Musings

Customer acquisition relies on conversion funnels. Of the X people you start with, X/10=Y click-through to a product. Of the Y who click-through, Z = Y/10 actually buy something. To get to 50 purchases, marketers may send out 100,000 emails, and sometimes 1 million emails. It is understood that the conversion from an unsolicited message sent out to a purchase is less than 1% in most cases. That is, the growth strategy for most traditional marketers is to assume that the vast majority of potential customers will ignore an email.

Why have people ignored an email? If you run an A/B test on a subject line, the content and the display style, you may find that some messages feel spammier, or more welcoming than others, and that your product itself actually has more “natural” demand than you thought. You may find, in fact, that they ignored the first email, but after the second, or the third, they succumb to the pressure of clicking through, and ultimately do what you hoped.

And then, of course, there is attrition. As any Mailchimp or Constant Contact user will tell you, every time you send out an email, you will lose a handful of people, who have unsubscribed from your service. This is the unsubscribe rate, and of those, there are buckets of “stop emailing me for a while” and “ stop emailing me ever”. It is understood in marketing that you will lose a few in the process of getting a few. These are the basics. Once you start adding complexity, the implications are interesting. Let’s look at LinkedIn as an example.

LinkedIn does a masterful job of growing, understanding that the game they are playing is a quantitative one. There is an email notification for every possible action that happens on LinkedIn, and a user has to manually unsubscribe from every one of them. And for those lists which you stay on (opt-out, not opt-in), they do a slow trickle of emails: every couple of days, optimized for the time of day given where you work, optimized for how many mutual connections you have with the actor which has precipitated the email. If you unsubscribe from one type of notification, they slow your trickle to once a week, instead of every day. They tune it algorithmically to respond to you without your realizing you need responding to, to ensure that they are extracting the best conversion out of you as possible without pushing you away.

They collect email addresses whenever a member OAuth’s one of their email accounts. During the flow of adding contacts, they invite users to pay the “stupid tax”, where they give you the option to add all of your friends to LinkedIn in what I might generously describe as a “less-than-intuitive” UI. A lot of my friends have accidentally added their entire address books this way. From LinkedIn’s perspective, this UX is perfect. It is exploitative enough that every person who does it feels legitimately violated, and likely reminds himself to avoid LinkedIn and warn others against it. But it is just rare enough (say, 1 in 100 of your contacts will do this) that the bitterness isn’t viral, so they can keep getting away with collecting email addresses this way. They have been fine-tuning this process for almost 10 years now, and I believe it is the special sauce that has made them the business they are today. They now have 100M users and a market cap of > $30B.

The best technical marketers are building responsive models that slowly create conditions like LinkedIn’s, and tune them to squeeze the most engagement and growth possible out of users. I know this because at Doostang we tried doing this, but ended up looking more like spam than like LinkedIn. It’s a quantitative science which we incorrectly treated like an art. If you want to employ this type of strategy, you are doing something closer to computer science than to brand-building. The question I have, given this, is this method worthwhile? Is it death to your brand by a thousand cuts? After all, I would guess that LinkedIn’s brand in my network is at best neutral, with a few outliers that are strongly negative (and a few outliers that are strongly positive… reasons for which I’ll save for a later post). Or is it just good applied behavioral science and statistics? After all, $30B market cap is nothing to sniff at. And if you take a quick look at your spam folder, the ‘conversion funnel’ strategy is big business and not going anywhere, is it? Despite having some scars, my instinct leans (sadly) to the latter: bury ‘em (quantitatively) with email, and you will grow and keep growing.

I was on a thread with my colleagues Craig, Alberto, and Ryan where Craig proposed that “social enterprise was eating the enterprise” and it got me thinking. Is that true? Why? I work for Collaborative Fund because I believe it, but what motivates that belief? Here is how I’m currently aiming.

It’s amazing to me to meet with entrepreneurs from outside the Silicon Valley. I’ve found that in New York, Los Angeles, Seattle, and even Boston, the valuations are lower, the entrepreneurs are more likely to be thinking about revenue, and the fundraising is proving (on average) more onerous. But really, it pales in comparison when meeting with an entrepreneur from outside the United States. A friend of mine joked that a Demo Day in Europe looks like a series of entrepreneurs each trying to outdo the next not on growth, or revenue growth, but on “time until we are break-even”, so as to assuage the frightened, risk-averse angel investors that their money is in good hands.

