Chris Dixon

There has been a lot of discussion lately about the markets for startup financing. Many of the discussions use words like “valuations” “bubble” “crunch” etc. Words like that generally mean the writer is discussing the world through the lens of finance. This is a useful lens, but I’d like to suggest there is another lens that is also useful: the product lens. First, some background.

Two markets

Startups sit in the middle of two markets: one between VCs and startups, and one between startups and customers. These markets are correlated but only partially. When the financing supply is low but customer demand is high, entrepreneurs that are able to finagle funding generally do well. When financing and startup supply is high, customers do well, some startups do well, and VCs generally don’t. And so on.

When VCs get too excited, people talk about a bubble. When VCs get too fearful, people talk about a crash. Historically, downturns were great times for startups that were able to raise money because competition was low but customer demand for new technology remained fairly steady. Downturns also tended to coincide with big platform shifts, which usually meant opportunities for entrepreneurs.

These markets shift independently between different stages and sectors, although there are connections. The amount of financing available is relatively constant, because of the longevity of VC funds and the way most VCs are compensated (management fees). Less financing in one sector or stage usually leads to more financing in others.

The stages are related because the early stages depend on the later stages for exits and financings. The result is a bullwhip effect where changes in later stages (the latest stage being public markets) lead to magnified changes in early stages.

Smart VCs understand these dynamics and adjust their strategies accordingly. Smart entrepreneurs don’t need to think about these things very often. Fundraising is necessary (at least for companies that choose to go the VC route – many shouldn’t), but just one of the many things an entrepreneur needs to do. The best advice is simply to raise money when you can, and try to weather the vicissitudes of the financial markets.

The product lens

Good entrepreneurs spend most of their time focusing on the other market: the one between their company and their customers. This means looking at the world through the lens of products and not financing. This lens is particularly important when you are initially developing your idea or when you are thinking about product expansions.

The product lens suggests you should ask questions like: have the products in area X caught up to the best practices of the industry? Are they reaching their potential? Are they exciting? Are there big cultural/technological/economic changes happening that allow dramatically better products to be created? Sometimes the product lens guides you to the same conclusion as the finance lens and sometimes it doesn’t.

For example, there has been a lot of hand wringing about a financing crunch for consumer internet startups. One theme is that investors are pivoting from consumer to enterprise. The finance lens says: for the last five years or so, consumer was overfunded and enterprise was underfunded – let’s correct this. It also helps that enterprise IPOs have performed much better than consumer IPOs in the last year or so.

The product lens is tricky. My sense is that, at least for the non-mobile consumer internet, the product lens and financing lens agree. Anyone who has had the misfortune to use enterprise technology lately will tell you that the hardware and software they use at home (iPhone, Gmail, etc) is far and away more sophisticated and elegant than the software they use at work. It feels like the enterprise tech is way behind in the product upgrade cycle.

Mobile seems like a case where the lenses disagree. The finance lens says: billions of dollars have been invested in mobile apps. It has become hit driven and there have been very few “venture-scale” startups created.

The product lens says: the modern smartphone platform began about four years ago when the iOS app store launched. This is clearly a major new platform. Platforms and apps interact in a push-pull relationship that takes decades to play out. Innovative new apps, designs and technologies are created all the time. It would be surprising – and contrary to all the historical patterns – if the mobile product evolution were already played out.

That is not to dismiss the finance lens. It could be painful along the way:  financing markets might dry up, and profits might accrue to the platforms over the apps. But clearly mobile is just getting started.

Some of the biggest mistakes I’ve made as an angel investor stemmed from being beholden to the finance lens. The finance lens feels more scientific and therefore appeals to analytical types. It might sound unsophisticated to say “the products for X are crappy, and I have an idea for how to make them great.” But in many cases, it’s actually that simple.

Andrew Parker had a great post a few years ago where he sketched out all the startups going after pieces of Craigslist:

Startups that have tried to go head-to-head against the entirety of Craigslist (the “horizontal approach”) have struggled. Startups that have tried to go up against pieces of Craigslist (the “vertical approach”) have been much more successful (e.g. StubHub, AirBnB).

