Chris Dixon

You’ve either started a company or you haven’t.  ”Started” doesn’t mean joining as an early employee, or investing or advising or helping out.  It means starting with no money, no help, no one who believes in you (except perhaps your closest friends and family), and building an organization from a borrowed cubicle with credit card debt and nowhere to sleep except the office. It almost invariably means being dismissed by arrogant investors who show up a half hour late, totally unprepared and then instead of saying “no” give you non-committal rejections like “we invest at later stage companies.” It means looking prospective employees in the eyes and convincing them to leave safe jobs, quit everything and throw their lot in with you.  It means having pundits in the press and blogs who’ve never built anything criticize you and armchair quarterback your every mistake. It means lying awake at night worrying about running out of cash and having a constant knot in your stomach during the day fearing you’ll disappoint the few people who believed in you and validate your smug doubters.

I don’t care if you succeed or fail, if you are Bill Gates or an unknown entrepreneur who gave everything to make it work but didn’t manage to pull through. The important distinction is whether you risked everything, put your life on the line, made commitments to investors, employees, customers and friends, and tried – against all the forces in the world that try to keep new ideas down – to make something new.

[Google CEO Eric] Schmidt said that while the Google Instant predictive search technology helps shave an average of 2 seconds off users’ queries, the next step is “autonomous search.” This means Google will conduct searches for users without them having to manually conduct searches. As an example, Schmidt said he could be walking down the streets of San Francisco and receive information about the places around him on his mobile phone without having to click any buttons. “Think of it as a serendipity engine,” Schmidt said. “Think of it as a new way of thinking about traditional text search where you don’t even have to type.”  - eWeek

When users type phrases into Google, they are searching, but also expressing intent. To create the “serendipity engine” that Eric Schmidt envisions would require a system that infers users’ intentions.

Here are some of the input signals a mobile device could use to infer intent.

Context

Location: It is helpful to break location down into layers, from the most concrete to the most abstract:

1) lat / long – raw GPS coordinates

2) venue – mapping of lat / long coordinates to a venue.

3) venue relationship to user – is the user at home, at a friend’s house, at work, in her home city etc.

4) user movement – locations the user has visited recently.

5) inferred user activity – if the user is at work during a weekday, she is more likely in the midst of work. If she is walking around a shopping district on a Sunday away from her home city, she is more likely to want to buy something. If she is outside, close to home, and going to multiple locations, she is more likely to be running erands.

Weather: during inclement weather user is less likely to want to move far and more likely to prefer indoor activities.

Time of day & date: around mealtimes the user is more likely to be considering what to eat. On weekends the user is more likely to be doing non-work activities. Outside at night, the user is more likely to be looking for bar/club/movie etc.  Time of days also lets you know what venues are open & closed.

News events near the user: they are at the pro sporting event, an accident happened nearby, etc.

Things around the user: knowing not just venues, but activities (soccer game), inventories (Madden 2011 is in stock at BestBuy across the street), events (concert you might like is nearby), etc.

These are just a few of the contextual signals that could be included as input signals.

Taste

The more you know about users’ tastes, the better you can infer their intent. It is silly to suggest a great Sushi restaurant to someone who dislikes Sushi. At Hunch we model taste with a giant matrix. One axis is every known user (the system is agnostic about which ID system – it could be Facebook, Twitter, a mobile device, etc), the other axis is things, defined very broadly: product, person, place, activity, tag etc.  In the cells of the matrix are either the known or predicted affinity between the person and thing.  (Hunch’s matrix currently has about 500M people, 700M items, and 50B known affinity points).

Past expressed intent

- App actions:  e.g. user just opened Yelp, so is probably looking for a place to go.

- Past search actions: user’s recent (desktop & mobile) web searches could be indications of later intent.

- Past “saved for later” actions:  user explicitly saved something for later e.g. using Foursquare’s “to do” functionality.

Behavior of other people

- Friends:  The fact that a user’s friends are all gathered nearby might make her want to join them.

- Tastemates: That someone with similar tastes just performed some actions suggests the user is more likely to want to perform the same actions.

- Crowds: The user might prefer to go toward or avoid crowds, depending on mood and taste.

