Zach Perret on Founder Mode, Product Expansion and Building Plaid

Plaid co-founder and CEO Zach Perret joins Harry Stebbings to discuss product expansion, management, financing and investing. His comments on Plaid’s prospects and valuation reflect his role and financial interests in the company.

Perret recalls agreeing to sell Plaid to Visa in 2020, followed by US regulatory scrutiny and strong growth during the pandemic. He says Plaid had the opportunity to walk away at the end of the process and chose to remain independent; the proposed sale did not close.

For him the difficult question was whether a sale would limit the company’s long-term mission. He describes the original decision to sell as close, and the subsequent decision to remain independent as much clearer after the business grew.

Plaid had already begun developing new products before the proposed acquisition. In hindsight, Perret says launching three new business areas, expanding internationally and moving upmarket simultaneously stretched the organization; he would sequence those changes.

Go-to-market broke first: competing product teams wanted sales attention, while sellers needed help explaining unfamiliar products. Moving from one product to several required specialists, sales engineers and a more complex operating model.

Perret sees resource allocation as increasingly central to a CEO’s work as a company matures. At an early stage finding product-market fit may be the overriding task; later, allocating people and capital among opportunities becomes crucial.

For new ideas he favors a small “atomic team”: the minimum mix of product, engineering, design or data skills needed to test a hypothesis. Give it resources, protection from distractions and one or two design partners willing to experiment.

He prefers adding resources when evidence warrants it rather than making teams wait for a rigid internal funding cycle. A team might quickly discover a real customer need or conclude that an idea will not work; either result should inform the next decision.

Perret calls Plaid’s early bank integrations a “grinder problem.” Banks often lacked APIs, so the company built and maintained many different screen scrapers. He argues that sustained difficult work can create a product competitors are unwilling or unable to replicate.

He agrees that very early startups may not yet have defensibility, although experience can create an advantage. What matters to him is whether the work can eventually produce durable differentiation through scale, brand, network effects or another mechanism.

Grinding alone is not valuable if customers do not care. He contrasts fragmented banking markets, where many integrations can matter, with markets dominated by a few banks, where financial infrastructure presents different challenges.

Perret initially disliked recruiting but came to treat it as a competitive search: find a strong person, determine whether they are right for the role and persuade them to join. He believes a founder must learn to recruit well to build a strong company.

His outreach is short and specific: explain what impressed him about the person, mention the opportunity and ask for a conversation. He says many people ignore or decline a cold email, but repeated, genuine outreach can work; for recruiting he favors fewer than five sentences.

Plaid hires for “spikes”: exceptional strengths in some areas, balanced by complementary teammates even if a person has weaknesses elsewhere. He values unusual backgrounds and different ways of thinking as part of that mix.

He has changed his mind about experience. Early Plaid was wary of very experienced hires it thought might lack urgency, but in specialized areas such as fraud engineering he now sees deep experience as valuable, provided candidates also fit the company’s working values.

A recurring hiring mistake is rushing because a search has taken too long. Sometimes a necessary role must be filled temporarily, he says, but the company should recognize the compromise and be prepared to revisit it.

On Paul Graham’s “founder mode” essay, Perret sees useful ideas but worries it could be misread as a license to undervalue executives or micromanage everything. He advocates effective delegation: stay close to important decisions without trying to control every detail.

A practice that worked for a particular leader at a particular company may not transfer wholesale. Perret says founders should apply advice to their circumstances, revisit decisions and identify where their company genuinely needs a distinctive approach.

For instance, he does not believe every enterprise startup must immediately hire a vice president of sales. He led sales for a long time with a small supporting team, which he says helped him learn about customers and product needs.

He also questions applying OKRs mechanically to software teams. Goals matter, he says, but founders should choose a system suited to their business rather than copy another company’s process without reflection.

Asked whether speed is most important from zero to one, Perret puts being right first. Before product-market fit, teams need to gather evidence and choose a promising direction; once they have it, they should execute quickly.

