The Edge of Democracy

If you haven’t already read Edward Snowden’s interview with the Guardian I suggest you take a few minutes and dig in.  For the uninitiated I will summarize it in two sentences. 1.) The US government is collecting a surprising amount of information on its own citizens. 2.) We should be concerned about the use of this data once our technology becomes more advanced.

I have been thinking about the second point for the past few weeks.  It is hard to predict what technology will exist in the future, but there are signposts showing the general direction.  Think back to 2001 when the original iPod was state of the art technology. Five million years of human innovation went into creating a device that was basically a hard drive with an audio jack.  A decade later we have pocket computers that weigh a few ounces, communicate via wireless broadband, run on solid state memory, and consume a fraction of the power.  Innovation is happening exponentially.

Modern devices can be used to record live events as they are happening, like Mitt Romney’s infamous “47%” speech, and broadcast them to everyone in the world using social media.

So how can this technology harm us?

In the wake of the Paula Deen squabble I read an online interview where Paula blamed the Jews for her press debacle.  It was a horrible story, with a dozen direct quotes.  I researched the source and the story turned out to be a hoax.  However, it still reached a thousand Facebook comments within minutes of being posted.  The same thing happened last year when Paul Krugman, a notable supporter of Obama’s fiscal policy, declared personal bankruptcy.  These fake stories were not even a blip on the public consciousness, but it shows how easily we can be misled.  And these were just simple text articles.

The good news is that audio and video are much harder to forge than text.  It took Apple and SRI over ten thousand man-hours to produce the SIRI voice, and it still sounds like a robot.  But that is changing too.  I recently met with a company that can create a similar voice automatically using a few hours of source material.  It is only a matter of time before we have the processing power and algorithms to make cloned speech indistinguishable from actual speech.  And after that cloned video will be indistinguishable from real video. My guess is that in 10 years, an average person using a wearable computer will be able to literally put words in people’s mouths and publish it to the entire world anonymously.

This is a worrisome version of Snowden’s future.  An independent media is vitally important to a Democracy – whenever there is a political coup, the first move a dictator makes is to take control of the airwaves.  But imagine a future where we have a free media, but every piece of content is manipulated.  Every recording is questionable.  Every video is a forgery or a hoax.  And there is no way to tell the difference.

Maybe this will once again produce a golden age for journalism, although instead of writers the media companies will employee fact checkers.  Or maybe the NSA, with their giant databases of human behavior, will be able to verify fact from fiction.  Either way I think it is going to be a bumpy ride.

Why Apple Doesn’t Make TVs

Have you ever wondered why a cable box, Roku, and PC are not embedded into your television set?  On the surface it seems like a logical next step – hardware is cheap, devices are small, and convergence has been a buzzword for a decade.  Apple has even been rumored to have an iTV in the works for four years.  So what gives?

I hate to be the bearer of bad news, but there is a reasonable chance that a unified TV/PC will not happen.

I sat down last week with the head of Global Business Development for Samsung to get a sense of where things stand.  As it turns out, Samsung first tried to embed basic WiFi in television sets ten years ago.  Many consumers weren’t able to get it to work with their home network, and in the end they returned the TVs for another model.

Seems like a small issue, but it was a disaster for Samsung.  Every returned television set cancelled out the profit of six TVs sold at full price.

All television companies have learned this lesson at one point or another.  They now hold back technical advances until they are foolproof.  Good luck getting a feature approved if it risks confusing the customer.

In January I had a chance to meet Steve Ballmer, who spent 15 minutes discussing Microsoft’s television strategy.  Microsoft purchased Perceptive Pixel in 2012, which makes giant touchscreen displays, and many people assumed MSFT would go into the television business too.

Instead Steve Ballmer rattled off the problems with television manufacturing:  seven year replacement cycles, low margins, expensive returns, high shipping costs.  He said that TVs are furniture, and Microsoft has no intention of going into the furniture business.