Recruiting looks like it’s going through a similar evolution. Last-generation products like LinkedIn are broad but not deep. Everyone I know who recruits uses LinkedIn, but none of them think it has solved their recruiting problems. Now we are seeing the rise of vertical solutions that are significantly better, e.g. Stack Overflow for developers and Behance for designers (at least that’s what I believe – I’m an angel investor in both).

The benefits of focusing are: 1) you can create a dramatically better user experience when it’s tailored to a specific use, 2) you can do unscalable hacks when starting out (e.g. AirBnb paying photographers to take pictures of apartments), 3) you need far fewer users to get to minimum viable liquidity, and 4) brand building is easier when you solve a straightforward, narrow problem (e.g. “I need a place to stay this weekend”).

This pattern – horizontal first, vertical second – is common. But you need to be careful. Back in 2003-2004, there was a lot of speculation that vertical search engines would eventually take down Google. A few categories worked (e.g. travel), but Google adapted in other categories (e.g. video, news) and lots of startups suffered.

VCs are experts at analyzing industries and identifying new opportunities, which is why it’s odd that the VC industry itself has so stubbornly resisted change.

Two years ago I wrote a post where I argued that innovative new VC firms are finally starting to change this:

Top tier entrepreneurs are frequently selecting their investors, not vice versa. The VCs most sought after are mostly new firms: big firms like Andreessen Horowitz, Union Square Ventures, and First Round, and micro-VCs like Floodgate (fka Maples), Betaworks, and Ron Conway.

Since then, the trend has become even more pronounced. VC is only partly about investing. It is primarily a service business whose purpose is to help entrepreneurs.

When Andreessen Horowitz (“a16z”) started out three years ago, like a lot of people I thought “OK, really interesting entrepreneurial founders, but how will they be as investors?” Then I started hearing chatter among entrepreneurs that they really wanted to raise money from them. “We’re talking to X, Y, & Z — but Andreessen is the firm we really want” became an increasingly common refrain.

Earlier this year I got to meet the a16z team and observe the operation directly. There are over 60 people at the firm. Only six people do traditional VC activities: investing, joining boards, and helping out. The rest are exclusively focused on helping entrepreneurs.

The “startup idea” behind a16z is: instead of spending the bulk of the fund fees on partner salaries, spend it on operations to help entrepreneurs. There is a marketing team (=helps you get noticed), a talent team (=helps you recruit), a market development team (=helps you get customers), and a research team (=helps you figure stuff out).

Spending time there, I had the same feeling I have whenever I meet a great startup: “This is obviously the future, why didn’t someone do it before?”

So I’m super excited to say that I’m joining a16z as their seventh General Partner. I’ll specialize in consumer internet investments but will be open to anything ambitious that involves technology. I’ll be based in California, but plan to do a lot of investing in NYC.

I’ll miss seeing my Hunch colleagues on a daily basis. Many of us have been working together for eight years, through two startups. I’d also like to thank everyone at eBay for being so welcoming and supportive.

“Agency problems” are what economists call situations where a person’s interests diverge from his or her firm’s interests.

Large companies are in a constant state of agency crisis. A primary role of senior management is to counter agency problems through organizational structures and incentive systems. For example, most big companies divide themselves into de facto smaller companies by creating business units with their own P&L or similar metric upon which they are judged. (Apple is a striking counterexample: I once pitched Apple on a technology that could increase the number of iTunes downloads. I was told “nobody optimizes that. The only number we optimize here is P&L in the CFO’s office”).

If you are selling technology to large companies, you need to understand the incentives of the decision makers. As you go higher in the organization, the incentives are more aligned with the firm’s incentives. But knowledge and authority over operations often reside at lower levels. Deciding what level to target involves nuanced trade offs. Good sales people understand how to navigate these trade offs and shepherd a sale. The complexity and counter-intuitiveness of this task is why it’s so difficult for inexperienced entrepreneurs to sell to large companies.

Agency problems also exist in startups, although they tend to be far less dramatic than at big companies. Simply having fewer people means everyone is, as they say in programming, “closer to the metal”. The emphasis on equity compensation also helps. But there are still issues. Some CEOs are more interested in saying they are CEOs at parties than in the day-to-day grind of building a successful company. Some designers are focused on building their portfolio. Some developers are only interested in intellectually stimulating projects. Every job has its own siren song.

