Andrew Duffy

Andrew Duffy runs SparkPlug, a San Francisco software company built on a deceptively small idea: the person who actually sells you the thing should get paid when you buy it. Not the regional manager. Not the brand. The human at the register who talked you into the better bottle, the bigger cut, the second round. SparkPlug measures that influence and turns it into cash, automatically, in real time. As of its Series A, the platform had already routed over a million dollars of supplemental income into the pockets of frontline retail and restaurant workers, and Duffy is not shy about wanting that number to read in the billions.

He describes his own career as a sequence of obsessions. "Behavioral science nerd turned finance nerd turned startup nerd," is how he puts it, and the line does a lot of quiet work. It explains why a guy who once built financial models for the Central Bank of Mexico now spends his days thinking about budtenders and baristas. The thread running through all of it is the same question that has followed him since he was a teenager: what happens when you actually pay attention to the people doing the selling?

The origin story is almost too neat. As a teenager waiting tables at Cactus Cantina in Washington, D.C., Duffy sold more Gold Cadillac Margaritas than anyone else on the floor. Then he did it again the next month. And the month after that. Three months running, he was the top margarita salesman in the building, and somewhere in that streak he noticed the thing that would later become a company. Being measured felt good. Being rewarded for it felt better. The restaurant had no real way to capture any of it, no system to notice that one server was quietly moving more high-margin product than the rest of the staff combined. The information just evaporated at the end of each shift.

Years later, he would build the machine that catches it.

The Detour

Harvard, a hedge fund, and a luxury tea brand

Before SparkPlug, Duffy collected an unusually varied set of credentials. He studied Psychology and Economics at Harvard, where he met Jake Sky Levin, the man who would become his co-founder and, by both their accounts, his best friend. The two have run SparkPlug on a deliberate 50-50 split ever since, a rare thing in a startup world that loves a single hero on the cap table.

From Harvard he went to Bridgewater Associates, the largest hedge fund on earth, and proceeded to do almost everything. He started in Client Service, helping sovereign wealth and pension funds understand markets. He built models for Japanese and Australian equities, for Mexican central-bank policy, even for the early cryptocurrency markets. He became the first product manager of the firm's Disruptive Technologies Lab and led a machine-learning team optimizing trading strategies. It was a finance education by total immersion, and it gave him a vocabulary for incentives, behavior, and systems that he has been spending ever since.

Then he left to sell tea. Sky & Wyatt was Duffy's luxury whole-leaf tea brand, complete with a sustainable supply chain, a direct-to-consumer community, and retail distribution. He has been spotted styling his title there as "CTeaO," which tells you something about how seriously he takes himself, which is to say, not very. The detour mattered. Selling a premium consumer product through brick-and-mortar shelves taught him exactly how much a motivated employee at the point of sale is worth, and exactly how little the existing tools did to find or reward them.

The Company

Commercializing the trick that sold the tea

When Sky & Wyatt sold, Duffy and Levin did not look for the next product to sell. They looked at the method. The incentive playbook that had moved their tea, the small, well-aimed rewards that turned ordinary shifts into selling competitions, was the actual asset. In 2020 they spun it out as SparkPlug, a platform that lets brands and retailers run gamified commissions, contests, goals, and product education for the frontline staff who represent them. A vendor can sponsor a reward. A store can run a leaderboard. A budtender can earn a bonus for hitting a real sales target, approved and paid out automatically, with the point-of-sale data doing the refereeing.

The founding mission is stated plainly on the company's own wall: realign the incentives of brick-and-mortar commerce so everyone can earn like an owner. Duffy's conviction is that businesses which invest in their frontline simply perform better, and that the people closest to the customer have been handed the worst tools in the building. SparkPlug is his attempt to flip that. The team he has assembled around the idea is a deliberate mix of behavioral economists, data scientists, market researchers, and retail veterans, which reads less like a typical SaaS roster and more like the faculty of a small, strange university.

In November 2022 the company closed an $8 million Series A led by Lightbank, with Industry Ventures joining alongside existing backers TenOneTen Ventures and the angel investor Jason Calacanis. It brought the total raised to $11.5 million, and Duffy announced it with characteristic dryness, calling it a round closed "in the most tumultuous fundraising and macroeconomic environment in recent memory." The capital went where you would expect from a systems thinker: executive hires, plus deeper investment in data, engineering, and behavioral-science expertise, the three disciplines that have to agree before any bonus check goes out.

The Worldview

Why a margarita streak became a mission

It would be easy to file SparkPlug as another piece of retail software, a dashboard with confetti animations. Duffy keeps pulling it back toward something bigger and more stubborn: a belief that the modern economy systematically undervalues the people doing its most visible work. He talks about burnout in retail and restaurants, about the churn that comes from treating frontline jobs as disposable, and about a future where those roles are understood as skilled labor worth competing for. The product is the argument. Pay people for the value they create, measure it honestly, and the rest, retention, performance, even dignity, tends to follow.

He is also a pragmatist about how change actually spreads. His pitch for adoption is almost mischievous: make the tool easy enough and the math obvious enough, he argues, and the demand becomes unstoppable on its own. "If we could make this scalable and make this easy," he has said of one category, "then I can't imagine an alcohol brand in the world that wouldn't want to be able to do this." It is the voice of someone who has watched incentives bend behavior up close, at a hedge fund, on a tea shelf, and on the floor of a Washington cantina, and decided to point that force somewhere useful.

The partnership with Levin is worth lingering on, because it is the quiet engine under everything. The two met as undergraduates and have been close ever since, close enough to bet their twenties and thirties on each other and to do it without the usual founder arithmetic about who gets the bigger slice. A 50-50 split is a statement of trust as much as a cap-table line. It also reflects how Duffy thinks about value generally: that fairness is not charity, it is good design. The same instinct that splits the company evenly between two friends is the instinct that wants the cashier to share in what the cashier creates. The logic scales down to a single shift and up to an entire company, and Duffy seems most comfortable when the two ends of that range agree.

There is a tidy irony in where he landed. Duffy spent years at Bridgewater learning to read incentives at the scale of nations, modeling how central banks and sovereign funds respond to pressure. He could have stayed in that world, where the units are billions and the players wear suits. Instead he pointed the same lens at the smallest, least-watched corner of the economy, the few square feet around a point-of-sale terminal, and decided that was where the interesting inefficiency lived. The bet is that the gap between what frontline workers produce and what they are paid is not a fact of nature but a measurement problem. Close the gap, and you do not just make a nicer workplace. You unlock real economic value that was sitting there unclaimed the whole time.

For all the behavioral theory, the through-line is human and a little sentimental. Duffy is still, at heart, the teenager who liked being noticed for selling the most margaritas. SparkPlug is the grown-up version of that recognition, built so the next kid on the floor does not have to wait for anyone to notice. The score is kept automatically now. And at the end of the shift, it pays.

In His Words

Six lines that explain him