So my belief is that we will always have devices attached to TV sets, which bodes well if you are an investor in Roku or Fan TV (in our portfolio).  I also expect a lot of new entrants in this area over the next five years, particularly from Android-based devices.  Companies like Comigo out of Israel are lowering the bar for carriers and operators to get into the hardware business.

The only hope of reversing this trend is if Apple can convince consumers to pay twice as much for a truly converged TV set.  Apple has done it before, but not under the current leadership.  And the likelihood is that they would not accept returns.

Fools Gold

SB_BuckI have decided in the past few days that I am going to create a new currency.  It is going to be called “Ian Bucks”, or just “IB” for short.  If you are interested in purchasing some IBs please write below in the comments and I will contact you about it.

Here is how it will work.  Initially, 1 IB will equal $1 USD.  I have created an elaborate scheme to make more IBs, with 1 IB produced every 10 minutes.  Unless at some point I change my mind and make more.

I am calling the process of making IBs “mining” because that is how gold is made.  If you want a mining permit please call me and I will send you a floppy disk with instructions.

If you wish to purchase IBs you will have to give me US dollars and I will issue you IBs.

There are currently no merchants that accept IBs.  This is a speed bump and it will be fixed over time.

If you wish to redeem IBs with me, I reserve the right to give you Bitcoin in return.  This as a win-win because Bitcoins only go up over time.  In the event Flooz comes back I also reserve the right to give you Flooz as well.  There is no guarantee that Flooz will come back and I have no knowledge of Whoopi Goldberg’s future involvement.

I am planning to create an exchange for the trading of IBs.  Unfortunately, “Magic the Gathering: Online Exchange” is already taken by Bitcoin (aka Mt. GoX), so I am thinking of other cool MTG names like “MTG: Shivan Dragon Exchange” (aka Mt. GSDX)

Shivan Dragon


“MTG: Mahamoti Djinn Exchange” (aka Mt. GMDX):

Mahamoti Djinn


I just can’t decide.  A lot of people play Magic the Gathering so this will definitely go viral.

I promise to write some software so IBs become hard to counterfeit.  I am not particularly good at writing software but I know people who can do this for me.

I have a detailed marketing plan to help get support of IBs.  My car has been wrapped with the IB logo and I purchased a super bowl ad in February.  I might also use sponsored posts on Twitter because I got a free allocation.

You may be thinking to yourself, “Why would anyone want IBs??”  I was wondering the same thing, so I came up with a couple use cases:

1.) Buying drugs.

2.) Evading sales tax.

3.) Funding Al-Qaeda.

4.) Speculation.

5.) Getting money out of China.

6.) Getting money into China.

7.) General money laundering.

There are many more uses. Some of them are even legitimate.

I have not raised outside money for IBs yet.  I plan to keep it closely held.  If, however, you know any VCs I would be happy to meet with them because this is a big idea.

NOTE: This article is a work of satire.

Selling Courage

How to Succeed In the musical “How to Succeed In Business Without Really Trying”, JP Finch finds himself in a predicament.  In a span of just a few days he ascends from the mail room to the corner office, but in the process he accidentally becomes the Vice President of Advertising (against the advice of his trusty manual).  The advertising role forces him to come up with “good ideas” and his easy career advancement comes to a screeching halt.

The original book for “How to Succeed” was written in 1952, and senior marketing jobs have only gotten harder in the past 60 years.  Public companies change their CMOs every 18-24 months on average, making it the shortest job tenure of any C-level position.  Over the next decade the CMO will become responsible for more technology decisions than almost any other group in the enterprise, from marketing software platforms, to housing customer data, consumer privacy, and the endless breadth of digital point solutions.

The Assisted Future

In the past few months we have seen five or six start-ups going after the virtual assistant category.  The idea is kind of like Apple’s Siri, but super charged with people who do stuff for you on the other end of the line.  This is a future where we all use on-demand assistants to help us achieve Tim Ferris’ illusory “Four Hour Workweek”.

