Betting on things that never change (2017)
collabfund.com
collabfund.com
Technical founders often focus on a particular cool way of solving a problem and burn lots of capital building that thing. Sometimes the thing is the right thing and everyone makes money. But sometimes it’s not. Yet if you stick with the same problem space for long enough, and aren’t too connected to your particular solution, I think you have a greater chance of succeeding.
Okay, go ahead and poke at my pontifications now. I’m ready.
Is a microprocessor or a mobile phone one thing in many different problem spaces?
If you build microprocessors, you better understand how your clients are using them to create electronic devices, and provide all the services and documentation they may need. A less potent microprocessor paired with a website that perfectly explains how to use it and what use cases it supports will win any day over a more powerful chip that no one understands.
Most of the startup opportunities in the last couple decades were built off the back of several deeper tech breakthroughs like mass mobile phone adoption and virtualization. I think more founders have been successful just riding these tech waves than sticking to a problem space. Brian Chesky didn’t work in hotel management and Travis Kalanack didn’t work in transportation. They were just familiar with cutting edge consumer tech and applied it to problems they had no familiar with before they started.
Tesla and many other EV companies started with low-volume expensive sports cars, with high margins so they could be basically hand-crafted. (Tesla actually bought the chassis/body/interior from Lotus) If they had started with the Model 3, they'd have gone bankrupt long before they sold any.
However, the tax discount drove ALL the small, local technical booksellers out of business. It might have happened anyway since local sellers had to pay retail rent and so Amazon had a structural advantage. However, Amazon also gained a straight 8% advantage for not having to pay tax.
I'm assuming you are saying "not paying sales tax". Amazon was operating like every other mail order bookseller that did not operate stores at the time. Retailers have complained for a century about mail order catalogs not collecting local sales tax. This was not new, or novel.
> the tax discount drove ALL the small, local technical booksellers out of business
Demand aggregation is a real thing. Localized, there's not a lot of demand for any specific niche technical book, and that leads to sub-optimal inventory. Why carry something you'll only turn once a year? You'd be better off putting that floor-plan money into gum, mints or batteries at the cash register. Quick access to product is why I shifted to buying technical documentation online. The local electronics store or bookshop could order books and it would often take 4-6 weeks for the book to arrive. Even early Amazon did 3-4 days. Demand aggregation is how Amazon won.
That was never true - sure they didn't have high street stores, but they invested big time in warehousing. That's why you could get a book from Amazon in days, whereas a traditional seller it would take weeks if it needed to be ordered.
They also invested in the wider distribution network.
They replied that their business exists to solve a need -- a need that exists regardless of the specific regulations. Should the regulations shift, a company motivated beyond a specific solution will adapt to keep meeting its customers' needs.
Where to start (concretely) and what you’re actually going to do, and if it will work for customers in a way they will pay you more for it than it costs to do.
Discovering that, and if it is going to work, is what a startup does.
You can’t do it without deciding on something concrete eventually to deliver, and pursuing it with enough vigor to get a workable product in someone’s hands.
Tech startups that build an entire product without understanding the space are leaning too heavily in one direction.
Sales led startups that get a customer base and start collecting money without a concrete product plan are leaning too hard in the other (and even more, add big red flags for scams and fraud).
Doing all the things and integrating it into a workable collection of customer needs they’ll pay to solve, along with an actual solution that works to solve it, while not going bankrupt in the process is why startups are hard.
Not being a scammer or committing fraud in the process? Harder still!
It’s the unknown unknowns and navigating all the bullshit on the way that also make the risk very high. Fundamentally, it’s impossible to know everything until you’ve already done it, and by then, it’s already changed under you.
Today, most estimates I could find put AWS at 25%-50% of Amazon's worth.
This nicely summarizes the current hype in Bitcoin or deep learning.
But I'm not convinced that this advice can be used to pick your fundamentals, your core business model.
Amazon opted for more choices and lower prices.
Barnes & Noble stuck with faster solutions to problems (get your book today not after a week of shipping as was common when Amazon started), deeper human interactions (you can go into the store and ask someone for recommendations, meet authors, etc.), and increased confidence/trust (you know what you're getting because you can hold it in your hand and read it before you buy it).
Low-cost carriers have moved in on traditional airlines because while it's true that people will never stop caring about added comfort, it turns out they care about lower prices much more.