From my (very) limited experience, this seems to happen all around the world. In South Africa, which I know best, my friends and family in entrepreneurship have lamented that “venture capital” there means buying half of a profitable company, while in the Silicon Valley it looks more like buying 10% of a ‘growing’ company, often pre-revenue. This seems like a combination of structural and cultural differences.

In the United States, in particular in Boston and Silicon Valley, the industry is built on more than three decades of cycles, with booms, busts, exits, and plenty of vintages of venture capital data to attract big patient capital. In the newer markets like Seattle and New York, highly liquid anchor companies (Microsoft, Amazon), world-class engineering schools (UW, NYU, Columbia), and government investment accelerate the ecosystem, encouraging early stage investment, and thus entrepreneurs, with a different risk tolerance.

But there are highly liquid companies listed on the London Stock Exchange, or on the Tokyo Stock Exchange… Korea and regions of India have more world-class engineers per capita than most of the major cities in the United States… Are the investors in Kenya, where M-PESA, ushahidi, and i-hub are creating a very vibrant ecosystem, or in Brazil, where the government, huge populations of connected users are doing the same, and in Germany, where companies like Soundcloud are breaking through, ready to back super risky startups? Are they doing so already? Is there a culture of risk that is unique to the United States? These are questions I’m struggling with this morning.

Some industries tend to have winner-take-all dynamics more than others. These dynamics share the characteristics of “defensibility”. When a product is properly defensible, only a herculean effort or an enormous sum of money can shake it from its position of dominance, protecting it from start-ups and big companies alike. But not all defensible products have monopolistic tendencies, which is the “winner-take-all” dynamic I hope to explore a bit here. Let’s look at internet companies, in particular.

Content businesses are the furthest from winner-take-all. For a content business to succeed, it must produce unique, compelling content consistently, because page views are like snowflakes in the fall: they melt as soon as they land. The fastest growing content businesses, and the most successful, are content farms, often deploying thousands of free or unpaid writers to produce on their behalf. It doesn’t scale well.

Products are better, in this regard, because a well-designed product that captures magic in a user experience can lock a user into habitual behavior, and at scale that starts to look monopolistic, sure. But products are easily copied, and even if you have a strong brand, like Kleenex or Q-tip, you aren’t protected against competition. Technical innovation is a good to stay ahead of the market in this category, but innovation is fairly evenly distributed, and not zero sum. Better marketing is a great way to distinguish yourself, but its mad expensive, and an art not a science.

Social products, or networks, are even better! Here, the experience is not only in the product itself, but the fact that you are using this product with others. In these cases, your experience is not only tied to the design of the product, but also to the community that has gathered around the product. As networks become more dense, they become more interesting, and “network effects” describe the process whereby they become increasingly difficult to unseat. But Dunbar’s number is a real phenomenon, and some networks break at scale, thereby making them the best thus far, but still not the most defensible to the point of being monopolistic.

Enter Marketplaces. A marketplace is a network where you are transacting! On a marketplace, it is not only the value of the other people using it that adds to the experience, but the fact that the people may have things you need. Marketplaces start to look like winner-take-all when they reach liquidity. Reaching liquidity in a marketplace means reaching the point where an average customer finds what they want in the marketplace most of the time, *but not all of the time*, and yet still comes back. If they still come back, you’ve locked them in profoundly. It’s extremely hard to change their behavior once you’ve done that. I still shop a bit for on-demand ride-sharing, but it looks like 9 rides Lyft, 1 ride anything else. I still shop when looking for a car on carsharing platforms, but I’m doing so less as the industry is consolidating. AirBnb seems to be there. I don’t even go to hotels anymore, once I’ve decided I want to use AirBnb for travael. Thanks to experts like Simon Rothman of Greylock and Jeff Jordan of Andreessen Horowitz, every marketplace founder today understands that if they reach liquidity they win, but until they do, they lose. 

The debate seems to be settled that vertical marketplaces tend to work better than horizontal marketplaces, because they allow the platforms to focus on user experiences that are optimized for the specific type of product transacting, allow infrastructure to reduce the friction (in insurance, trust and safety, and payments), and because they focus customer demand, which is helpful for acquisition. But eBay is horizontal. Craigslist is horizontal. They are the most important marketplaces in the history of the web, with no signs of slowing. While it may be easier to get to liquidity as a marketplace by staying vertical, perhaps once you reach liquidity, you can be horizontal? Does that mean AirBnB becomes Air? Food for thought.