After digging into the market I have to say I am left scratching my head.  There are some exceptions, but the typical customer of this service, by definition, does not already have an assistant.  Most early adopters fall into one of two camps:  A.) they need a full-time assistant but can’t yet afford one, or B.) they have a couple tasks here and there that require assistance but not enough for a full time assistant.  It turns out that both of these users are hard to scale.  The price sensitive consumer jumps from service to service, and likely churns after a few months.  The less needy consumer struggles to find continued use of the service, and he or she also churns regularly.  There are some fixes to help bridge the gap for both users, but it is a tough problem to solve.

It dawned on me that there is another idea out there that nobody is focused on (at least not anyone I have met).  Rather than try to convert people who don’t currently have assistants, how about developing software for people who do have assistants, but want to empower them to do more?

There are millions of professionals in the US with full time assistants – virtually every lawyer, banker, doctor, and business executive has one.  That is an easily quantified market.  And the ROI seems pretty clear – just by avoiding last minute airfare hikes a business traveler would save thousands of dollars a year.  And how much could you save on hotel rooms, rental cars, or airline change fees?  For professionals who charge by the hour, how much more money could be made by predicting when a customer or a patient is going to cancel their appointment?  All it takes it a little software.

On the other side, assistants need help answering email, logging contacts in multiple CRM systems, and even remembering hundreds of personal preferences.  The tool could learn from experience, automatically completing tasks and suggesting the right answers.  If you are having lunch with Jim, who is vegan and lives in the West Village, the system knows where to go.  If you are flying across country with multiple stops, the system knows how to book airfare based on weather delays, destination changes, and even the potential for cancelled meetings depending who you are visiting.

Maybe this is a pipe dream, but I know that Microsoft Outlook is not going to get there anytime soon.  This is the last mile of productivity, and it seems like a big opportunity to me.

Is LinkedIn the New Facebook?

facebook-linkedinWhen was the last time you added a new friend on Facebook?

I still go to Facebook every day to browse links and baby pictures, but like many people I know, I am not adding many new friends.  The problem is that my personal information on the site has built up over the past few years, and I have second thoughts now about sharing things like wedding photos with anyone outside my inner circle.

This is the paradox of Facebook.  As time goes on I have become more vested in the platform, which is what Facebook wants to happen.  I add photos of friends and family, and I comment on my friend’s stories.  But the flip side is that I find myself second guessing who I want to share this information with.  As of this morning, I am sharing wedding photos with 1,059 friends.  I think that is probably enough.  And yes, I know there are Facebook privacy settings to solve this problem, but it is a pain to go back and edit six years of content.

When was the last time you added a new connection on LinkedIn?

As of this morning I have 2,434 connections on LinkedIn.  I add a few new connections every week.  Even my friend Michael Lazerow, who is the most active Facebook user I know (and built a business on Facebook), has more connections on LinkedIn than Facebook.

This sample size of two may not be statistically significant, but the growth across the network is undeniable.

Does this mean that LinkedIn will topple Facebook for social media supremacy?  Maybe not.  The type of personal media on Facebook creates more page views and interaction than the professional sort of media on LinkedIn.  In a world where attention is traded for advertising dollars, that edge goes to Facebook.  But the two sites continue to evolve in different ways, and I can’t help but think that the number of individual connections will be a key component of value creation down the road.  At the very least, I believe it will help LinkedIn grow faster than Facebook for the foreseeable future.

“Cloud” to “Crowd”

Crowd ComputingFor the past decade we have all benefited from cloud computing.  You may not realize it, but most Internet companies use some component of the cloud behind the scenes.  Cloud computing cuts down on the cost of launching applications, and as a result we get free services like Instagram, Facebook, Pinterest, and Dropbox that come with infinite storage and more frequent updates than they otherwise would.

At the moment we are entering a new innovation cycle, where the labor market is taking on characteristics of the technology market, with scalable, on-demand labor resources.  This trend is early and doesn’t have its own vocabulary yet, but at Greycroft we call it crowd computing.  There is a big efficiency element to crowd computing, but more importantly it has the potential to create new services.

