57: Immad Akhund - The Art of Irrational Effort

1. Key Themes

Founder-as-Product-Spec: Building From Personal Pain, Not Market Research

Immad's entire approach to Mercury stemmed from being an entrepreneur first and a banker second. He built a list of everything that sucked about being a founder over years, and banking was just one item on that list. "So, but even then I was like, okay, I'm busy with my startup. So, so it wasn't until 2017, uh, when I, you know, had my list of ideas and this was one of them" 00:14:58. Critically, he didn't set out to invent a novel feature — he built the obvious product done well: "we didn't launch when we launched with like saying like, we do this thing that no other bank does. It was much more like we did the thing that you might expect your bank to do... we did it like much nicer" 00:25:51.

Banks Are Enterprise Lending Companies in Disguise

Immad reframes the entire banking industry through a single insight: banks are not consumer product companies, they are mid-market/enterprise sales organizations whose core product is loans. "Banks exist and make money on lending... 80, 90% of like the revenue most banks make... So these are really like mid-market enterprise sales companies selling loans as a main product. So if you put it in that context, then like I think the rest of why banks are kind of make a lot more sense" 00:21:10. This explains why deposit/checking products are neglected: "the deposit product is like... never been a place where they're trying to differentiate or be competitive. It's like the minimum possible... It's an alignment issue" 00:22:02.

Cohesion Compounds; Feature-Sprawl Decays

A recurring theme is that most companies degrade over time because they ship disconnected features that don't work together, while Mercury tries to do the opposite — every new product must strengthen the whole system. "Very few companies have stopped that march of like degradation... it's really important for me to think about like, how do I build a company that like gets better over time, ideally, not worse over time. And I think Apple is really good at this" 00:01:16. The bar for a new product is not novelty but integration: "why, why would this product work better at Mercury than it would work anywhere else? It has to be like a combination of we have the money, we have the data, we have the users" 00:38:42.

AI Collapses the Cost of Software, But Raises the Value of Everything Else

Immad argues AI is shrinking the "easy 50%" of building a startup (writing code) while the "hard 50%" — trust, brand, network effects, regulatory moats, thoughtful integration — becomes proportionally more valuable and differentiating. "Maybe when Mercury started, it was like 50% like software... And now... the first 50, like the building of the software gets to like, you know, 10X easier than it becomes like, you know, the other part becomes 95% hard... that's where like most of the value goes" 00:39:38. He also notes AI product tools only recently crossed a threshold: "AI for product development really only got amazing in like December last year" 00:45:06.

Betting on the Next Generation of Companies, Not the Existing Ones

Mercury's strategic logic is a bet on new company formation rather than displacing incumbents among mature companies. "About 25% of people do switch from whatever bank they're using, but actually the rest of the people are coming as the first company, first bank account kind of thing" 00:48:28. He extrapolates this into market dominance: "the earlier you are, the more likely you are to use us... for the next 10 years of companies, like hopefully half of them are using Mercury. And that would be like a really powerful position to be in... the future is defined by those companies, not the existing companies" 00:01:38.

Product-as-Trust-Building for Cross-Selling New Verticals ("Right to Win")

Mercury's expansion into personal banking, Mercury Books, invoicing, and bill pay isn't opportunistic bundling — it's built on a deliberate "right to win" framework: proprietary data advantage, distribution advantage, and inherited brand trust. "Is there like a distribution advantage kind of thing? And for us, it's like, Hey, most people set up banking before they set up an accounting solution... we can even help, right?... we'll start categorizing your transactions like from day zero" 01:07:29. On trust specifically: banking requires unusually high trust ("send me all your money"), and once won, it transfers: "a user's already in like a trusted relationship. They like more likely to be responsive to like products that also require trust" 01:08:52.

Frictionless, Human-Free Product Design as a Trust Signal

Immad's personal aversion to talking to people shaped Mercury's entire self-serve philosophy, down to the public demo environment. "I really hate talking to people if I don't need to... I hate calling service providers... I've always thought of Mercury as like the thing you come to if you don't want to talk to people... because it's like so easy to use" 00:57:31. The demo.mercury.com product, born from an internal front-end/back-end mock data architecture, became a powerful, measurable growth lever: "when people go to that demo before they sign up, I think there's like a 30 or 40% higher chance of them like becoming active" 00:58:27.

Culture as an Operationalized, Written-Down Personality Filter

Culture, to Immad, is not vague — it's a specific, hiring-enforced list of traits established at four employees. "It really came down to that for me. It's like, what are the personalities that I want at Mercury?... we wrote like six of these things down, which is like changed over time, but the nucleus has been like fairly similar... you've got to write them down. You've got to hire against them. You have to encourage them" 01:16:26. He explicitly contrasts Mercury's low-ego, helpful culture against a hard-charging alternative like early Uber's "crush your competitors" ethos, framing culture as a deliberate choice, not a universal ideal 01:17:22.