One of the reasons The Wire is such a great TV show is that it shows in realistic and persuasive detail how agency problems in large organizations consistently thwart well intentioned individual efforts. The depressing conclusion is that our major civic institutions are doomed to fail. Those of us who are technology optimists counter that the internet allows new networks to be created that eliminate the need for large organizations and their accompanying agency problems. Ideally, those networks recreate the power of large organizations but operate in concert like startups.

There’s been a lot of speculation lately about why big companies spend millions of dollars acquiring startups for their technology or talent. The answer lies in the economic logic that big companies use to make major project decisions.

Here is a really simplified example. Suppose you are a large company generating $1B in revenue, and you have a market cap of $5B. You want to build an important new product that your CTO estimates will increase your revenue 10%. At a 5-1 price-to-revenue ratio, a 10% boost in revenue means a $500M boost in market cap. So you are willing to spend something less than $500M to have that product.

You have two options: build or buy. Build means 1) recruiting a team and 2) building the product. There is a risk you’ll have significant delays or outright failure at either stage. You therefore need to estimate the cost of delay (delaying the 10% increase in revenue) and failure. Acquiring a relevant team takes away the recruiting risk. Acquiring a startup with the product (and team) takes away both stages of risk. Generally, if you assume 0% chance of failure or delay, building internally will be cheaper. But in real life the likelihood of delay or failure is much higher.

Suppose you could build the product for $50M with a 50% chance of significant delays or failure. Then the upper bound of what you’d rationally pay to acquire would be $100M. That doesn’t mean you have to pay $100M. If there are multiple startups with sufficient product/talent you might be able to get a bargain. It all comes down to supply (number of relevant startups) and demand (number of interested acquirers).

Every big company does calculations like these (albeit much more sophisticated ones). This is a part of what M&A/Corp Dev groups do. If you want to sell your company – or simply understand acquisitions you read about in the press – it is important to understand how they think about these calculations.

A common way to think of business regulations is by analogy to sports: the rules are specified up front, and the players follow the rules. But real regulations don’t work that way. Regulations follow business as much as business follows regulations.

Sometimes the businesses that change regulations are startups. Startups don’t have the resources to change regulations through lobbying. Instead, they need to start with regulatory hacks: “back door” experiments that demonstrate the benefits of their ideas. With luck, regulators are forced to follow.

Nextel was one of the all-time great regulatory hacks. In the late 80s and early 90s, the FCC’s rules banned more than two cellular operators per city. As Nextel’s cofounder said, “the FCC thought a wireless duopoly was the perfect market structure”. Nextel (called Fleet Call at the time) circumvented these rules by acquiring local (e.g. taxi, pizza truck) dispatch radio companies, which they then connected to create a nationwide (non-dispatch) cell phone service.

Predictably, the cellular incumbents tried to regulate Nextel out of existence. From a 1991 New York Times article:

In a move that could threaten cellular telephone companies, the Federal Communications Commission may decide on Wednesday to grant a small radio company’s request to provide a new form of mobile telephone service in six major cities, including New York. If the request is approved, the action could inject new competition into the industry. At the moment, Federal rules permit only two cellular systems to operate in any city. But the new proposal could open up a regulatory back door, allowing companies that provide private radio service for taxi fleets and delivery services to offer mobile telephone services to individuals…. The proposal has alarmed the industry, which has heatedly opposed it and enlisted support in Congress late last year to delay the F.C.C.’s decision.

The incumbents argued that Nextel’s service would interfere with public safety frequencies and therefore endanger the public. They also argued that Nextel’s service would be too expensive:

Some analysts contend that the radio handsets for Fleet Call and its imitators will be more expensive than cellular units. The technical features of cellular equipment are now standardized nationwide, making it possible to bring down costs through higher selling volumes. Specialized mobile services are currently different in each city.

And their call quality would be inferior:

Some analysts contend that Fleet Call’s local service is likely to be inferior as well. “It is highly unlikely to be as good as cellular service,” said Denise Jevne, telecommunications analyst with T. Rowe Price Associates in Baltimore.

