The Gong Show

This is the third post in my series on Got It Wrong posts. Posts about how I really missed a particular trend, market development, company, etc…  (Post #1 and #2 if you’re catching up.)

When considering a prospective investment, I’m looking for a variety of characteristics that Fred and Brad have previously written about extensively. One criteria I want to highlight for the purpose of this post is the “taste-maker risk.” Is the service in the business of trying to pick winners that will appeal to users, or is it a platform in which the users self-select and surface the content that interests them?

For example The New York Times is in the business of picking winners in content and as such is subject to the taste-maker risk. The company’s staff has to pick an op-ed board that they think will appeal to their audience. The editors have to decide what makes the front page and what should get buried. This editorial taste-maker risk is hard to due diligence, and is subject to changing rapidly as key personnel leave or as a fickle audience changes their tastes.

By contrast, a site like Craigslist has no taste-maker risk.  They don’t have an editorial staff that determines what posts are good or bad and which ones are in violation of their ToS.  Instead, they setup light governance and a flagging system and then let their users decide what is best-of-craigslist, what is ham, and what is spam.  Most UGC sites in general avoid the taste-maker risk because the best content in the system will rise to the top through the social architecture of the service, and you as the site owner can sample the data exhaust your users are leaving behind to aid in the curation process.

The taste-maker risk is one risk that has been hard for me to ignore or overcome, and as a result, I have been pessimistic early on about the possibilities of some fantastically successful companies.  For example when I first joined USV in ‘06, Softbank and Greycroft closed a $5M round in The Huffington Post.  I didn’t understand the investment thesis in that deal because I was distracted by the taste-maker risk.  But, in retrospect the growth has been remarkable.

I think part of the success of The Huffington Post is the fact that they address the taste-maker risk using a very analytical and quantitative methodology. Jonah Peretti presented at the Tech Meetup a year ago the technology back-end that the editors use to iterate on each story in real-time, and I was thoroughly impressed. For example, there is an overlay view of the site that editors can use to determine which headlines are attracting clicks and which headlines are duds, and then change them in real-time. I’m sure this is a methodology that seems obvious in retrospect, but at the time of Jonah’s presentation, there were audible gasps of amazement across the room.

So, in hindsight, I think I have a blind spot when it comes to first-party content and editorial choices in web services.  The taste-maker risk is a risk, but it’s not nearly as important as I thought it was, and additionally, it’s a risk that smart technologists can navigate well.

Just to drive this point home: one company in the USV portfolio has always been subject to the taste-maker risk (and is still today). It’s Zynga. All social gaming companies that build their own games are subject to this risk. Fred in the comments on his blog awhile back that Zynga is one of the best investments he ever made.

This is the second post in my series on Got It Wrong posts. Posts about how I really missed a particular trend, market development, company, etc…

When I interviewed for my gig at Union Square Ventures back in ‘06, part of the process was a 45-minute interview with Fred and Brad.  We talked about 3 different subjects: user experience, product management, and mobile browsers.

If you think about the timing… the state of the art in mobile phone browsing was either the Blazer Browser on the treo or pocket IE or the RIM browser on the BB 7250… none of which were setting the world on fire. The best way to browse on a handheld device at the time was the Opera add-on browser to the Nintendo DS, which had just be launched in Japan 3 weeks prior to my interview; the the initial reviews of Opera on NDS were pretty mixed.

So, Fred asked me, “Do you ever think we’ll have a mobile browsing experience comparable to the experience on a laptop in the next 5 years?” And I said, “No.”

I didn’t think we would get there because we were all carrying around mobile computers that had more processing power and memory than the computers we all first used to browse web pages in the 90s.  So, it wasn’t an issue of technology… and even if it was a technology issue, Moore’s Law would fix that in a year or two.

So, if we didn’t already have great mobile browsers, then what we holding us back? Clearly not secondary technology progress.

I concluded that the form factor of handheld devices was too small to deliver a usable web browsing experience that could ever rival the experience of the web on a laptop. I tried to imagine A) reformating pages on the fly for mobile consumption B) designing dedicated mobile web pages or C) zooming in and out of real web pages on a mobile phone… but none of these options seemed reasonable to make more than a tiny portion of the web useful on a phone.

Then, a year later, the iPhone was released, and I ate crow. I think my original analysis did properly account for secondary technologies, but I failed to account for Steve Jobs.

I’m going to start a new series of posts on this blog. All my favorite VCs in the industry are generally pretty humble and intellectually honest. In that spirit I’m going to recall a few occasions over the past few years working in VC, where I was quite wrong (and, hopefully, what I learned since then).

Today’s “Got It Wrong” post is on privacy.  When I first joined Union Square Ventures I was adament about protecting user privacy.  I was a user of tools like AdBlock Plus (to protect both privacy and attention), and I cleared my cookies with reasonable regularity.  I was picky about the companies with whom I trusted to handle my data. I ran system-wide searches for my own credit card info, lest it be stored in an autofill field.

All of those behaviors are subject to the eye of the beholder… to some they’re good hygiene and to others they’re over the top.

But, where I “Got It Wrong” regarding privacy was assuming that other users felt similarly to me. I didn’t think users were as proactive as I was about privacy hygiene, but I did think they had similar instincts. Watching companies like Mint bloom with a mainstream audience was informative. You have to enter your username and password for your credit card just to get started, talk about a high barrier to entry, and yet lots of people did so readily.