2. Contrarian Perspectives

The Two-Year Product Timeline Is a Feature, Not a Failure

While Silicon Valley celebrates speed, Immad deliberately spent a year and a half building Mercury pre-launch and two years building Mercury Books, believing the market rewards depth over speed for complex financial products. "So, you know, our perspective was really like you launch and that's when the work starts... we literally launched the first possible moment we could have launched. It just took a year and a half to get to that" 00:26:49. This runs against the "ship fast, iterate" dogma common in fintech and SaaS circles.

Fintechs Have Made Zero Progress on the Part of Banking That Actually Matters

Despite a decade-plus of fintech disruption narrative, Immad points out the entire category has ignored the actual profit engine of banking. "What's 80% of revenue in all financial services and fintechs haven't really made any headway into it kind of is like an interesting kind of DNA thing" 00:23:27. This is a quiet admission that even Mercury itself is playing in the "easy" 10-20% of banking economics (deposits/payments), not the lending core — a contrarian signal that most fintech "disruption" is cosmetic relative to the industry's real cash flows.

A 2% "Fuck Yeah" Response Rate Is a Green Light, Not a Red Flag

Conventional pitch-testing wisdom says lukewarm reception before launch is a warning sign. Immad's experience inverts this: talking to 100 companies pre-launch and getting only 2 enthusiastic responses was actually the correct predictive signal for Mercury's eventual explosive growth. "I think the interesting thing I learned there is like, actually if 2% of people are like, fuck yeah, that's actually a lot of people if the market's big enough... it has to be the 2% that really want it, like have to really, really want it" 00:00:28.

The Bank Sponsorship Path Beat the "Real Bank" Path for 8 Years — Deliberately

Most founders assume acquiring full regulatory infrastructure (a bank charter) early is de-risking. Immad explicitly rejected this, arguing product-market fit must be proven before absorbing that fixed cost. "It wasn't even obvious that... you have to build a great product that people love. And it's not obvious that like you have to do all this like fixed cost work of like getting a bank charter before you can even figure out whether the product is anything people want" 00:32:17. Mercury only pursued a charter after achieving massive scale — inverting the "derisk regulation first" instinct common in fintech.

Talking to Customers on the Phone Is a Bad Metric, Not a Good One

Traditional banking (and much of enterprise SaaS) treats "relationship" and "conversations with customers" as a positive KPI. Immad flags this as actively broken: SVB relationship managers were "literally optimizing to talking with their customers" 00:57:57 as a success metric, while Mercury optimizes for the customer never needing to talk to anyone at all — treating human-required support as a product failure rather than a relationship win.

3. Companies Identified

Mercury — Banking platform for startups and businesses (in process of becoming a chartered bank). Central subject of the episode, discussed extensively as an example of product cohesion, culture, and customer-obsessed design. "One in three startups use us" and "more than 40% of like pre-seed companies use Mercury" 00:49:08.

Silicon Valley Bank (SVB) — Mentioned as Mercury's former head-to-head competitor for early-stage startup banking market share before its collapse. "When the whole SVB thing went down, we were like pretty head to head in terms of like market share in the early stage space. I would say like, you know, we were like 20-ish percent. And they were like 25%" 00:51:44. Also cited negatively for optimizing relationship managers around customer call volume rather than customer ease.

Starling / [Y Combinator debit-card-for-seniors company] — A small YC company (two founders) that built a debit card for seniors using a sponsor bank deal with Sutton Bank. Mentioned as the pivotal proof point that made Immad believe a small team could realistically enter banking. "It was a debit card for seniors... it was literally two people and they went and did the sponsor bank deal... at that point I was like, okay, maybe I could do it" 00:14:19.

Sutton Bank — Named as the sponsor bank behind the above YC company's debit card product, illustrating the sponsor-bank ecosystem model Mercury itself later used.

Apple — Cited as the gold standard for a company that avoids product/quality degradation over time. "I think Apple is really good at this. Like the... whatever you launch, you should never, you should have a high bar for it... you should feel like you're giving delight with everything you launch" 00:41:02.

Amazon — Referenced as a counter-model of "two-pizza team" independent domain launches (contrasted with Mercury's more integrated approach). "A lot of things that Amazon does, right? They like, they launch a new domain and like, that's like, you know, like it's a true, like two pizza team... that's not that hard and maybe we could do more of that" 00:38:13.

Uber — Cited as an example of a different, equally valid but distinct cultural model ("crush your competitors") contrasted against Mercury's low-ego helpfulness. "My sense of Uber is like it's at the start, it was like... do whatever it takes to win, crush your competitor. And like, obviously worked for them" 01:17:22.

Stripe, Gusto, Slack, AWS — Named collectively as examples of startup tooling that got dramatically better between 2006 and 2017, everything except banking. "In 2006, everything was bad. Right. There was no Stripe or Gusto or Slack" 00:13:26.

QuickBooks — Referenced as the entrenched incumbent accounting software Mercury Books must displace, illustrating the switching-cost challenge. "It'll be quite hard to get you to switch from QuickBooks or whatever you're using" 00:48:28.

4. People Identified

Immad Akhund — Founder/CEO of Mercury; serial entrepreneur (fourth company, building startups since 2006). Central figure of the episode; his personal history (developer-tools founder, financially strained through early startups, deep personal aversion to bureaucratic friction) directly informs Mercury's product philosophy. "I've been doing startups since 2006... this is my fourth company" 00:09:37.