The FCC eventually decided not to block Nextel. Nextel grew to become a top five US cellular operators before it was acquired by Sprint in 2004 for $35B. Their service turned out to be cost-competitive, high quality, and safe. The only thing endangered were the incumbents’ profits.

What Nextel faced in 1991 is very similar to what many startups face today. Uber is being threatened by the taxi industry, Aereo by the TV broadcasting industry, and Airbnb by the hotel industry. Some industries, like finance, are so heavily regulated that almost any new idea runs into regulatory objections.

Of course regulations that truly protect the public interest are necessary. But many regulations are created by incumbents to protect their market position. To try new things, entrepreneurs need to find a back door. And when they succeed, it will all look obvious in retrospect. Today’s regulatory hack is tomorrow’s mainstream industry.

Building an enterprise software company used to be largely about sales, because enterprise software was sourced and purchased by high-level business people. Those business people needed to be charmed and convinced, an activity that was distasteful to many technologists.

Internet-based delivery (“SaaS”, “cloud”) dramatically lowered installation costs, letting individuals or small groups buy software on discretionary budgets or use basic versions for free. As adoption spread throughout the organization, the value of the software eventually percolated up to high-level business people who could write large checks to get features big companies need, such as administration, security, integration, compliance, and support. This ”bottom-up” approach was pioneered by Salesforce and open source companies like MySql. Recent enterprise success stories also follow this model, e.g. New Relic, Yammer, Twilio, and Github. Many of these companies have processes that would have seemed crazy ten years ago – e.g. sales people only handle inbound inquiries or only call customers who already use their product.

Thus enterprise software went from being about sales (one-to-one) to being about marketing (one-to-many). Marketing requires crafting a compelling message, figuring out the right channels and then optimizing. But the most effective marketing is a compelling product that can be easily tried. As a result, as Benchmark’s Peter Fenton said recently: ”We’re seeing a fundamental shift from sales-driven companies to product-driven companies. The companies that are leading the way there let this consumer and product focus permeate the culture of their companies.”

One of the most visible manifestations of this shift is the refreshingly accessible language on modern enterprise websites. Sales-driven enterprise software companies speak the arcane language of CIOs. Marketing-driven companies talk directly to business users (e.g. Yammer) or developers (e.g. Github).

This is good news all around. Enterprises are more likely to get software that incorporates the advances made over the last decade in consumer software. Startups get a shot at creating this software, and get to do so on a fairly level playing field. The product and marketing focus should also attract a lot more technologists who were turned off by sales. The only losers are incumbents who continue to pursue the old model.

The best way to think of Facebook’s stock is as the sum of two businesses: the existing display ad businesses, and a probability-weighted option on a new line of business. This is how Wall Street views it. For example, here is a section of a recent Goldman Sachs analyst report on Facebook:

Optionality not in the model: further potential upside

While not in our model, as [Facebook] has not publicly expressed pursuit of these areas, we believe there are three obvious opportunities that the company could leverage its platform to capitalize on:

- Developing an external ad network

- Monetizing paid search

- Entering China

Of the three options, search is clearly the most interesting. An external ad network is inevitable. Google proved this model with Adsense. With an already huge base of advertisers bidding on CPCs, it is impossible for most other ad networks to compete on publisher payouts. But Facebook’s traffic is so great now that an external ad network might increase their revenues by 2x or so. The same goes for entering China. They might get another half a billion users who monetize at lower ad rates than US users. Neither move would put them in Google’s revenue range. They need a better business model for that. The only (known) models that deliver RPMs high enough to compete with Google are search, payments, and e-commerce.

At TechCrunch Disrupt last week, Mark Zuckerberg talked about possibly entering the search business. Investors had been concerned that maybe Zuckerberg really meant what he said in his IPO letter – that he just didn’t care that much about making money. By expressing an interest in search, Zuckerberg signaled that he understood Facebook’s immensely valuable embedded option and was thinking about ways to exercise it.

Eric Ries uses the phrase “vanity metrics” to refer to metrics that founders cite to demonstrate progress but that are actually false signals. A related concept is “vanity milestones”: achievements that are more about making you feel good than helping your company. Vanity milestones include:

- Raising money from famous people/firms who aren’t really going to help your company (e.g. Hollywood celebrities).

- Partnerships with brand name organizations that aren’t really going to help your company.