But Apple built a staggering market cap relying on the assumption that while people care about lower price, they care about great control of your time (just works) and higher social status more.
So it seems like this article provides a good framework for thinking about your business, but doesn't give any answers as to the actual strategy you should use.
Or am I missing something?
I read TFA’s take on fundamentals as neutral on the examples given. There was mention that Amazon survived where Beenz did not because they picked the right mix of changing and unchanging. On the other hand, the juxtaposition of Andreeson and Buffett says that you can find business success with seemingly opposite fundamentals.
The vision is you realize the universal truth (in Amazon’s case people want things cheap and fast).
The strategy is how you deliver on making that visions reality. This is what you call (rightly so) your core business model.
Only experience people tend to get all caught up in the big ideas of the vision (especially middle managers) but spend far to little time on making the hard choices that come with deciding in strategy: it’s as much, if not more, about deciding what not to do as it is commitment to what you will do to realize your vision.
Actually there is one thing people are starting to desire -- less choice. I think some kind of curated content for anything has a lot of potential because mostly I do not care if I get the 100% perfect thing, I care about it being 80% right and not a headache. The time investment is not really worth the trouble to find the perfect thing.
Crap has a place sometimes. But I want the same things to be grouped.
IE I don't need one knockoff product with 30 different brand spellings of random letters taking up 30 slots of space in my viewport.
Show me:
- that option, price range, ability to dig deeper into that if I want.
- 29 other things.
Perhaps I'd understand what you meant with an example? What products are you finding is too varied?
Some brands build a lot of reputation on quality, and then cash it in for higher profits by suddenly switching to low quality (American appliance brands were notorious for this, with a sort of round robin selection on which brand had actual quality behind it for any given year).
Agreed with this, which is essentially why Amazon's e-commerce is probably doomed. Reviews have become unreliable and brands lack consistency on their site (is it fake? is this an officially branded product? who is actually selling me this?)
This is why I've got my bet on Shopify.
> What products are you finding is too varied?
I would argue virtually every category suffers from too much choice: https://en.wikipedia.org/wiki/The_Paradox_of_Choice
The paradox of choice would be mitigated by better trust, and better user interface. These aren't necessarily easy problems to solve, but Amazon should have the resources to solve them, if it only had the will.
Given the externalities involved in Amazon style delivery, I personally wish they would optimize more for throughput than latency.
Maybe not less but easier choices. It's hard to tell which frying pan is the best when there's 100 nearly identical listings all with 5000+ 5 star ratings which may or may not have been gamed. There's not a great way to tell the objective quality of a product before you buy it.
Given that today’s sellers make it nearly impossible for a consumer to compare items (store exclusive SKUs, component swaps in same SKUs, astroturfed reviews, and even inventory co-mingling), a curated marketplace is worth so much.
- Faster solutions to problems ("buy the best widget for me" button > two hours of amazon + google + reddit research)
- Increased confidence/trust (trusting that "buy the best widget for me" really is the best widget for me, not the one with the highest affiliate-link returns)
Also Bezos (paraphrased): "While working at D.E. Shaw, I heard that the internet was growing at 2,300% per year. And I asked myself: 'what kind of business makes sense to start in the context of that kind of growth?'"
The Holy Grail is when a big change allows you to get a foothold in solving a fundamental, unchanging problem better than it has been solved before - and then leveraging that foothold with rapid iteration and growth.
For every ten IT trends/fads, roughly only one has staying power in my observation. The others often move into niches or fade.
If anyone was that good at predicting the diamonds from the duds, they'd be golfing with Warren Buffett instead of poking around Hacker News.
Note that I thought the internet had big potential in the late 90's, as I was getting hooked on Compuserve (a precursor to webbish things). I just didn't think it would ruin biz CRUD/GUI's, rather staying in the consumer lane.
Yes, a loose analogy can be made in that there was a hyped technology that some people thought might become huge, others thought might merely become large or medium sized, and most others were pretty unaware or didn't care.
Unlike with crypto, during the late 90s internet/PC/web usage explosion, though, there were decades of successful smaller scale use (BBSes, academic use of the internet, etc). It was just a matter of scaling it up. The future was already here, just unevenly distributed.