The NY Times touched on the concept of crowd computing earlier this week, with coverage on how companies are turning to human intervention to improve applications and make sense of mountains of data.  Most companies have teams (ranging from one person to 100,000) who spend all day in front of a terminal manipulating data.  This covers every industry, from healthcare to financial services to consumer packaged goods.  And it is equally common for both large and small businesses.  Even at Greycroft we have two teams of people who spend all day entering data, one focused on portfolio company financial information and the other on pipeline information for our new investments.

If you want to build one of these teams, you either open a satellite office and hire workers, managers, and quality assurance professionals or you go to a business process outsourcing (BPO) provider that has already done some of the legwork.  Either way it is time consuming, inflexible, expensive, and hard.

In 2011 we seed funded a company to fix this problem, called Crowd Computing Systems (CCS).  They offer two different solutions – one is a “public” crowd and another is a “private” crowd.  For people who are familiar with cloud computing this distinction is self-explanatory, but the general idea is that a public crowd is a shared resource that many companies use at the same time, while a private crowd is a dedicated team that only performs work for one company.

CCS’s technology has three components. First, they have a reputation system that scores workers on error rates, availability, throughput, and a number of other factors.  This determines a crowd’s ability to process tasks.  Second, they have tools to break complex processes into simple tasks that can be easily performed in a few seconds.  This is akin to a production line in manufacturing, but for data and complex business processes.  The third and final component is an artificial intelligence system that learns over time from observation.  This allows companies to focus human effort on high skill tasks while pushing redundant work to machines.

While this is still early days for CCS, we recently closed a larger investment to scale the business and they are off to a good start with a handful of Fortune 500 firms including Thomson Reuters, Walmart, Amazon, and Coca-Cola.  I am still waiting on case studies to share but the early data is exciting.  Overall I expect this space will be one of the best areas to watch over the next decade, as well as create major economic change for many companies

The Future of Human Resources

In the last few months we have seen a number of companies attempting to transform the Human Resources space.  The big opportunity exists with large companies that are straining under the weight of hiring and on-boarding so many new employees.  For instance, a company like Accenture hires seventy thousand people per year.  In the process they screen a million candidates across a global network of recruiters, agents, job boards, software, and services.  The budgets are large and the systems are antiquated. There is also a nagging problem with mobile – job seekers research positions on their phone today, and the current applicant tracking systems (Taleo, Kenexa, etc) were built before the advance of the mobile web.

Once a candidate arrives for work he or she represents the company, which opens a whole new series of questions.  Does the employee describe their company in a flattering way on LinkedIn?  Is the company’s name even spelled correctly?  Is the employee connected to the right people?  The Internet has already democratized “what you know”, and it is only a matter of time before it democratizes “who you know”.

The Future of Video

longtailI am very proud to announce Greycroft’s recent funding of Longtail Video.  Longtail owns the JW Player, which is the world’s most popular video player, and the company is a key part of our overall thesis in the online video space.

According to Comscore, there were approximately 40bn online video views in the US last month.  13bn of those views were on YouTube.  Hulu, Facebook, VEVO, Yahoo, and the rest of the major media sites accounted for another 4bn or so.  And the remaining 23bn views – 58% of the entire online video audience – were scattered across millions of websites, forming the proverbial long tail.

Scaling Teams

The hardest part of building a start-up is hiring good people.  I realize this may not be intuitive with high unemployment nationally, but there is a shortage of talent when it comes to software development, sales leadership, marketing, web design, and finance positions that are the building blocks of Internet companies.

Companies typically come through my office looking to finance growth, which inevitably means hiring a lot of people, so I deal with this on a daily basis.  A common story is that a company has 10 or 15 employees at the time we invest, with a budget calling for 20 or 30 new employees in the first year.  Looking back I have found that companies almost always fall short of this hiring plan, and it gets worse as time goes on.  As a result I have developed a few rules and tips when it comes to hiring:

2 of 9