Immad's father — Ran a garage in Pakistan; unable to hold down menial jobs in London because "I can't work for people." Cited as an early formative influence on Immad's own inability to work for others. "It just like really stuck at the back of my mind that like, oh yeah, maybe I can't work for people either" 00:11:01.

Immad's sister (older) — Cited as the source of his personal definition of "helpful" as a life purpose. "If you're not going to be helpful, like, well, you know, what is, what is like, what's your purpose?" 01:18:23.

Ben Horowitz — Referenced as author of a culture book Immad considers the best he's read on the topic. "Ben Horowitz has a great book on it. Like who you are is what you do... that's probably like the best culture book I have read" 01:16:26.

Stewart Brand — Referenced (via the host) for his writing on maintenance and the effort required to resist entropy/decay, used as a parallel to Immad's philosophy on product decay. [01:41:00 context, host reference]

Sonia Huang (Sequoia) — Quoted by the host regarding software costs trending to zero and the resulting value of scarce assets like bank charters, used to frame the discussion of bundling. "As the cost of build software asymptotes to zero, we should expect to see more bundling and consolidation of previously independent categories of software around the things that are actually scarce post-AGI, like a bank charter" [00:35:51 - host paraphrase].

Ashwin — Twitter/X commentator quoted around the Mercury Books launch, noting "Mercury's unfair advantage is that they just do the obvious things that every bank should have done 15 years ago" [00:54:01 - host paraphrase].

5. Operating Insights

Build Front-End and Back-End Independently to Enable Free Optionality Later

Mercury's demo product wasn't planned — it emerged because the engineering team had architected mock data separately from the live backend for development speed, and that decoupling later became a major, measurable growth channel. "We wanted to be able to like develop the front end and the back end separately... it just so happened... we'd built up this like set of mock data and like a front end that were completely independent of the back end" 00:56:23. Operating lesson: infrastructure decisions made for engineering velocity can produce unplanned product/growth assets years later.

Use an "Investment Model" to Gate New Product Bets Like a Portfolio

Immad manages new product lines the way an investor manages a portfolio — pre-PMF products get minimal resourcing, and investment scales only with demonstrated traction. "This is like books, like before it launches, like a pre-park market fits startup. Like how much do we invest in that based on the idea and whatever. But then like as these things like do well, we invest more in them" 00:42:52. This caps simultaneous big bets deliberately: "there's only so many new things you can have going at a time."

Structure New Verticals as Autonomous Units That Don't Cross-Distract

When launching personal banking or Mercury Books, Immad explicitly designs organizational boundaries so that teams don't need to coordinate constantly with the core business. "Think about them as like autonomous kind of units that can like go after like a separate vision to what the rest of what we're doing is" 01:04:45 — while still sharing the underlying product backbone. This balances speed (autonomy) against cohesion (shared infrastructure), which is a nuanced operating tension most companies get wrong in one direction or the other.

Interview for Product Sense Across All Roles, Not Just PM/Design

Immad runs a product-thinking interview exercise for candidates in non-product roles as well, testing whether they can reason through feature development scenarios regardless of function — a signal that product judgment is treated as a company-wide competency, not a departmental skill. "I have a product interview that... I do for almost everyone, including non-product roles... if you were developing this feature... how should they develop it?" 01:20:35.

6. Overlooked Insights

The 90-Day Transaction History Limit Reveals How Legacy Bank Software Monopolies Shape "Normal"

Immad's offhand observation that most banks cap visible transaction history at 90 days isn't a deliberate policy — it's a shared technical limitation baked into legacy core banking software vendors that the entire industry runs on. "They all use the same backend software and that's been a limitation from like 20 years ago or something. But yeah, there's no law about it for sure" 00:54:16. This is a strong, underappreciated signal about market structure: nearly the entire US banking industry's user experience is bottlenecked by a small number of decades-old core-banking software vendors, meaning any fintech that builds independent infrastructure (as Mercury eventually will with its charter) has an enormous, structurally-protected UX advantage that has nothing to do with "innovation" per se — it's simply not being shackled to 1990s-era mainframe cores. This is a much bigger structural moat opportunity than the conversation's framing ("banks are just lazy") suggests.

Sponsor Bank Ecosystem Peaked and Is Now Shrinking — A Signal of Fintech Consolidation

Buried in the middle of the regulatory discussion is a striking data point: the sponsor-bank ecosystem that enabled the entire 2017-2021 fintech boom grew to roughly 120 partnerships and has since declined. "It probably like a lot more came online and did a lot more partnerships and it probably peaked 120. And that's actually kind of somewhat declined since then" 00:31:19. Neither speaker dwells on this, but it implies a meaningful structural contraction in the neobank/BaaS space — likely tied to increased regulatory scrutiny post-2023 (the era of Synapse's collapse and increased OCC/Fed enforcement against sponsor banks), suggesting the "build a fintech on top of a sponsor bank" playbook that created hundreds of startups in the last decade is now materially harder and more consolidated, favoring incumbents like Mercury who have the scale to pursue direct charters.