Perret tries to raise funding infrequently because fundraising consumes time. Plaid sometimes went two and a half or three years between rounds; he says early operations happened to be profitable, so the company did not always need external cash.

He treats money as a resource to obtain when it helps solve a real problem, with efficient terms. He disagrees with the idea that a founder must always be fundraising, though the appropriate timing and dilution depend on a company’s circumstances.

His hardest round was the seed. After failed attempts at a consumer budgeting product, the founders pivoted to the underlying infrastructure. A planned $500,000 round fell apart when the lead backed out, leaving the company nearly out of money.

Perret says three investors named Justin collectively invested $60,000 after the founders explained the lead had withdrawn. That bridge kept Plaid alive until it gained customers and raised a later seed round; he recalls personal debt and months without pay, but does not recommend that risk.

Following the abandoned Visa deal, Perret wanted as many employees as possible to participate in a secondary sale. He says a large majority could participate, though not everyone, and that the amount employees could sell was limited.

Employees had understandably imagined the liquidity from the proposed $5 billion sale, he says, and walking away created disappointment and uncertainty. The later secondary offered some liquidity, but he did not promise an amount he could not guarantee.

Asked about the subsequent $13.4 billion valuation, Perret says private company value is difficult to determine and technology multiples have changed. He wants investors to earn returns and believes Plaid can grow into the hurdle, while acknowledging that will take time; this is his view, not a valuation guarantee.

Perret says the proposed Visa arrangement would have left Plaid operating as an independent unit with resources, but he prefers the independent path. He describes large M&A as difficult under the regulatory environment at the time and does not predict when it will change.

On an IPO, he says Plaid would like to become public over the long term but was not in a specific rush in October 2024. He views an IPO as a financing event and milestone to use when it benefits the company, not an announced schedule.

Perret uses a hypothesis-driven decision style: offer an initial answer, then revise it as evidence arrives. He wishes colleagues would ask how confident he is and how much he cares, since a founder’s confident-sounding first reaction can carry more weight than intended.

He identifies a leadership weakness: failing to recognize progress when the result is still short of excellent. Going from two out of ten to four is an improvement worth acknowledging, he says, even though four is not the desired final standard.

A founder friend suggested asking teams what first milestone would show they are on track toward a larger goal, then celebrating that milestone. Stebbings compares it to breaking a marathon into five-mile intervals and rewarding progress.

Perret says his independence helps him think clearly alone but can make him difficult for others to read. He tries to share more of what he is thinking; quiet time on a plane with a notebook remains valuable to him.

Fatherhood changed him personally, he says, especially by showing him a new side of his wife as a mother. It made him more attentive to how people he knows can grow and change over time.

He defines success as excellence in the craft rather than simply being bigger or faster. Though competition can be motivating, he increasingly compares Plaid with what it could be at its best instead of focusing on rivals.

Stebbings challenges whether investing outside Plaid distracts from the company. Perret says angel investing originally kept him close to early-stage founders, provided ideas and helped him meet people, but also demanded substantial personal time.

He co-founded Mischief with partners who operate the fund full-time; he says he spends a few hours a week on it. He argues that the structure lets him learn from founders, share lessons with Plaid and mentor others, while acknowledging the distraction concern.

He expects average venture returns to face pressure if capital grows faster than the supply of exceptional companies. This is his investment outlook, and he has a direct interest as a venture fund co-founder.

He also recalls a Plaid funding round at risk of falling apart when NEA chose to support the company with additional capital at a higher valuation. He credits that decision with helping the company; NEA is a Plaid investor.

In quick-fire advice, he warns against applying OKRs or revenue targets too early, and against delegating to executives without staying informed. He reiterates that specialized experience can matter and describes promising consumer experiences in AI, smart glasses and one-click checkout.

Looking five years ahead, Perret hopes consumers can use their financial data to access credit more easily and that fraud can decline. He imagines a one-tap loan application; these are company ambitions, not proven outcomes or guarantees of credit access.