How to Identify Interesting Markets
What this covers
Elad is the only speaker, and he works through how an investor can find markets that others overlook. He starts with a three-part Non-Obvious Market Typology. The first type is a new market created by a technology shift, which is the rarest. The second is a market that looks crowded but is still early because the products are immature, as with search before Google or cloud storage before Dropbox and Box. The third is a market opened by a distribution or segmentation shift, such as Zynga riding Facebook, SaaS entering at the low end, or Tesla entering at the high end. He adds practical signals: 20 to 30 percent monthly word-of-mouth growth, and big brands adopting a product on their own. He also argues that a positive founder reference check means something while a negative one does not.
The middle of the episode is about how companies scale. Using Microsoft, Google, and Facebook, he describes a Product-to-Distribution Playbook in which one product's channel is used to launch or acquire the next. He also pushes against received wisdom. He calls the long tail argument largely false, since a concentrated group of power users often carries the economics. He says Silicon Valley overlearned winner-take-all, when many markets settle into oligopolies of around three. He also covers self-inflicted wounds, and he criticizes holacracy and Valve-style flat organizations.
He then moves outside software to two under-served areas. One is longevity biology, where he draws on worm genetics, rapamycin, metformin, and parabiosis to argue that aging is a perturbable program, and he points out that pharma has largely ignored it apart from a Novartis trial. The other is custom ASICs for machine learning, which he forecasts could support a company worth tens of billions.
Elad argues that early-stage investors should weight market and real user pull above team, that successful companies usually scale through one distribution channel before adding products, and that growth can mask many operating mistakes.
- Early signals such as organic growth, unprompted enterprise adoption, and usage despite rough products predict success better than conventional metrics
- Companies win by finding one distribution channel, then broadening into new products and M&A
- Non-obvious markets are new technology shifts, crowded markets that are still early, or openings in distribution or segmentation
Distribution, not any single product, must drive growth, and a focused distribution strategy beats build-it-and-they-will-come.
- Google once took its own homepage down for a day to serve traffic from a new partner, Yahoo or AOL, which had far more traffic than Google, illustrating how aggressively it pursued partnership distribution.
“Google actually took down its own homepage for a day so that they could serve that new traffic from that partner”
- The right distribution hires depend on the channel: enterprise sales needs different leaders than growth marketing, and growth-driven companies need quantitative people who optimize small details like email conversion.
“depending on the type of distribution that's working for the company, you're going to need pretty different skillsets”
- To build a lasting franchise, a company must shift from thinking of itself as product-centric to also being distribution-centric, because scaling one product alone tends to cap the business.
“you need to shift from thinking of yourself as a product-centric company, into a company that's also distribution-centric”
- Companies usually try three or four distribution methods, but a company typically grows off only one channel that works very well, such as Facebook's viral person-to-person and email-scraping growth or Google's early word of mouth that quickly moved to partnerships.
“in general, a company will only grow off one distribution channel that works really, really well”
- Companies that reach around five billion dollars instead of fifty billion usually fail to be aggressive enough on distribution, M&A to diversify, and developing new product lines.
“the companies that tend to sort of end up as a five billion-dollar company, which is amazing, but they could have been a $50 billion company”
- Microsoft, Google, and Facebook each built a dominant product, used its distribution to launch or acquire new products such as Office components, Chrome, Android, Instagram, and WhatsApp, and this pattern separates companies that keep growing from those that tap out.
“start with one great product that's working in a singular way and then add more”
- Microsoft crushed competitors in the 1990s by bundling products with its operating system, which led to antitrust hearings, showing that dramatically better distribution can overwhelm rivals.
“Microsoft effectively did that in the '90s, which is why they had all the antitrust hearings - they would literally just crush their competitors through distributing things as a bundle with their OS”
- Strong distribution professionals share a common trait of being aggressive but ethical, which means a right type of greed aimed at driving product success.