How should an algorithm weight all these signals? It is difficult to imagine this being done effectively anyway except empirically through a feedback loop. So the system suggests some intent, the user gives feedback, and then the system learns by adjusting signal weightings and gets smarter.  With a machine learning system like this it is usually impossible to get to 100% accuracy, so the system would need a “fault tolerant” UI.  For example, pushing suggestions through modal dialogs could get very annoying without 100% accuracy, whereas making suggestions when the user opens an application or through subtle push alerts could be non-annoying and useful.

Startups that raise seed funding face the risk of not being able to raise additional money. This is what is sometimes known as “financing risk.”

If you are a company that just raising seed funding, financing risk should be top of mind.  Here are some tips for mitigating it:

- Start by thinking about the next round of financing and work backwards.  What milestones do you have to hit to get VC funded at an upround?  If you are a consumer internet company, the milestone probably involves getting a certain number of users.  If you are building hardcore tech, it probably means building a working prototype.  Basically you want to take the main risks that exist at the seed stage and eliminate as many as you can. A good way to discover what milestones you need to hit is to talk to as many VCs as possible. Experienced seed investors can also advise you on this.

- Raise enough seed money. How much money will it take to hit those milestones?  A good rule of thumb is 18 months – 3 months to get going, 12 months to execute, 3 more months to raise VC.  But it really depends on the specifics of the milestones, your operational plan, etc. which is why you need to figure those out first.

- Preserve cash.  Pay only subsistance wages but be generous with equity for great people (this also provides a screen for hiring people with the right startup mindset).  Keep legal fees low (try to keep incorporation and financing costs to $10K or lower – this is one reason I prefer convertible notes).  Act like a scrappy startup.

A rule of thumb is a successful Series A is one that is led by quality VCs with a pre-money at least 2x the post-money of the seed round.

There is an amazing amount of useful, free information available on tech blogs for fledgling tech entrepreneurs (this list is a great place to start). I think sometimes we techies forget that this wealth of content is unknown to the non-startup world. I was reminded of this recently when I met a first-time entrepreneur who said when he was first starting out he tried finding books on Amazon, Googling for stuff etc. He described it as an epiphany the first time he stumbled upon Fred Wilson’s blog, which then led him to Brad Feld, Mark Suster, Eric Ries, Venture Hacks, etc.

So this weekend I thought I’d try an experiment. I took about 100 of my blog posts (the ones that I thought were most “evergreen”), bundled them as a PDF and submitted them to the Kindle Store. The Kindle submission process was surprisingly easy. You give your book a name and upload the PDF and then choose pricing.  They force you to charge a minimum of $0.99.  Also, strangely, if you charge less than $2.99, Amazon takes 70% of the revenue, but if you charge between $2.99-$10 they only keep 30%.

I decided to price my book at $2.99 and donate all of the proceeds (~$2 book) to HackNY, a non-profit that “keeps the kids off the Street” (encourages college students to join/start tech startups instead of working on Wall Street). All of the content in the book is available for free on cdixon.org. The only reason to buy the book is to get this blog in a different format and to support a good charity. It is available in the Kindle Store here.

I don’t expect many people to buy the book but maybe some first-time entrepreneurs will stumble on it and from there discover more tech blogs. Think of it as “Kindle SEO” for tech blogs.

Finally, I am having trouble getting the links to work on the Kindle version. I’m not sure if this is an Amazon policy or if I am just doing something wrong (the links work fine in the PDF I uploaded to Amazon). So here is an alternative version on Scribd that has working links.

The TV industry is a major segment of the consumer electronics industry and Apple is the leading consumer electronics company in the world. Thus far Apple has entered the TV market with a stand-alone device, Apple TV. There has been speculation about whether Apple might enter the TV market by creating an actual TV. The most convincing objections to that idea cite the unfavorable industry structure: the power of the cable operators, the low margins on TVs, the infrequency of people buying new TVs, etc.

