Mercury 曾公开说不需要 PM，后来却建起了产品组织和多产品引擎

Rohini Pandhi explains that before roughly two years ago Mercury had no one with the official PM title, even as it approached 400 employees and roughly half in R&D. CEO Immad had publicly celebrated not having PMs, and business leads absorbed cross-functional coordination while engineers and designers took on PM duties. Pandhi argues this can work because in seed-stage companies the founders should be the PMs until they literally cannot do the job anymore. Mercury’s early success came from hiring product-obsessed engineers and designers and giving autonomous teams room to make local decisions. Today Mercury has about 30 PMs, with more than half senior.

The trigger to hire PMs was scale plus complexity, not a sudden belief that PMs are inherently good. Decisions bottlenecked on co-founders; fintech work required compliance, legal, and risk reviews; and design and engineering were spending time on PM work instead of their crafts. Pandhi says the breaking point likely arrived before the 100-employee mark, and Mercury created “business lead” roles to keep projects moving. Those business leads eventually became PMs about two years ago, but they had not been trained in strategic planning, roadmapping, prioritization, goal setting, or deep customer research.

In hindsight, Pandhi says Mercury should have made the official transition to PM roles sooner. The lesson is not that every startup needs PMs early; it is that founders should watch for the moment when they become the bottleneck, or when other people are doing core PM work at scale. A great PM can make product velocity and quality dramatically better, which is why the common complaint that PMs slow teams down usually reflects a team that has not worked with a great PM.

When Pandhi joined about a year ago, PMs were asking basic questions about what good looked like, so her first step was to define the discipline internally. Mercury built a PM career ladder that combined company-wide values such as humility and low ego with PM-specific expectations. The ladder had two main vectors: building products customers genuinely love, and making work that has a positive, outcome-based impact on the business. It also covered long-term vision, communication, and cross-functional leadership in a regulated environment.

Mercury then rewrote hiring and interviewing to test the same attributes: product sense, customer and quality obsession, analytical skill, and strategic or system-wide thinking. Pandhi wants to add a written interview component because writing demonstrates depth and the ability to put thoughtful thoughts on paper, especially for senior PMs. She borrowed from Block, talked to senior product leads at other companies, and worked with Mercury’s executive team to separate table-stakes PM skills from company-specific superpowers. Her professor’s line—‘well-stolen is half done’—captures the approach, provided the borrowing is deliberate.

The interview process includes a generic Mercury presentation prompt: talk in depth about something you are interested in. Pandhi says the panel is testing attention to detail, craft and care, and the desire to go deep; one PM presented on an Airbus or Boeing plane from structure down to the bolts and why they mattered. On timing, Pandhi advises founders not to hire PMs too early. The signal is when work is not getting done without the founder, when design and engineering are doing core PM work instead of their own, or when the company moves into adjacent product areas with different customer segments.

Pandhi rejects the idea that ‘PMs suck’ based on one bad experience because product management has many flavors. She uses a Square-learned framework: pioneers are strong at zero-to-one work, building from nothing without infrastructure; town settlers grow a product that has some infrastructure but still needs experiments and foundations; city planners excel at mature products and efficiency at massive scale. Founders who say they do not want PMs may be imagining a city planner when they actually need a former founder or pioneer for the next product line.

The right mix depends on the bet. Mercury’s core banking is mature, so the PM there must measure three times and cut once, and the company needs people who have operated at that scale. Smaller companies may lean toward pioneers, but Pandhi warns against hiring from resumes alone. Interviews should drill into the kind of experience a candidate is actually good at and what they want next, because ‘the resume without the receipt is just a piece of paper.’

Answering a question from Mercury COO Jason Zhang about what attracts great PMs, Pandhi starts with table stakes: viability, market pay, product resonance, and culture fit. Then she thinks about recruiting the way a PM thinks about a customer. A mentor’s S-curve framework plots ambiguity on the Y-axis and scale on the X-axis: early-career PMs start with one low-ambiguity project, then juggle multiple low-ambiguity projects, then hit an inflection point with a highly ambiguous problem such as 10Xing revenue in 12 months, and eventually handle ambiguous strategic problems at scale as a VP or CPO. Companies should know where a candidate sits on that curve and what challenge would move them forward.

Senior PMs want to know whether the problem area is big and meaty enough to challenge them and make them better, and who they will get to work with. Pandhi compares it to tennis: you improve by playing with people a little better than you, and PMs want strong design and engineering peers in the same functional structure. She is transparent about the good, the bad, and the ugly, while still selling the dream of why the work matters. Candidates are also testing the company through the interview, which is why Mercury revamped prompts and questions based on Pandhi’s own experience as a candidate.

Pandhi argues that the choice between metric-chasing experimentation and design-heavy artistry is a false dichotomy; the best companies sit in between. In a ‘boring’ category such as fintech, precise and intentional details breed the trust customers need to save and move money. Mercury’s Bill Pay product lets customers drag and drop an invoice, OCRs it, pre-populates fields, and zooms the right-hand screen to the relevant part of the invoice as they check the data. That detail may not move a single completion metric, but customers specifically tell Pandhi they loved using a product to pay bills.