- Getting press (e.g top lists) that focuses on founders and not your company.

- Almost all tech press (unless your product targets developers or tech companies).

This doesn’t mean it’s bad to hit vanity milestones. Good companies hit lots of vanity milestones along the way, and sometimes they can be a morale boost for employees. What is worrisome is when founders equate vanity milestones with success. The attention will go away very quickly if your company fails.

Ten years ago, startup financing was an insider’s game. Since then, the topic has been widely discussed on blogs, to the great benefit of entrepreneurs. Comparatively little, however, has been written about the important transaction at the other end many startups’ life, acquisitions. Here are some things I’ve learned about the acquisition process over the years.

- There is an old saying that startups are bought not sold. Clearly it is better to be in high demand and have inbound interest. But for product and tech acquisitions especially, it is often about getting the attention of the right people at the acquirer. Sometimes the right person is corp dev, other times product or business unit leads, and other times C-level management.

- Don’t use a banker unless your company is late stage and you are selling based on a multiple of profits or revenues. I’ve seen many acquisitions bungled by bankers who were either too aggressive on terms or upset the relationship between the startup and acquirer.

- Research the potential acquirer before the first meeting. Try to understand management’s priorities, especially as they relate to your company.  Talk to people who work in the same sector. Talk to industry analysts, investors, etc. If an acquirer is public, Wall Street analyst reports can be helpful.

- Develop relationships with key people – corp dev, management, product and business unit leads. The earlier the better.

- Don’t try to be cute. Leaking rumors to the press, creating a false sense of competition, etc. is generally a bad idea. Besides being ethically questionable, it can create ill will.

- What you tell employees is particularly tricky. Being open with employees can lead to press leaks and can annoy acquirers. Moreover, some public companies insist that you don’t talk to employees until the deal is closed or almost closed. Employees usually get a sense that something is going on and this can put you in the awkward situation of being forced to lie to them. I don’t know of a good solution to this problem.

- Understand the process and what each milestone along the way means. As with financings, acquisitions take a long time and involve lots of meetings and difficult decisions. Inexperienced entrepreneurs tend to get overly excited about a few good meetings.

- Strike while the iron is hot. Just as with financings, you need to be opportunistic. Waiting 6 months to hit another milestone might improve your fundamentals, but the acquirer’s interest might wane.

- There are two schools of thought on price negotiation: anchor early or wait until you’ve gotten strong interest. Obviously having multiple interested parties makes finding a fair price a lot easier.

- Deal structure: the cap table is an agreement between you and the shareholders that says, in effect: “If we sell the company, this is how we pay out founders, employees, and investors.” Acquirers have gotten increasingly aggressive about rewriting cap tables to 1) hold back key employee payouts for retention purposes, and 2) give a greater share of proceeds to employees/founders.  Some even go so far as to try to cut side deals with key employees to entice them to abandon the other employees and investors. In terms of ethics and reputations, it is important to be fair to all parties involved: the acquirer, founders, employees, and investors.

- Research the reputation of the acquirer, especially how they have behave between LOI and closing (good people to talk to: investors, other acquired startups, startup lawfirms). This is when acquirers have all the leverage and can mistreat you. Some acquirers treat LOIs the way VCs treat term sheets, as a contract they’ll honor unless they discover egregious issues like material misrepresentations. Others treat them as an opportunity to get free market intelligence.

- Certain terms beyond price can be deal killers. The most prominent one lately is “IP indemnification.” This is a complicated issue, but in short, as a response to patent trolls going after IP escrows, acquirers have been trying to get clawbacks from investors in case of IP claims. This term is a non-starter to institutional investors (and most individual investors). You need to understand all the potential deal-killer terms and hire an experienced startup law firm to help you.

- Ignore the cynical blog chatter about “acqui-hires” (or, as they used to be called, “talent acquisitions”). Only people who have been through the process understand that sometimes these outcomes are good for everyone involved (including users when the alternative is shutting down).

Finally, acquisitions should be thought of as partnerships that will last long after the deal closes. Besides the commitments you make as part of the deal, your professional reputation will be closely tied to the fate of the acquisition. This is one more reason why you should only raise money if you are prepared for a long-term commitment.