Braden Dennis — How One Frustrated Investor Democratized Wall Street's Data

0:00 Cold open — "you own the data"

• The first five minutes are a preview clipped from the body of the interview (it repeats verbatim at 21:52–27:24): the moat is owning the data content rather than licensing it, and the reason it is possible now is an AI reasoning layer replacing a manual data factory.
• The host's summary of what changed: "the world of investing is changing at the fastest pace that I've ever seen. Not the core principles, but how we get there and how thorough we can be about research."

7:56 The host's frame — Ben Graham's 1955 Senate hearing

• Baranowski opens on Graham's testimony about "a market mystery that's resolved eventually by value recognition", and casts better tools as the mechanism: "the more people are capable, knowledgeable, and have the right tools, I think the markets can work just a lot better."
• He returns to it at the close — "I think Ben Graham would have been really impressed with what's available to us these days compared to what was available when he was playing this game."

9:02 A math kid outside Toronto, and engineering for optionality

• Middle-class upbringing near Toronto; 30 years old now; "I knew from a very young age that I was a math kid" and carried it into an engineering degree "knowing that that would serve probably a career with the most optionality."
• On why numbers grabbed him and essays didn't: "the feedback loop on something that is numerically correct versus abstract opinion has completely different incentives on how you want to study or better understand the rules of the game."

10:29 The mission — capitalism is not zero-sum, so own assets

• The formative observation, from a large extended family: "the people in my family… that were doing the best were not the ones that made the most money" — the ones investing and acquiring assets were.
• His stated mission beyond the product: "you can benefit from this system much much better if you own assets… you feel like the system's cheating you if you don't own assets."
• "Capitalism is not a zero-sum game. It's not because I made money and because I invested that means I inherently stole it or took it from someone else."

13:02 The host's formative book — One Up on Wall Street

• Baranowski's first investing book was Peter Lynch's One Up on Wall Street, and the line that stuck was ownership: "I can become an owner of a business, a couple of shares with very little money, and I'll be on the same boat as the people that founded this business."
• His answer to the "is now a good time?" question he gets in his Talking Billions office hours: ownership of assets "is a lifelong pursuit."

14:27 The leap — build it on nights and weekends first

• "That's all I could really think about was not if, but when" — but he had already built enough on evenings and weekends "where I wasn't having to take a complete leap into the unknown."
• The advantage he thinks people misprice: "there's just not that many people willing to do the work on some of this stuff… if you are, you're already kind of filtered out."

15:57 Persistence math — episode 21, and Mr. Beast's hundred videos

• He has been podcasting about stocks for roughly ten years. The stat he keeps: "back in 2019 it was 95% of podcasts did not hit episode 21. So that means you're in the top 5% if you just get to episode 21."
• The Mr. Beast version: "I will give you free consulting and help you make videos after you've made a hundred" — and nobody comes back, because by then they have learned it themselves.

18:12 Chapter one — Yahoo Finance on steroids

• Three distinct eras; they are in chapter three now. Chapter one's brief: "can I build Yahoo Finance on steroids? A non-ADR-riddled experience, good data, not being limited, not feeling so retail, but still so accessible."

19:05 Chapter two — the viral AI chat product, 60,000 signups in 48 hours

• They hooked the ChatGPT LLM API into a finance product and "it went viral overnight… we got like 60,000 people signed up in the first 48 hours," with venture investors calling.
• Not just the first application of the technology in finance but "one of the first applications of that technology period, in any vertical."
• What they concluded anyway: retrieval, manipulation and summarisation "was probably going to commoditize to zero as the LLMs became better and bigger and cheaper and faster."

20:09 The arms dealers of a $46bn industry

• Rebuilding "the airplane engine as it was flying" — the entire data infrastructure — while the chat product ran.
• The realisation: "these five big multi-billion dollar incumbents are the arms dealer in this industry… taking most of the unit economics of the entire industry" — and Fiscal was a customer of their data feed too. "Financial data is a $46 billion a year business."

21:15 Killing a working product to go after the data layer

• Contracts about to be signed that "were going to take the business to multi-millions of ARR from say 1 million" were cancelled — "we said no to all of it."
• The trade: "there's a much bigger prize of solving the data layer here." Chapter three is "can I build AI native FactSet better, faster, cheaper across the board. And the answer is proving out to be yes."

22:17 Why owning the content is the moat — the buy-sider's compliance test

• Why every "Bloomberg killer" has failed to take material share: "they are licensing that data and don't own the content."
• The objection that killed deals before: "I love the UI. I love the UX. This is way better, faster, cheaper than FactSet. But you don't have the ability to audit the number where it came from. My compliance team says I need this to subscribe… I'm a buy-sider. I need to put my neck out on this. I need to be able to click through to the filing."
• "Before we owned our own data content, there's no technology that could solve that problem. There's nothing we could code up to solve that problem."