It is rampant, and extremely hard to avoid. I met with a number of entrepreneurs yesterday whose backgrounds were highly varied, and whose products were all inspired. I saw a couple kids who went to college together and had been tinkering on websites and launching multiple successful Kickstarter campaigns. I saw a twenty-something who had previously founded and been the principal of a Japanese high school where nobody wears shoes and Ambassadors send their kids. I listened to a pitch from a young man from Owensboro, KY who spent multiple years in high school homeless, and who had more grit, tenacity, and hustle than anyone I had ever seen. They caught me on one of those days where I had been reflecting on my own unconscious biases, and so I was paying particular attention to that. I want to take a moment to list a few of the ones I’ve heard, seen, or had.

Oh, you don’t drink elite coffee? You must not be a good designer.

Oh, you aren’t based in – or have strong ties to – a major NFL city? The press will ignore you.

Yikes, that watch is tacky. Your angel network must be small.

Wait, Latino engineer? Surely not. (I kid you not.)

That accent must make you a bad salesperson.

Your market can’t be that big, because nobody who looks and sounds like me and my friends is a part of it.

You dropped out of Stanford? You must be a genius.

You dropped out of [Insert non-USNews top school here]? You must be psychologically compromised.

These unconscious biases, fair or unfair, happen all the time, with all of us, and investors are no exception. And its concerning. Those entrepreneurs who raise early stage capital successfully do so because they navigate the silent social cues that we all look for when building relationships – particularly in the very early stage, when you can’t just point to growth as all the evidence you need. But some of these cues are impossible to navigate. The community is slowly recognizing the awful gender bias, and creating systems to counteract it. The community is working hard on creating infrastructure to counteract race bias (particularly for blacks), but there are many of these, and the most insidious ones exist between the lines of class and elitism, because they are harder (or more taboo) to describe. Just like culture can often be used as a proxy for “people who are like me”, taste can also be a way of discriminating. And I’m vocal in protest of the former, but acknowledging that I’m often guilty of the latter. It’s not good, and it’s not right. The last thing we want is an anonymous non-identity in our entrepreneurship class. The more diverse the problem solvers, the more kinds of problems that will be solved. So I’m working on being more conscious. Join me!

This morning I naively asked on Twitter: “Can someone write a post-mortem on HTML5 taking over mobile? Was it a technical performance failure, or a business strategy failure?” The collection of responses were almost unanimous in saying: it was performance. Users want their apps to be fast, and HTML5 apps weren’t performing well enough. Compiled language implementations (native code like Objective-C on the iPhone) perform better than interpreted language implementations (like HTML5 for mobile web).

Quick, very basic primer on what that means:
Compiled language sends directions directly to the machine. Interpreted language sends directions to another program, which then adjusts the machine code. The former allows you to optimize, runs faster (and in some cases much faster), and scales with fewer performance issues. The latter allows for faster, easier implementation, and since it has at least one extra layer of abstraction, can be used cross-platform much more easily.

It’s still early in HTML5’s life, and mobile interfaces are moving quickly enough that today’s status quo in devices, leading programming paradigms, and applications, won’t be tomorrows. But there’s a lesson in the above for other aspects of a start-up’s processes.

As a builder, if you can get a system up and running quickly and easily, there’s an apparent advantage, because you can test quickly, iterate, and stay lean. But it comes with a real cost. Those builders who optimize more carefully, build more fundamental systems, may end up with a user experience that is simpler, faster and easier, and that’s where the long-term advantage lies. Companies of all sizes struggle with this bargain, not only in their technical infrastructure, but across the user experiences that they design.

The answer to this challenge is not so simple as one or the other. My hypothesis: move quickly to learn, but then choose a point of view, and move slowly to manifest that point of view well.

Thanks Zach, Albert, Kane for your perspectives.

Every time someone designs a piece of hardware, she also needs to design a manufacturing line to produce that hardware. This is a non-trivial problem. The early lines of hardware products are always highly un-repeatable, because the designers are still experimenting with different mould injections, circuit designs, CAD software, not to mention the software itself. In this early stages of engineering, where they are doing skews of <100,000 units of their product, they have to pay a manufacturer to design a new line for every change that they make. And in the early stages, those changes are significant, therefore expensive. NRE, or non-recurring engineering, is a central part of every hardware product’s lifecycle. In talking about “things that don’t scale”, YC founder Paul Graham talks about NRE very briefly, referencing the “importance of a screw" It’s worth noting that the idea isn’t really new. 