I thought it would be interesting to go back and look at the reasoning analysts used to predict the failure of the iPhone before its launch in 2007. Some predicted it would fail because the other handset makers would successfully compete with Apple:

The iPod also conquered the problem of small screens and cheesy navigation. With its newfound popularity, the company was also able to get music publishers to agree to its terms. Unfortunately for Apple, problems like that don’t exist in the handset business. Cell phones aren’t clunky, inadequate devices. Instead, they are pretty good. Really good. Why do you think they call it a Crackberry? Because the lumpy design and confusing interface of the device is causing people to break into cars? No, it’s because people are addicted to it. Samsung has scoured the world’s design schools and hired artists on three continents to keep its phones looking good. Motorola has revived its fortunes with design. KDDI, a Japanese carrier, has a design showcase in the teen shopping area of Tokyo just to be close to trends. And Sharp doesn’t skimp when it comes to putting LCD TVs on its phones. Apple, in other words, won’t be competing against rather doltish, unstylish companies like the old Compaq. The handset companies move pretty quick and put out new models every few weeks. [emphasis added]

Other analysts predicted Apple’s phone was doomed because of the mobile phone industry structure – mobile operators commanded so much power via subsidies, retail distribution etc:

Apple will launch a mobile phone in January, and it will become available during 2007. It will be a lovely bit of kit, a pleasure to behold, and its limited functionality will be easy to access and use. The Apple phone will be exclusive to one of the major networks in each territory and some customers will switch networks just to get it, but not as many as had been hoped. As customers start to realise that the competition offers better functionality at a lower price, by negotiating a better subsidy, sales will stagnate. After a year a new version will be launched, but it will lack the innovation of the first and quickly vanish. The only question remaining is if, when the iPod phone fails, it will take the iPod with it.  [emphasis added]

I am not citing these analysts to mock them. Hindsight is 20/20 and it was quite reasonable at the time to assume that a new phone from Apple would confront the same issues that new phones from other companies confronted. What Apple ended up doing, however, was creating a phone that was so incredibly desirable to consumers that it completely restructured the industry, causing a massive shift of power away from the carriers.

Regarding the TV industry, here is what Steve Jobs said last year at AllThingsD:

Q: Is it time to throw out the interface for TV? Does television need a new human interface.

A: The problem with innovation in the TV industry is the go-to-market strategy. The TV industry has a subsidized model that gives everyone a set top box for free. So no one wants to buy a box. Ask TiVo, ask Roku, ask us… ask Google in a few months. The television industry fundamentally has a subsidized business model that gives everyone a set-top box, and that pretty much undermines innovation in the sector. The only way this is going to change is if you start from scratch, tear up the box, redesign and get it to the consumer in a way that they want to buy it. But right now, there’s no way to do that….The TV is going to lose until there’s a viable go-to-market strategy. That’s the fundamental problem with the industry. It’s not a problem with the technology, it’s a problem with the go-to-market strategy….I’m sure smarter people than us will figure this out, but that’s why we say Apple TV is a hobby.

So Jobs doesn’t believe an “additional box” is a viable strategy for seriously entering the TV industry. This leaves three places to enter: 1) integrating into set top boxes, 2) integrating into other TVs, or 3) Apple creating its own TV. Regarding #1, the last thing the cable operators want is for internet-delivered programming that bypasses their cable channels to become widespread – they see that as the fast track to become a dumb pipe. Re #2: This just seems very unlike Apple – the most vertically integrated company in tech, and famous for wanting to control every aspect of the product and user experience.

Re #3, let’s imagine Apple develops a TV that is as groundbreaking as the iPhone was. The biggest problem “smart TVs” have today is that they need clunky IR transmitters to control set top boxes because the cable operators won’t willingly interoperate. So a new Apple TV would have to drum up such incredible consumer demand that the operators would feel compelled to support it. This does indeed seem harder in the TV than in the mobile industry. At least in the US you had 4 nationwide mobile operators at the time of the iPhone launch. In TV, consumers normally have at most two real choices for traditional cable programming – cable and satellite – and two real choices for two-way internet – cable and DSL/FIOS.

Perhaps Apple won’t enter the market due to its structure. But that didn’t stop them in mobile phones where the structure was similarly difficult. The mistake analysts made about the iPhone was to assume the current industry structure would be sustained after Apple’s entry. I’d be wary of making the same assumption about the TV industry.

Mark Twain famously quipped that “80 percent of life is showing up.” Running a startup, I’d say it’s more like 90 percent. For example, I frequently hear from founders how it’s hard to recruit programmers. It is indeed hard. But great programmers are out there, and can be found in places where other people simply aren’t showing up.