Unlike crypto, a lot of the usages were not just emotionally exciting, but actually practical. You could look up a stock quote right away instead of looking at an outdated quote in the paper, or calling a broker on the phone, or having a dedicated fancy service just for that. You could email a friend half way around the world. You could read the full text of a book for free, if it was from before 1923. In general, you could transmit information freely and cheaply across the world -- as long as both recipient and sender were on the internet somehow.
All of those things previously were not possible that easily. There were faxes and landline phones already, but they had their obvious limitations.
Crypto is totally different. It's been hyped for two decades and, except for illegal activities, still hasn't enabled anything that could not be done more efficiently with SQL database to track who owns what, and a system of courts and law enforcement to enforce property rights and enforce contracts.
If you think they are similar you probably don't remember just how different it was to not have a single simple easy way to get and receive information. You had to actually go specific places or talk to specific people. You would do stuff like check movie times in the newspaper, and make specific plans to meet people at specific times.
I still remember when Tell Me came out. It was a phone number interface to some basic internet applications like weather, news, sports, movie times. You called 1-800-TELL-ME and followed the prompts. It was massively popular.
Where is crypto's TellMe moment? Nowhere. The only useful thing I've ever been able to do with crypto is accidentally profit $8000 because I forgot I had owned some I had bought as a joke... which came directly out of someone else's pocket later when it crashed.
If you're looking for a historical analogy, tulipmania is a good one. So is the boom in joint stock companies in the 1600s. So is the plank road boom of the late 1800s.
Nobody had as strong opinions about the internet being a horrifying scam, like many do about crypto. Because it wasn't, and there was no reason to think it was. The main "issue" was, would it catch on? Was it too nerdy? I mean, it's not like now, where the internet is almost slightly cool to some young people. Buying a modem was extremely uncool. And it cost a lot of money. So it was more of a question of "Is it really ready for prime time yet or does it need to become easier to use and cheaper?" much like EVs today. And, again, like EVs today, the problem was solved gradually but exponentially... each succeessive cost decrease or simplicity increase lead to another wave of converts, which made the whole thing more useful, solving the chicken and egg problem gradually
"Shuffling, you were dealt the, five of cloversh, and the, shix of hearts. Say yesh for another card, or no to stay!"
I was a child then and I used to dial TellMe any time I encountered a public payphone, just to see it work.
I'm skeptical of crypto as well...but when reading your post about the internet it actually reminded me more of crypto than not.
Although the internet looks revolutionary in retrospect, it may have felt incremental at the time -- because the telephone already existed. Sending an email vs. making a phone call. Driving 30 mins to the library vs. using an internet browser.
What about sending money to a friend in Germany today? Sure, it can be done. But it's going to take several days, involve currency conversions, fees, and likely some phone calls. With crypto they just need to message you an Ethereum address and it's done.
What about sending $20k from one US bank account to another? It still takes 3 days. Isn't the crypto speedup as significant as emailing a friend vs. making a phone call?
You make a fair point in theory, but this kind of superficial thinking is what crypto shills count on (no offense, I see this kind of argument a lot).
Here's the difference: With crypto, that speed comes at a huge and disturbing cost-- the inability to reverse a fraudulent transaction and appeal to the court system to fix other misunderstandings. Sorry, but it turns out that's essential to everyday business, not "silly" friction. There have been many cases demonstrating this by now, not to mention the huge enablement of ransomware, etc.
Crypto is never going to be able to solve that problem because if they do, they stop being crypto, by definition. The friction exists for legal and regulatory reasons. If the legal regime wants to make instant transfers possible, well they can, which is why you can now Venmo or PayPal people instantly. With the right treaties and such, it would be possible internationally too. The technology is there, but the political will isn't.
Crypto gets around needing that, which seems nice, but the cost is too great for anyone but hobbyists, criminals, and speculators. It's no different from a car company being able to undercut all other manufacturers by not including seat belts, air bags, or brakes. Cool proof of concept, but of limited utility except for niche applications like NASCAR.
If I actually need to send money to my friend in Germany, I will use Western Union (or whatever the modern cheaper quicker equivalent is -- a lot of movement in this space lately!). It will be same day, it will cost a small fee, and the small fee will probably be worth it to avoid the risk of scam, theft, and crypto volatility.
Idk about that. Some of the more mass appeal stuff of those days (you can look up stock quotes and movie times, etc blah blah) like I mentioned, was definitely more incremental.