The compounding effect of many small, intentional improvements becomes a competitive advantage, and it also attracts talent and endears customers. Pandhi says quality has to be a cultural value that comes from the top; a PM cannot bolt it on if the founders do not believe it. She also sees B2B product design moving beyond simply making things look like B2C: the next generation must be simple and easy to use while still powerful enough for enterprise work on a small mobile screen. She points to ugly, billions-dollar incumbents in categories such as procurement as an opening for companies that compete on experience rather than more features.

Mercury now thinks of itself as more than a bank. In 2024 it launched Mercury Personal Banking in April, Bill Pay in May, Invoicing in August, and updated spend controls, spend management, and employee reimbursements in October. Pandhi’s favorite signal is that each subsequent launch took less time than the previous one from concept to launch, as the company learned how it builds and launches new products. The bigger lesson is organizational: new product areas placed too close to core banking or payments get pulled into mature-product prioritization, maintenance, and impact work.

Pandhi’s expansion org exists to give seedlings air, like a small fire that needs oxygen to become a flame. Mercury treats these teams as seed-stage companies, reviews every quarter or half-year whether to keep investing, and sets goals and financial projections without killing a team for missing a number if the learning points in a promising direction. Seed teams are full cross-functional squads of engineering, design, product, and data science, with central CS, risk, and ops teams leveraged where possible. Personal banking needed roughly 3X the engineers of the invoicing line because it went after a different market and had to build more banking infrastructure.

The operating cadence differs from mature products: the expansion teams moved to two-week sprints and weekly check-ins instead of monthly development cycles to stay nimble. Customer contact is the heartbeat; Pandhi expects at least one customer conversation per week, and teams in discovery may do 20 to 25 calls. Mercury uses a UXR team, dedicated research weeks, and sidecar participation in relationship-manager, sales, and CS calls to keep the voice of the customer close. Pandhi warns against confirmation bias: customers are often polite and may blame themselves for UX problems, so PMs should observe like a documentarian, look for repeated patterns, and validate them with prototypes, betas, and numbers.

For the next product line, Mercury looks for a solution adjacent to its existing space, not just a new feature. The team starts with a target segment or ICP, asks what adjacent segment would want a deeper product experience, and de-risks the bet by looking for early traction. Invoicing is the example: Mercury already had Payment Requests, where a customer could send a URL and get paid directly to their bank account, so it could talk to people already doing invoice-like work. That gave Mercury both a customer signal and a distribution advantage: it could test a fuller invoicing solution with existing Mercury customers instead of running a full go-to-market effort.

Mercury also uses its current UX as distribution. When a customer is sending a payment, it highlights Bill Pay if they need more complex vendor payments, fitting into the natural flow rather than forcing a separate sales motion. Pandhi says launch is only the first hurdle; the next is finding product-market fit. The company deliberately does not rush from launch to revenue optimization, and is willing to give away the second or third product for free to encourage usage, engagement, and retention while customers fit it into their workflows.

Pricing still matters, but it is tied to the customer journey and willingness to pay. Core banking and core money movement should stay free; invoicing metadata on top of a money movement can stay free; a fuller invoicing solution that works with accounting and a CPA becomes paid when a customer needs that complexity. Pandhi says Mercury did a willingness-to-pay process and targets the customers at the higher-complexity end, not everyone using its banking product. She also drew on Madhavan’s pricing conversation—water is free at a fountain but costs more at a hotel or on a plane—to show how context changes value perception. The missteps she names are trying too many things at once, understaffing new efforts so seedlings do not get enough care, and wrestling internally with pricing model questions such as bundling versus à la carte.

Outside Mercury, Pandhi co-founded Transparent Collective, a 501(c)(3) nonprofit that helps underrepresented founders access resources and connections. It grew from a conversation with a co-founder from Michigan about how fundraising in the Midwest, New York, and the West Coast differ, and the desire to share tribal knowledge transparently so one generation of founders could help the next. Its spotlight program flies 8 to 10 founders from across the U.S. to San Francisco for a week of mentoring, workshops, and mock investor meetings, selected from 200 to 250 applications. Pandhi says the organization has 90-plus alumni across 11 batches, about 65% to 70% have raised venture funding, total seed or early-stage fundraising is around $125 million, and about 10% of alumni companies have been acquired or exited.

In the lightning round, Pandhi recommends Vectors: Aphorisms & Ten-Second Essays and The Inner Game of Tennis, the latter for its idea of relaxed concentration and flow. She also names fiction such as Pachinko, Cutting for Stone, and Song of Achilles, and says she has learned to stop books she cannot finish. Her recent TV favorites are Apple TV shows including Shrinking, Bad Sisters, Slow Horses, Severance, and Presumed Innocent; her favorite product remains Waymo because the self-driving experience becomes comfortable almost immediately.

Pandhi shares a life motto from Adam Robinson: connect with others whenever possible, create fun and delight for others with enthusiasm, and lean into each moment expecting magic or miracles. As an angel investor, she gravitates toward product-minded founders and cites Payitoff, a student-loan repayment company that was acquired, because the founder showed deep care for a real financial pain. She can be found on X at @rohinip and on LinkedIn, Mercury is hiring at mercury.com/jobs, and she asks listeners to keep sending product feedback; Mercury also offered 25 podcast listeners a chance to skip the waitlist for Mercury Personal Banking. Transparentcollective.com has donate and application buttons for people who want to support or apply.