25:42 Why now — a reasoning layer replaces the data factory

• The incumbent process, described from the inside: "Joe, you're taking the income statement this company just reported. Sally's taking the cash flow. I got the balance sheet… and by the way, we have 200 other companies that just reported" — which "leads to slowness, inaccuracies, shortcuts being made, especially with some of the smaller companies."
• What is new: "with AI I have a reasoning layer on what we're doing when we're extracting the data and properly doing the standardizations into the different industry templates."
• Why nobody funded a challenger before: "I'm going to disrupt one of these companies. I just need a billion dollars so I can hire 10,000 people. That's not getting funded, and that's why the data moats were so strong."

28:11 The scale — 96 terabytes, 2.4 million SEC filings

• The host's numbers, unchallenged: 96TB of data stored and 2.4m SEC filings. "A human being cannot process it. Just cannot."

29:49 Two days to two minutes — and the three ways data is consumed

• "Those three key areas people consume it are on a terminal, in an Excel add-in type product, or on a data feed API. And we play in those areas." Other fintechs — "say you're the Robinhood of country X, Y and Z" — license the feed.
• The incumbent wait is "on average two days, sometimes longer for small caps" for cleaning, standardising, consolidating and QA. "What takes them two days, we do in two minutes."
• "Why should I wait two days for public information that the market has already gotten?… How you actually bake the cookie matters in terms of how it tastes."

35:34 No scoring, no verdicts — the feature customers ask for and don't get

• "We don't provide any sort of scoring. We've steered clear of all of those features that our customers have asked us for, by the way. It's so important to listen to users. It's also so important to know when to ignore their recommendations."
• "I don't want to build a scoring system that tells you if a stock is good or bad or valued or undervalued. I'm just here to present you all the information."
• Where he thinks AI does not belong: "letting GPT manage your portfolios hasn't really been linked to good returns and active trading doesn't necessarily lead to great returns." The host's phrase for it: the technology "doesn't allow you to outsource judgment."

38:25 Company-specific KPIs — the Substack-screenshot problem

• The origin of the feature that became his differentiator: reading thoughtful research on Substack and noticing the author had pulled the real business metrics into Excel by hand, "because no one had done it for them."
• "I thought it was just crazy that the core fundamentals of the company were not accurately being tracked anywhere for people to get up to speed on something."

39:50 The Uber reversal — take rates doubled while trips tripled

• He was wrong out loud for two years: "Uber is a ZIRP phenomenon, venture-backed subsidized, terrible unit economics business."
• What changed it was two KPIs moving the wrong way for that thesis: "take rates doubled while the trips tripled. That's a rare kind of KPI you see from a business, to be able to double their pricing power and usage triple."
• He bought before the consensus turned — "back then the consensus on the street was that Uber's never going to make a dollar. And that's not that long ago." His point about accessibility: "you don't have to be an investment analyst to really understand those core business metrics."

41:11 The exit trigger — when they stop disclosing the number

• The KPI is the holding thesis and the sell discipline: an autonomous-vehicle threat "I'm going to see it there. That's where you're really going to see things get dicey."
• "Or even worse, they stop reporting or stop disclosing the number. We flag right on the platform that they've discontinued that number. So if that were to happen, I would be very much so looking for the exit before other people are."

43:17 Fundamentals only — technical analysis as "astrology"

• The product is "completely useless if you're looking to draw some technical analysis on a chart and make some sort of prediction on price movement… a lot of that stuff that I call astrology more than investing, to be honest. We don't provide any of that."
• His observation about the two user cohorts: a professional running billions and an individual running their own money are "looking at the same things… the use case is identical."

47:21 The eight-things template — simplify in order to hold

• Running the business forced him to turn over more stones, but it also "simplified my process a lot… there's just a few key things that I really want to track."
• The pattern he built a feature for, from users who work the same way: "these are the eight things I care for every company, I want to create that as a template as a starting point for each name."
• Why it matters: "you can drown in — I love that I have the ability to pull all the data, but… I also want to just pull up a view and be narrowly focused on what I think matters."

48:36 Visa and Mastercard — two metrics that outrank every headline

• The disclosed position: "I've been a long-term shareholder of Visa and Mastercard… probably two of the most brilliant businesses ever created."
• The noise he is deliberately ignoring: recent negative news about a potential cap, which would "massively hurt the banks that administer those cards and… the end consumers."
• The dashboard that decides whether he cares: combined total transaction volume "at the 27 to I think combined 30 trillion scale" plus total cards in force across Visa, AXP and Mastercard. "If anything is threatening those two metrics for these three companies, then I'll pay attention."