But having access to such huge volumes of information that could not necessarily be found anywhere else was a big deal. You could download a long text file explaining how to hack the phone system. That would not be in any library or bookstore, at any price. Not to mention communities of people. You could discuss things that were very niche, something people take for granted today.
Crypto hasn't done anything qualitatively different like that. It just lets you send money instantly, and with no recourse if you made a mistake or got tricked or your counterparty fails to perform and isn't in your jurisdiction.
PS: Sorry for the double reply, not sure if that's a faux pas, but they were about very different aspects
The “what” is often timeless and stays the same.
The “how” the what is achieved changes.
I need to get from point a to point b (the what)
V1: walk/run/horse
V2: carriage
V3: train
V4: cars
Etc etc
There is another dimension to this as well.
Sometimes the what has to completely be re-architected since society has fundamentally changed.
As an example:
Consider education (what) this will always be required in society.
A naive version of innovation in this space would be to keep things the same and just change the delivery mechanism (online)
A better innovation would be to re-imagine what it means to be educated/schooled.
The best place to start is identifying the assumptions inherent in the current system. , and asking are they still relevant?
Long-term compounding of wealth indeed comes from betting on things that won't change over the course of a lifetime.
> which is that an insurance company selling directly would have a cost and convenience advantage over those paying brokers.
Put it into software, to me it is about eliminating abstraction/boundary/communication. Frontend and backend is an invented boundary. Product and Dev is an invented boundary, CS/QA/Documentation and dev is also an invented boundary. Whoever break the most of the imaginary boundaries is going to make software building so much faster
Do boring things well.
It's easy to chase the newest technology, the newest features, the newest markets. But unless you pay attention to the boring things too, you've set your feet on sand instead of stone.
Interestingly that is kind of what people have started saying. Perhaps not in those exact words, but in the form of “Jeff, don’t outcompete all the Mom & Pops” and “Jeff, treat your workers better”.
1) the rune goldberg machine of suffering of amazon laborers
and, more selfishly:
2) everything is so cheap, that it's useless. I would rather have the choice between two or three premium options rather than having thousands of cheap knock offs.
I would be surprised if >10% of the US would prefer to pay Amazon more so that others could be treated better.
Amazon is doing exactly what companies in a capitalist society are geared to do.
I'd grant that Amazon has a large selection and that's an advantage over competitors.
As for pricing, I wouldn't say Amazon is expensive but it is not consistently low cost and it is not unusual to see something on the shelves at Target for $50 that is selling for $70 on AMZN. I think after all these years, and particularly with Prime, AMZN has gotten people in the habit of looking to AMZN first and you'd expect them to cash in on it.
The better selection than other retailers is a durable advantage. I'd call out Best Buy for having a poor selection in comparison, particularly being overly wide (they sell washing machines, grilles, and mirrorless cameras and parts to build PCs) but not sufficiently deep (no selection of lenses for those cameras, no quality tripods, a very limited supply of PC parts.) It begs the question of "Why am I going to drive to Best Buy where they might have some of the things I need when I can order them all online from Amazon or a specialist retailer?"
The flip side is that the marketplace puts a lot of junk products in the AMZN catalog. I'd never buy anything that is likely to be counterfeit on AMZN (Nike only from the Nike store) and I learned the hard way not to buy a thermal printer that is marketed with a size in mm instead of inches.
Andy Jassy's 2020 re:Invent opening keynote goes a bit deeper into these ideas, which makes for a fascinating watch: https://youtube.com/watch?v=xZ3k7Fd6_eU
That said, reading his early interviews it hard not be impressed by the foresight and insight. I personally do think he got some things wrong ("if you can't measure it, you can't manage it" mantra being one of them ) or at least got others to misunderstand it, but it is hard to deny the results he helped manifest.
It is the same impression I have of Thiel.
Could you explain this point more. I'm genuinely curious.
<< "if you can't measure it, you can't manage it"
1. It is wrong in a literal sense. You obviously can. Maybe you should not ( which is a valid question to ask ), but you absolutely can manage things and not measure everything.