“they're aggressive but ethical. And so there's sort of the right type of greedy”
- Facebook's early distribution was email scraping and viral person-to-person spread, while Google's early growth came from word of mouth before it rapidly switched to paid and negotiated partnerships.
“In Facebook's case, early on, it was basically email scraping and really focusing on sort of viral person to person distribution.”
- The long-tail argument that most internet businesses rely on niche customers is largely false, since even Google and Apple depend on concentrated user groups, and mass-market retail like Apple is the closest exception.
“most internet businesses are really based on the long tail. And in reality, that's really false.”
- Distribution is underrated in business lore, and companies that focus on a singular, well-chosen distribution strategy outperform those relying on a build-it-and-they-will-come approach.
“you need to shift from thinking of yourself as a product-centric company, into a company that's also distribution-centric”
Internal breakdowns like poor hiring, infighting, and flat structures cause companies to fail until they professionalize.
- Holacracy and Valve-style flat organizations are flawed and tend not to work, since removing hierarchy without another coordination mechanism produces poor results.
“the holocracies, and all the various - you know, the valve style approach - like all these things I think are actually pretty flawed in terms of their approaches, and tend to not work”
- Some companies with strong product-market fit succeed despite poor org structure, and are then cited as proof of alternative approaches, though many could have been ten times more successful with proper organization.
“some companies could have been 10 times more successful if they'd actually organized properly, but they didn't”
- Human nature has not fundamentally changed over ten thousand years, so organizations still need direction, coordination, and processes to deal with people who act badly, even as innovation continues.
“when all is said and done, our basic drivers are the same”
- Fast-growing companies can survive every kind of operating mistake, and growth masks these errors until the company professionalizes; a Cisco manager's maxim was that growth covers up a lot of mistakes.
“growth covers up for a lot of mistakes”
- Beyond strategy, hiring, and capital allocation, a CEO is ultimately the company's chief psychologist and must spend substantial time on unexpected people issues.
“you're also ultimately sort of the chief psychologist of the company”
- Elad credits his career to many small acts of kindness in Silicon Valley, where people introduced him to companies while he had no money, reflecting an ethos of taking a shot on smart people.
“I've benefited from a 1,000 small acts of kindness that have sort of built up over time”
- First-time founders often burn out as companies scale because they keep doing tasks they hate, such as sales compensation or HR, unless they build a team to handle them, and the inability to delegate is the second breaking point.
“one big sort of breaking point is just doing stuff that you hate”
- Most companies die or slow down from self-inflicted wounds such as choosing the wrong people, leaving them to make repeated mistakes, founder or leadership infighting, and toxic culture, which are hard to fix.
“most companies tend to die of self-inflicted wounds or to really slow down of self-inflicted wounds through bad management, bad people, choices and bad culture”
- Companies tap out when they have one strong product but never add a second product line, fail to reinvent, or iterate too slowly, letting new entrants or incumbents slow their market.
“the companies that tend to tap out are the ones that have one product that works really well, but then they never add that second product line”
- Organizations need direction and coordination, because without roles, a group making dinner can end up with five desserts and nothing else, so removing hierarchy entirely produces lots of enthusiasm but poor execution.
“Otherwise, you end up with five desserts.”
- Metrics such as negative churn are useful signals of traction, alongside talking to customers, though Gil considers them standard criteria rather than novel ones.
“looking at metrics like negative churn”
New technology, channel shifts, and waves of incumbent dissection reveal non-obvious markets and company-building opportunities.
- The rarest non-obvious market is a brand new one that grows unexpectedly fast after a technology shift, such as cryptocurrencies or the mid-1990s browser wave that let software run inside browsers.
“it's a brand new market, but it's growing really fast and it kind of comes out of nowhere and surprises people”
- A third type of non-obvious market opens when a new sales, distribution, or channel shift appears, such as Zynga riding Facebook, SaaS entering at the low end, Tesla entering at the high end, or steel mini-mills melting scrap.