50:22 Decades of holding, checked quarterly

• "Until then, I'm just holding them for, you know, decades potentially… finding the simplicity in the data as well too is really what makes you a long-term patient investor."
• The cadence, in his words: "we're all tracking the score every quarter… the right amount of frequency to stay on top of the story every 3 months without being spoonfed it every single day so that you have decision paralysis."
• The host's version of why this matters: "there's no shortage of great ideas… There's a shortage of the ability to hold them long enough. And it's not just patience, it's the conviction."

52:48 The walk-through — financials, then segments and KPIs

• Overview tab: top-level ratios, business description, trailing-month news, a bulls-bears summary. Financials tab is "the guts": three statements, ratios, KPIs, adjusted and custom metrics, forward estimates, 20 years of history, switchable to quarterly.
• Click any number and "boom, it builds you the exact data visualization" — overlay revenue with diluted EPS, or a KPI against revenue on a separate axis.
• Segments and KPIs: revenue by geography and by line item, plus the reported operating metrics — "16.7 trillion dollars of transaction volume just really gives you a sense of scale" for Visa.

57:08 Filings and IR — query the document, not the whole model

• The "investor relations" tab holds the unstructured content — PDFs, reports, earnings-call transcripts — in one viewer, with a side panel for one-shot or custom AI summaries.
• The distinction he draws against downloading a transcript into a general chatbot: "I don't want to reach in the mind of the entire LLM. I want to reach into the mind of the LLM's extraction of this specific document."
• The host adds his own trick: listening to the earnings audio, because "the voice itself tells a story."

58:49 Every number traced back to the filing — and the $40,000,000,000 that looked fake

• "Every single number we trace back in the cell to where it came from in the filing. And this isn't just for US companies. We have it all for Canada, UK, and Europe too."
• Why it is the trust mechanism, not a feature: "there's no such thing as a perfect data set in the world, but we're asymptotically have kind of the most institutional one now. And trust is built over me showing you instead of me telling you."
• The demo: Visa's September-25 quarter net revenue "almost looks fake. It is 40 billion. Not 40 billion 362. It is 40,000,000,000" — and the 10-Q confirms it. The host's framing: a junior analyst brings you a number that looks off, and now you can check it "without a trip to the other office."

1:02:30 Screen to exclude — GARP by style, and no junior mining

• "I love stock screeners. I always have… Not because, oh, here's a list of things that I should invest in, but here's a list of things that I know I might be interested in."
• The inversion: "it's really helpful to filter not what's in the screen, but filter out what's maybe not a good use of my time."
• His own style, stated: "somewhere in that growth at a reasonable price type area" — no steady topline growth history and, special situations and spinouts aside, "I'm very very unlikely… going to be a long-term shareholder."
• The one sector he names and removes: "I don't like junior mining. I think it's just so boom bust, hard to make money. It kind of feels more like a lottery ticket unless you're in the industry. I want to exclude those industries. So I can go in here and exclude metals and mining and now rerun the screener."

1:04:59 Cross-comparison — the travel aggregators, and the answer he didn't expect

• Saved charts he keeps: trading platforms, the payments one (transaction volumes across Visa, Mastercard, American Express), and "travel aggregators" — gross bookings across Expedia, Booking Holdings and Airbnb, "off which base, off which take rate."
• Airbnb is "starting to look pretty interesting for me… I've been doing some more research" — but the chart contradicted him: "what I actually realized is Booking's growing the fastest off the COVID lows. That is not a thesis that I would have come in with."
• The general use: check a company's own account of its industry against its peers' disclosed numbers. The host's version — if one company says it can't push pricing while three peers are growing volumes and margins, "maybe the company you're paying attention to is facing some other challenges."

1:08:16 Dashboard, watchlists, notifications — and what's coming

• The dashboard holds watchlists with user-chosen columns (he sorts by forward EV/EBIT), an ownership breakdown that aggregates your positions "if it was a specific business", and a notifications panel — Taiwan Semiconductor's Q4 posting is the live example.
• The roadmap is personalisation plus geography: today "there's this non-symmetrical experience if you're looking at a US company versus some micro cap in Taiwan", and the engineering work is to make it uniform. "We don't want to just be focusing only on the US, Canada, UK."

1:11:03 Success — 50 employees, all of them shareholders

• On the moving goalpost: "I'm very rarely looking out the rearview mirror and always forward… I wish I was better at this honestly, but it's just not who I am."
• The ambition is not modest: "I also want to build a really sustainable heavy business that is the next really big powerful business inside of financial data."
• "Every single full-time employee has some equity in the business too… what success really looks like for me is that all of those people get really good outcomes and not just me." The host closes the loop back to the asset-ownership mission he opened with.

1:14:45 The free plan now gives 10 years of history

• "We just recently made the free plan — so throw away the trials… we now give everyone 10 years of data. It used to be just four or five."
• The claim attached to it: "of all the ones that exist there, no one is offering 10 years of historicals. There is not one of them doing it."