Mark Leslie, founder of Veritas Software talks about the sales learning curve. At the beginning of an organization, or a new product cycle, the product-market takes a while to be defined, but even once it is, the sales cycle takes a while to really gel. While you may know what problem your product solves, and your customers may even understand it relatively well, or demonstrate a huge market for it, it takes a while to learn how you price them, what story you tell to sell to them, and the channels you need to employ to best access those customers. This explains why big splashy launches make so little sense. No engineer would in their right mind do a run of 1,000,000 units before significant smaller skews of NRE. All business processes have a learning curve, and at the beginning of the learning curve in any process, it is very clunky and very manual. I need to remind myself of that as I work on both product and sales for guild, and grateful to Paul Graham for the reminder.

If you haven’t already done so, read Peter Thiel’s article “End Of The Future”. It is darkly brilliant, and extremely compelling. It speaks to the fact that the energy revolution we were promised in the 70s seems to have stalled; that cars planes and trains don’t move any more quickly today than they did 40 years ago; that the leaps and bounds in biotechnology that we should have made by now seem to still elude us. Why, he asks? He presents a number of cases:

 – The libertarian case: government regulation has stifled innovation. The FDA, the TSA, the government subsidies on oil & gas, Congress, etc. have made it impossible for free market innovation to thrive today, the way it must have in times past. 

– The neoconservative case: the military industrial complex drove most paradigm-shifting innovation in the modern era, from gunpowder to the internet to advances in transportation, science, and energy. We’re too afraid of arms races to innovate anymore.

– The Krugman liberal case: Since Carter’s presidency was largely viewed a failure, Reaganomics, which brought along a deep fear in government spending, has starved the economy from investing in energy subsidies, and better education infrastructure, etc.

It’s a fascinating article, and one that I was really glad to have read before I took Thiel’s amazing class at Stanford last year. But I felt as though something was wrong with the argument. For one, as many others have described far more eloquently than me, Thiel includes the dawn of the Information Era as something of a concession, or a footnote, which is crazy. The invention of the Internet, the World Wide Web, and the mobile phone are all “printing press” scale innovations, and they happened within years of each other. And they permanently transformed every industry that they didn’t simply destroy. But that’s not my argument. After all, the conditions that catalyzed two, if not three of the Information Era’s core innovations were tied to military-driven speculative bubbles. The software folks have done an admirable job of debating that side themselves. I have a different beef.

Peter Thiel talks about “us” vs. “them”. His view on the technology industry speaks about how “we” have stopped innovating, and in the same breath it refers to “we” as it relates to job creation, and “we” as it relates to immigration policy. It’s clear that he’s referring narrowly to the United States, but perhaps more broadly to Western Europe as well. Yes, the United States is the world’s last great superpower. Yes, there are 300 million people of moderate to high income here to serve as an anchor customer base. Yes, the United States and Western Europe were responsible for the majority of technology in the 19th and 20th centuries. But you want to know the real reason why innovation stopped in the 1970s, as Thiel likes to say? The answer is Chairman Mao.**

If you look at Hans Rosling’s dynamic chart of the world’s mortality rates and per capita income, you see that at some point between 1955 and 1980 the biggest dot in the world starts moving extremely quickly. In Hans’ words: “Mao brought HEALTH to China.” And it’s true, he did. In 1959, the average life expectancy in China was less than 35 years. By 1969, life expectancy in China was around 65 years: a 186% improvement over a population set of a billion in 10 years. For people who study macroeconomics and public health, you don’t see numbers like that… anywhere. Mao Zedong’s reign is without question the most significant 20th century event that the West has forgotten to talk about. In the same 10-15 year period, the entire continent of Africa went from under colonial rule to independence. That meant effectively 2 billion people added to the world economy, ready to contribute, and especially to consume the world’s resources. Eric Schmidt points to the fact that there is a mobile phone in billions of hands around the world as evidence of innovation not slowing down. He’s completely right. There are more people with smartphones today than were alive in 1860. That is no small feat.

The companies, like Google, by creating the Android Platform and spreading the smartphone across the globe; like Apple, who created the most extraordinary manufacturing infrastructure of the modern era entirely in Asia; like Samsung, Huawei, Sina/Weibo, and dozens of other Asian companies, innovation stopped because there was a greater incentive, rightfully, to move the needle across the horizontal axis (one product to very many people) than along the vertical axis (one product, vastly improved). Both are needed, but sometimes the world shifts very rapidly very fast, and one overwhelms aspects of the other.

