. The company reported 137 percent year-over-year revenue growth and introduced DiscoBeat, an API-first, white-labeled layer for consumer-facing software platforms, alongside DiscoMix for multiple media formats. The original post-purchase insight had widened into a claim about the transactional web: any business with meaningful consumer moments may be able to build a media business around them.
A CEO moves closer to the code
Sherline’s public writing in 2026 adds an unexpected second act. The marketer who became a network founder has been describing himself as an active AI-assisted product builder. He tested tools across coding, design, agents, infrastructure, deployment, and workflow, then worked his way into Disco’s repositories, data schema, eventing, authentication, and security model. The point was not to make attractive demos. It was to shorten the distance from product intent to production.
His analogy is personal and revealing. He compared the experience to playing The Sims at 12, staying up late after finding the cheat code for unlimited money because there was always another wing to add to the mansion. Decades later, the feeling returned through a system of agents and multiple models working across parallel projects. He wrote of directing that work from ordinary pockets of family life, then handing the results into an engineering process built to stage, test, and harden what he called “CEO code.”
The anecdote is playful; the operating doctrine underneath it is serious. Sherline argues that AI changes the old relay race in which an executive has an idea, a product manager translates it, a designer represents it, and engineers interpret the representation. A product-minded person can now make the experience tangible much earlier. Engineers spend less time guessing at intent and more time connecting it to durable systems.
“Perhaps we should just call it what it actually is, shipping quality product.”
Conner Sherline, on AI-assisted development
He is equally clear about the trap. Generated interfaces written in the wrong language, detached from backend services, or oblivious to security remain prototypes. The practical unlock, in his account, is a deployment architecture that preserves context and quality: repositories, databases, event streams, authentication, observability, evaluation, and the engineers who understand them.
Disco’s reported internal results are striking. Sherline wrote that from the fourth quarter of 2025 to the first quarter of 2026, the company generated 38 times more code, commits, and code cleanup while headcount stayed flat. More recently, he described moving Disco’s memory, workflows, evaluation loops, and processes into locally operated, model-independent infrastructure capable of routing work across many models. The institutional memory, rather than any single model provider, is the asset he wants the company to own.
That approach fits his longer career better than it first appears. His résumé has always crossed boundaries that organizations prefer to keep tidy: sales and product, marketing and business development, executive direction and hands-on building. In an earlier reflection on leadership, he wrote that influence required speaking less, asking more questions, and giving up the need to be the smartest person in the room. His current practice adds a productive tension: a CEO can empower specialists while also becoming more capable of expressing an idea in their medium.
Respect is part of the model
Commerce media can easily become a polite name for filling every spare surface with an ad. Sherline’s more interesting argument is that additional inventory alone has diminishing value. Relevance has to survive the expansion. A booking confirmation, refund tracker, or delivery receipt carries a different mood and purpose. The offer has to fit the moment, and the platform that owns the customer relationship should retain control of the experience.
That principle also sharpens Disco’s cooperative premise. A network is not valuable merely because it is large. It becomes valuable when participation produces better outcomes without eroding the identity and trust each member contributed. Sherline summarized the point when discussing the Gopuff partnership: the future would be won by “respect + relevance,” not scale alone.
The phrase connects the two versions of his work. In the first, brands collaborate rather than disappear into someone else’s marketplace. In the second, AI assists people rather than severing product work from organizational judgment. Both are attempts to gain leverage without handing away the thing that made the system useful in the first place.
Six years after co-op commerce began, Disco is larger, more technical, and aimed at a wider set of transactions. Yet its founder is still working on the same design problem he noticed after checkout: attention has value, but extracting that value crudely destroys it. The better system makes the exchange legible, returns control to the participants, and learns enough to offer the next person something worth seeing.