Back in 2005 when I was starting SiteAdvisor with Tom Pinckney (one of my cofounders at my last two startups and non-graduate of high school) we were trying to recruit great programmers. At the time, startups were certainly not the hot thing, especially on the East Coast. We were based in Boston so decided to spend time at MIT where we figured there must be smart programmers. We went to places like the Media Lab and basically just sat ourselves down at lunch counters and awkwardly introduced ourselves: “Hi, my name is Chris Dixon and this is Tom Pinckney and we are starting a company and would love to talk to you about it.” Most students ignored us or thought we were annoying. I remember one student staring at us quizzically saying “startups still exist?” Most of our trips were fruitless. At one point after a failed trip we were on the Redline back to our office in downtown Boston and joked, depressingly, that we felt so out of place that people looked at us like time travelers from the dot-com bubble.

Our first breakthough came after a series of trips when a particularly talented programmer/designer named Hugo Liu re-approached us and said something like “hey, actually I thought about it and your idea doesn’t suck.” Then his friend David Gatenby talked to about joining us. We eventually recruited Hugo and David along with a brilliant undergraduate Matt Gattis. We had finally broken through. Matt and Hugo now work with us at Hunch along with some of their friends from MIT they brought along.

People who say recruiting is easy are probably recruiting bad people. People who say recruiting is hard are right. People who say it is impossible just aren’t showing up enough.

It is widely believed that Facebook presents a significant competitive threat to Google. Google itself seems to believe this – Larry Page recently said that all employees would have their bonuses tied to the success of Google’s social strategy.

Why does Facebook present a threat to Google?  A few reasons:

- The utility of Google’s core product – web search – depends on the web remaining fragmented and crawlable.  Facebook has become the primary place web users spend their time and create content, and is mostly closed to Google’s crawlers.

- Facebook controls a large percentage of ad impressions and will likely launch an off-Facebook.com display ad network to compete directly with Google’s display ad business (built from its $3.1B acquisition DoubleClick). It is generally thought that display ads will become a larger portion of online advertising spend (versus direct response text link ads) as more brand advertising moves online.

- There are many other “wildcard” risks – e.g. Facebook competing with Google (and Apple, Paypal etc) in payments, Facebook gaining power on mobile (threatening Android), and the possibility of a greater share of internet intent harvesting happening on Facebook through not-yet-released features like a search and/or shopping engine.

When going after Facebook, Google has at least three key strategic choices to make:

Strategic choice #1: Should Google try to make social networking commoditized or new profit center? (For more about what I mean by this, please see this post on Google’s overall strategy and these posts on “commoditizing the complement” here, here and here).

The advantage of creating a new social networking profit center is obvious: if you win, you make lots of money. The advantage of commoditizing social networking is that although you forgo the potential direct profits, you open up a wider range of pricing and product options. For example:

- When you try to commoditize a product, you can offer a product for free that other companies charge for. This is what Google did with Android vs iOS and Google Apps vs Microsoft Office. Of course, making social networking free to users won’t work since Facebook doesn’t directly charge users (I say “directly” because they make money off advertising & payment commissions, among other ways). However reducing the cost to zero for 3rd-party developers like Zynga who have to pay Facebook large commissions would entice them toward a Google platform (note that, not coincidentally, Google invested $100M in Zynga).

- Interoperate / embrace open standards – Normals don’t care whether a product uses open standards, but by interoperating with other social networks, messaging systems, check-in services, etc., Google could encourage 3rd-party developers to build on their platform. If Google chose, say, RSS for their messaging system, it would already work with tens of thousands of existing tools and websites and would be readily embraced by hackers in the open source community. The web itself (http/html) and email (smtp) are famous examples where the choice to open them unleashed huge waves of innovation and (eventually) killed off closed competitors like AOL.

Strategic choice #2: How should Google tie its new social products into its existing products?

Besides a mountain of cash ($30B net, generating $10B more per year), Google has many existing assets on top of which to build. Google Buzz tried to build off of the “implicit social network” of Gmail contacts, which hasn’t seemed to work so far and raised privacy concerns.