Entrepreneurs, investors, journalists, and thinkers: don’t sleep on the rest of the world, like Mr. Thiel seems to be doing. Talent, market opportunities, and access to resources are all increasingly distributed evenly around the world, and in many cases unevenly against the United States. Germany runs on clean energy. Kenya runs on an entirely different (and superior) payment infrastructure from the West. Gaming culture in South Korea would be unrecognizable to your average Zynga user, or Xbox gamer here. If you have interacted with technology entrepreneurs in other parts of the world, very quickly as you hear them describe their businesses, you will notice that they are intimately aware of where they are in the world: “the biggest X in Europe, and excited to launch in Asia” or “we are focused on East Africa, but there’s a big opportunity in India”  or “we are already international, but haven’t launched in the US yet.” If you listen to entrepreneurs in the United States, Mr. Thiel included, the language is very different. The nature, source, and pace of innovation in our world is changing whether we like it or not. Take heed!

** No, Senator McCarthy, I’m not saying Mao or authoritarian communism was good for the world, or even necessarily for China. It was terrifying and in many ways awful, to be sure. I’m just pointing out that we conveniently forget his level of influence.

I wrote last year that early stage capital increasingly becomes a commodity. Here is a quick summary of why:

– You can do MUCH more with MUCH less. There are more free resources in business strategy and legal, in hosting, and prototyping in web and mobile, and even visualizing full UX
– The top ranks of entrepreneurs have unionized, working in guilds like YCombinator and 500 Startups, and using these guilds to lead the market, create protections around deal terms, and force transparency across the industry.
– The angel market has exploded. Increased public market exposure, an arms race and a talent crunch driving an active M&A ecosystem, and AngelList have made it easier than ever for someone who’s made some cash to get in the game of funding startups.

In response to this trend, it is inspiring to see First Round Capital create a new source of free, highly relevant content for founders, driving this trend further. AngelList is iterating like a startup, finding more and more amazing ways of transparently connecting capital with companies. pg created the most powerful brand in the early stage ecosystem by inventing the MBA for startups. Dave McClure and 500 Startups are reaching their arms across the world discovering and developing talent. The list goes on.** At Collaborative Fund, we have done our part to push the envelope in as many of these categories as possible, supporting crowdfunding, leading our content efforts with founders first in Fast Company and our video series, and microsites for the community.

But I think there’s another reason why venture capital is wise to innovate as quickly and as experimentally as it can: the management fee. My friend Richard Arnold of Crowdflower estimated that management fees (including carry), account for up to 500 basis points of economic activity around the world. That is trillions of dollars every year going to money managers. And for what? Their value is predicated upon information asymmetry. As an institution or individual, I let someone else manage my money because I believe they have access to better information than I do, and that their experience and network make them more capable of profitable investments than me. The former is surely no longer true. AngelList, Pitchbook, Mattermark, Crunchbase, Twitter, etc. have suggested that the information asymptotes to perfect. So then it must be that a money manager’s experience and network make them more capable. I recently heard from the managing director of a foundation that they are interested in “getting closer to the network” and that their analysts often read the same reports as the venture associates and analysts these days. In a mature social web, the networks matter more and matter less. 

When Marc Andreessen famously wrote about Why Software Eating The World, I have to imagine he realized private equity, and more broadly money management, was not going to be immune to this trend. That is more evidence still for his decision to overinvest in his firm’s experience and network. If venture capitalists want to justify their 1.5-3% fees and 15-25% carry moving forward, they will all have to look more like these innovative firms, market leaders, or else they will get squeezed, leapfrogged, and ultimately die. 

** The list is actually long. In each of those categories there are fantastic firms innovating.

Thanks to Ash Fontana, Rich Kerby, and Tommy Leep for putting eyes on this.

Yesterday morning, for whatever reason, I decided to buy a hard copy of the New York Times while buying my morning coffee. I carried it around with me throughout the day, never opening it, but holding onto it, since I had spent good money on it. At the end of the day I was waiting for the MUNI to head back to my apartment. I remembered that I had my paper, and that it was Monday, meaning the crossword was easy. I opened it up and decided to Snapchat a pic of the empty puzzle to a few random contacts on Snapchat. By the time I got on the train, I had a few snaps from people who knew a few of the answers. By the time I got to my apartment, 20 minutes later, I had 3 empty spaces, and within 2 minutes it was finished. The solutions had been crowdsourced over Snapchat. It was such a delightful “old-person” use case for Snapchat: totally ephemeral therefore lightweight, but a way to engage people I don’t otherwise communicate with anywhere except rather obliquely on Twitter. I may make it a weekly tradition!