2. It is wrong when taken metaphorically to a logical conclusion. The mantra seems to encourage ( and did encourage based on some of the things managers tried to measure -- lines of code come to mind ) measuring everything. In accounting, there is a concept similar law's 'de minimis' ( I kinda don't want to use that name, because I worry people wills start questioning that from tax perspective, which is its own animal I don't want to touch, but in practical terms, the name matches the concept exactly - basically 'too small to matter'), where accounting for some items in some processes is borderline pointless ( say a screw ). Every so often, people in accounting classes go 'but what if', 'but we can track it'. And the answer is invariably: 'You can, but is it worth it?'
Naturally, you could counter it with: 'Well, you don't know if you don't measure it.', but there has to be a point at which the tracking of atoms get too crazy.
So that is my basic thought on the matter. I can expand a little more, but I think that covers it. Note, I am not saying it is not worth it to measure stuff. I am saying it is important to be discriminating over what you pick to measure, because, people being people, will optimize for it.
One of the teams I manage is technical support (B2B SaaS). Support's core functionality in an organization is to prevent churn. If a customer reaches out to support, they have some issue that is blocking them from using your product effectively. If you are unable to unblock them, then you risk them switching to a competitor. The problem is that not every support ticket/interaction will result in churn so we can't really directly measure how effective we are at our core function. So instead we measure the number of tickets raised, customer satisfaction, adherence to SLAs, etc, which is as good of an approximation as we can get with the resources we have available to us.
In my case, those numbers all look reasonably good, so I'm treated as an effective manager, even though we don't (and maybe can't) really know how much churn we've prevented. I wish we could though - it would make my arguments to executives on behalf of the support team that much easier to say, hey we retained $XXXXXX in revenue over this period.
Toy stores (Toys”R”Us) used to be a thing. They were legally on the hook, if they’d sell contaminated toys. Now no one is. And unfiltered toys from China are shipped directly.
I did sell it in the past without all these issues. So things do change in terms of safety, and likely because there is at least some liability on Amazon
I've seen this when HN shifted from YC applicants, ie startup founders, to tech workers in general. There is quite a paucity of startup news that's not necessarily related to the wider tech ecosystem.
It wasn't esthetic, but got most CRUD jobs done quick and simple. I've never coded in it myself for production, but observed amazing productivity at multiple orgs. There's something magic about it.
Oracle originally had to cater to many OS's, so were parsimonious with GUI features. Thus, it has just enough to do typical CRUD jobs but not enough to distract, confuse, and over-complicate the tool. They didn't get feature-happy over time. Once you learned how to milk existing features, you realized you really didn't need boatloads of pluggins.
Current web stacks are bloated nightmares and business money sinks in comparison. I'm not saying bring back Oracle Forms as-is, but at least borrow its best lessons. We'd have more practical GUI web standards by now. (Most CRUD productivity don't need mobile UI's. Also, it may not do well for big "enterprise" apps, but for small and niche it groves.)
I wish more VCs thought about investing in our sustainable future on earth, the way Morgan Housel thinks about investing in sustainable companies.
The only assumption that keeps us from that step is that all people care about is their short-term rational self-interest.
of course, they do! what else would they care about? How can you even imagine a Market without that assumption?
Yeah... let's just continue to bet. I personally bet there are some "timeless" things that will definitely change soon.
Except that many consumers these days do look for business that are more socially responsible and pay their fair share downstream.
Personally I am seeking out business that pay good wages and are responsible to the environment. Obviously that'll lead to higher prices so in some way I am wishing for higher prices.
They both have the exact same approach, which is to seek companies that have or will create monopoly effects in their markets, using dominance of a market segment to raise prices and extract rents.
Same investment strategy that’s made basically every vast fortune since we invented the cotton gin.
Megalomaniacal monopolists such as these two seek to extract monopoly rents by cornering markets.
Their differences with each other are trivial, and the contrast between them and nearly everyone else in the normal business world is stark.
They are robber barons, it’s not some new and exciting story, it’s a tiny class of humans, there’s only a few dozen at any given time.
> the term was typically applied to businessmen who purportedly used exploitative practices to amass their wealth.[2] These practices included exerting control over natural resources, influencing high levels of government, paying subsistence wages, squashing competition by acquiring their competitors to create monopolies and raise prices, and schemes to sell stock at inflated prices to unsuspecting investors.
Sound familiar?
She was an entrepreneur of 1950’s that started a small retail business.
There may be ethical concerns.
- NINJA loans, swaps, CDOs, etc.
- housing markets collapses