Google’s recent mini-launch of its “+1″ button seems to be good use of the strategy known as “anchoring”. Google is apparently trying to create a federated network where websites embed +1 buttons  the way they embed Facebook’s Like button except the +1 button would be a signal into Google’s organic ranking algorithm (as an aside, this is where Gmail becomes useful as having millions of logged in users makes spamming +1 buttons much harder). Websites care a lot about their Google organic search rankings (which is why, for example, helping websites improve their rankings is multibillion-dollar industry).  A button that improved search rankings would likely get prominent placement by many websites. Making +1 appealing to users is another story.  The user value is much clearer for the Facebook Like and Twitter Tweet buttons – you send the link to your friends/followers. Providing value to users in addition to websites is a good reason for Google to acquire Twitter (something I think is inevitable if Google is serious about social – see below).

Finally, Android and YouTube are intriguing potential anchors for a social strategy. I’ll leave it to smarter people to figure out exactly how, but products with such large footprints always present interesting tie-in opportunities.

Strategic Choice #3: Should Google buy or build?

Historically, it is very rare to see tech companies adjust their “DNA” from within. Google’s best new lines of business over the past few years came through the acquisitions of YouTube and Android. Moreover, these acquisition were unusual in that they were left as semi-independent business units. Facebook’s hold on social is incredibly strong – besides the super-strong network effects of its social graph, Facebook has made itself core infrastructure (e.g. Facebook Connect) throughout the web. If Google really wants to catch up, they’ll need to go back to the strategy they succeeded with in the past of acquiring relevant companies and letting them run as separate business units.

* Disclosure: I’m an investor in a bunch of startups, so you could reasonably argue I’m highly biased here.

At around 5pm today two women who looked like professional models showed up at the Hunch office.  They had a delivery and said “we don’t know who this is from but it’s someone who likes the Taste Graph.”  Presumably this is an allusion to the taste graph we’ve built at Hunch and that I’ve talked about in various talks about graphs.

They hurried away without saying anything more.  In the box was a left worn Prada shoe:

A right shoe from Nike:

A USB drive containing the Dave Matthew’s song “you might die trying”:

Three books that were shut with magnets that made them almost impossible to open and mostly destroyed them once we did.  The books were Guns, Germs and Steel;  All The Money in The World; and The Grand Design:

Finally, there was an invitation to a fancy restaurant with the inviting people being two stars from Terminator 2:

Notice who it is from!  I tried calling the restaurant but no one answered.

Reminder that John Connor in T2 is the future leader of human rebels:

And Miles Dyson is a computer scientist who sacrifices himself to save humanity:

Now I see at least three possible explanations for this very strange delivery today:

1) it’s some company trying to pull a gig to raise money since they know I am a small-time investor

2) it’s some kind of weird practical joke

3) Although Hunch’s technology is only semi-sophisticated today, it will become so sophisticated in the future that people in, say, 2050 will feel compelled to go back in time and reward/punish us.

At any rate, I hereby proclaim today the day for time travelers to bring it on. If this blog post gets picked up by bigger blogs or even newspapers then I hope in the future they will remember to read those and know that April 4, 2011 was the day to make it happen.

I’m talking to you future robot motherf—ers who send weird deliveries.

I’ve downloaded and tested a few hundred iPhone and iPad apps.  One thing that I’ve noticed is that many of the top rated and ranked apps are pretty scammy.  Take for example “Night Vision.”

It’s a top app in under Utilities for both paid and free iPhone apps.

If you actually download and test the app, you’ll find it doesn’t work at all. In fact, I found it made objects darker, not brighter.  See these photos with and without the app of the exact same room in the exact same lighting.

The app tries to get you to download other apparently scammy apps.  I’m guessing this kind of “cross selling” is how Night Vision got  most of its downloads.

Another clever trick they play is when you look at the app customer ratings on the iPhone App Store you see that it has 4.5 stars:

But when you look on the desktop web you see the overall ratings are vastly lower and that they seem to game the system by releasing “new versions” to reset their ratings and then probably paying people to write positive reviews:

Companies like TapJoy let you pay to get in the Top 25, and then once you are there you can get “organic” downloads by being on the toplists.

Another platform, another way to game it.