“there is an opening from a sales or distribution or channel perspective”
- A market that looks crowded and game over can actually be in its early days if products or infrastructure are immature, since prior entrants identified real user needs but built janky products, as with search before Google and cloud storage before Dropbox and Box.
“what had happened is all these other companies that were kind of janky had identified a real user need, but they just hadn't built a product that addressed that need properly”
- Non-obvious markets come in three types: brand-new fast-growing markets created by a technology shift (the rarest), crowded-looking markets still early because the product is not ready, and markets opened by a distribution, channel, or segmentation advantage.
“it's a brand new market, but it's growing really fast and it kind of comes out of nowhere and surprises people”
- Taking a Fortune 500 company apart and identifying the functions it repeatedly performs can reveal about a dozen valuable companies, as Checkr did by upgrading background checks into an API.
“if you were to take a Fortune 500 company and just take it apart and ask what are the pieces that you have to do over and over and over again, you'd end up with a dozen really valuable, useful companies”
- Each technology wave creates a massive semiconductor company, such as Qualcomm and ARM for mobile, Broadcom for networking, and Intel and AMD for desktops and servers, and Bitmain is the current equivalent for crypto ASICs.
“each technology wave also creates a massive Silicon company”
- Custom ASICs that replace GPUs for machine learning, which are more power- and performance-efficient, form a large under-attended market that could support a company worth tens of billions of dollars.
“the semiconductor and systems layer for machine learning is a very large market that a lot of people aren't focused on”
- The internet, mobile, and cloud together have created markets far larger than anyone anticipated, with global reach at low friction, which explains why companies now reach scale faster than ever and some approach trillion-dollar valuations.
“the internet has created markets that are larger than anybody in the world ever anticipated”
Early organic demand and real product usage are the strongest signals of a company's potential.
- Early-stage growth of 20 to 30 percent a month driven purely by word of mouth is a strong sign of real demand, even when absolute numbers are tiny, and investors often discount it wrongly.
“if something's growing 20% or 30% a month, organically just through word of mouth, usually there may actually be something there”
- When a product looks rough yet people keep using it, that persistence is a strong positive signal, as Snapchat's early usage despite a rough UI illustrated.
“if something looks really janky and you're like, 'How can anybody use this?' But people are still using it. That's usually a really good sign.”
- Companies that had launched a product, or at least a crappy demo, when they began raising money performed better in Elad's portfolio, so investing on a PowerPoint deck tends not to work.
“they launched a product or at least had a crappy demo when they started raising money”
- For enterprise or SaaS companies, having one or two major brands adopt the product after finding it randomly is usually a very good sign of product quality and real demand.
“If they have one or two major brands that are using them that just found them randomly, then that's usually a very good sign.”
- In many cases the biggest things work very early, since once a product is available and needed it starts moving quickly, though enterprise and highly regulated areas take longer to build.
“the very biggest things tend to work very early”
- In early days, a company should delight a small number of users rather than create a so-so experience for everyone, because it cannot reach mass market without a product a small group loves.
“in the early days, you absolutely want to delight a small number of users versus create a so-so experience for everybody”
- Roelof Botha of Sequoia says every successful product is fundamentally driven by one of the seven deadly sins, so the question is which sin the product lines up against to shape product thinking.
“every product that you can think of is fundamentally driven by one of the seven deadly sins”
Market quality comes before team, and founder reference checks and stakes shape outcomes.
- Elad says he evaluates early-stage companies by market first, team second, and whether he likes the founders third, which differs from most angels who put team first.
“the biggest difference between me and most angels or people who write these really early checks is that I really focus on the market first”
- Andy Rachleff's rule holds that a great team in a terrible market loses to the market, a terrible team in a great market also loses to the market, and a great team in a great market produces something special.
“if you have a great team and a terrible market, the market wins. If you have a terrible team and a great market, the market wins.”
- Positive reference checks on founders are a good signal, but negative ones are neutral, because a person who underperforms in one context may excel in another.
“negative reference checks on the founders aren't a negative signal, they're a neutral signal”
- A founder seen as lazy at Twitter became one of the most successful startup founders from that community after saying 'I finally feel like my ass is on the line,' showing that stakes change behavior.
“I finally feel like my ass is on the line”
- Aging appears to be a biological program that can be perturbed, supported by genetic knockouts in C. elegans that extend lifespan two to three times, and by drugs like rapamycin and metformin that extend lifespan in several organisms.
“longevity is just a program that you can perturb”
Market economics are driven by head power users and a few players, not long tails or winner-take-all.
- Computing LTV and CAC on the average customer can erase the signal about what is actually working, so the analysis should be done by customer segment.
“if you do it for the average versus by customer segment, you actually may lose the signal in terms of what's really working”
- Rumor has it that in Uber's early days, five or ten power users in a single city, ordering higher-margin black cars constantly, could cover the cost of launching a new office.
“if they had like a handful of what they called whales in the early days of the business in a single city”
- Most internet businesses are not really driven by the long tail; instead, a head or torso of power users often generates a large share of revenue or margin, and they may cover the entire business economics.
“most internet businesses are really based on the long tail. And in reality, that's really false.”
- Many markets naturally collapse into oligopolies of two to four or two to five players, often settling at three, so investors should analyze the real market structure rather than assume winner-take-all.
“many markets naturally will collapse into oligopoly markets of two to four players or two to five players. And often it ends up at three players”
- Silicon Valley wrongly assumed everything was winner-take-all after social networking and search, which led investors to avoid markets with a clear leader because they feared the number two would be worth only a tenth as much; this is sometimes true but not always.
“it's fooled a lot of investors to not ever invest in anything where there looks to be a number one”
Basic habits, enduring human drivers, and aging biology findings shape longevity outcomes.
- The most reliable longevity practices are basic: exercise regularly, eat well, sleep well, and maintain meaningful relationships, which have real impact on how long and how well a person lives.
“exercise regularly, eat well, sleep well, form meaningful relationships with others”
- Within the aging community, people commonly take a baby aspirin daily, a statin, and in a smaller subset metformin, while avoiding human growth hormone, which Elad believes is a poor longevity idea.
“I just don't think are very good ideas from a longevity perspective”
- People have not fundamentally changed over the last 10,000 years of evolution, so their basic drivers remain the same even as evolutionary pressures and cultural change alter particular traits.
“I don't think people have fundamentally changed over the last 10,000 plus years of evolution”
- Knocking out certain genes in the C. elegans Dauer or Daf pathway, which integrates insulin, aging, and cancer signals, lets worms live two to three times longer as healthy adults before crashing at the end.
“if you knock out one of these genes and these little worms, they'll live two, three times longer as healthy adults”
- In parabiosis, giving young blood to an older animal restores certain rejuvenative capabilities, while giving old blood to a young animal harms it, showing that aging-related factors flow in both directions.
“if you take young blood and you give it to an older animal, they'll have certain rejuvenative capabilities restored”
Aging is a real market gap that traditional biotech and pharma have largely ignored.
- Despite decades of longevity data, traditional biotech venture investors and pharma have not focused on building drugs against aging, leaving a real market gap, with only small exceptions such as a Novartis rapamycin trial in elderly adults.
“there just hasn't been that focus by either the traditional biotech venture community or by pharma to build drugs against this area”
- Laura Deming's accelerator Age1, which Gil advises, aims to fund the next generation of regenerative medicine companies at scale for anti-aging startups.
“an accelerator she's building for anti-aging startups called Age1”
- Gil seeded the longevity company BioAge, for which Andreessen Horowitz led the Series A, and helped early on with Spring Discovery, which applies machine learning to longevity assays.
“seeding one company called BioAge which Andreessen Horowitz did the Series A for”