Also the growth of companies like Loopt (and later foursquare) and even Twitter was really interesting. Unlike with Facebook, in those services, egocasting was the default mode of usage and default to public. An older version of me would have been reticent to embrace Twitter when it first came out, but after a few months of diving into 5 new web services a day in deal-sourcing, I was already starting to evolve my stance on privacy and thus was quick to dive in.

I won’t go so far to say that users don’t care about privacy at all.  The success of companies like Lifelock are evidence of the fact that users are concerned about the security of their personal information. But, it’s remarkable what data people will willingly hand over when asked nicely, and it seems to me that privacy concerns are rarely the friction to adoption I expect they would be.

I’ve been listening to more electronic music lately and like what electronic acts are doing in a live setting.

Since large portions of electronic music are tweaked to perfection in post-production, its often difficult (or even impossible) to reproduce live. So, some live electronic acts are often accused of “just hitting play.”

One of the first live electronic acts I saw in NYC was Girl Talk.  The first thing Gregg Gillis said when he came out on stage was “Have you ever seen someone play a laptop before?  I think you all got ripped off.”  He’s poking fun at the common criticism right away and setting expectations according.  Smart showmanship.

But, the best electronic music live can deviate so far from the original work that only the anthems are recognizable, and the rest is a digital collage carrying between the major chorsus of multiple songs.  For example, Daft Punk’s Alive live album pulls off this megamix effect really well.  It’s everything you recognize yet it all sounds completely new.

Because of the real-time reproduction issues, electronic music is pushing what it means to perform live. The digital guys always have to push the presentation extra hard to win over audience.  For example, when I saw Diplo live, it gave me a whole new appreciation for what he does… he was mixing between wax he cut himself with his own remixes, and then had the whole act synced up to a big video projection he mixed himself of old Nintendo gameplay.  He tied in the theme from all the 8-bit classics seamlessly, and then washed the whole experience in a acidic warp that added a unique layer of his own.

If anyone has recommendations for other great live electronic acts, let me know.  My favorites so far have been the ones mentioned here and also M83, Hot Chip, and The Field.

It seems like Google is on a mission to make Android users feel bad about their purchases.

Every month, literally every single month, there’s a new flavor of Android phone that’s better than the previous one and makes the previous month’s Android phone look obsolete.

November: Motorola Droid

December: HTC Hero

January: Google Nexus One

February: HTC Desire

And now coming in Q2: the HTC Legend

How can you purchase an Android Phone (and likely lock yourself into a 2-yr contract in the process) with any confidence given the pace at which they release new phones?

I switched to an iPhone the first week of the 3GS release (back in June ‘09), and I’m still running the most current model available today because the iPhone is on an annual release cycle.  I’m sure in a couple months I’ll be blown away by the latest and greatest model by Apple, but at least I will have had one year of naive bliss in the interim.

My complaining in this post is of course a bit of a mock… I love seeing how rapidly Android is iterating, and I think it’s a real product development strength of theirs. But, it feels like a marketing mess.

The irony here is that being quickly eclipsed by the next generation of incompatible product development was Apple’s entire marketing strategy in the mid-90’s.  It felt like there was a new Macintosh Quadra released every 3 months or so… you could never feel comfortable with a Mac purchase for long.  Apple’s product marketing strategy has come a long way since those fumbles.

There’s a variety of third-party development efforts to innovate in the mobile browser market, such as: Skyfire, Fennec, and Opera Mobile.

Given how tightly coupled the browser is to the OS in the iPhone, Palm Pre, and Android, I don’t see how these mobile browsers are going to gain meaningful traction in the future of smartphones. The native, first-party browsers in all these OSes are too “default”… too baked into the interactions with other apps.

When the OS layer was so tightly coupled to the Browser in the desktop-era, the Justice Department had to step in and take action against Microsoft.

Yet, these days, it doesn’t seems like consumers and competitive companies in the mobile browser space are crying “foul” nearly as loud as during the Netscape vs IE drama. So, what’s the difference this time around?

For one, I think consumers these days actively seek out tightly-coupled, complete solutions.  Apple is praised for their unified product offerings with a comprehensive and consistent user experience across all apps. This is only possible when multiple layers of the software stack are integrated with each other and 3rd-party apps are limited in the damage they can do (such as no background processes).

Additionally, I think the open source nature of browsers make them feel more “open” even if they are being tightly coupled to a mobile OS.  The idea that a geek can go inspect the WebKit source code makes the process of deeply integrating browsers into the OS seem less evil.

But, like Google demonstrated with the launch of Chrome, there’s still much innovation left to do in the Browser layer of the stack, and I hope that in future Mobile OS releases it will be easier to swap in and out different browsers.

Why does content get more traction in some social media sites than others?

Since Buzz launched publicly last week I have been using it to syndicating my Tumblr posts.  I’ve noticed that sometimes I get more (and better) comments in the reply threads under my syndicated posts in Buzz than I get in my own comment threads or at Tumblr.

For example, when I responded to Dave Lifson about the “innovation premium” in big companies acquiring startups (specifically, Google acquiring Aardvark), I got the following interaction around my post:

• Tumblr: 2 Likes in Tumblr, and 2 reblogs.
• Disqus: 0 Comments.
• Buzz: 12 Comments and 3 Likes.

So, the quantity of activity of this post was highest at Buzz. And additionally, Greg Cohn jumped into the Buzz thread and left a really insightful comment. So, I’d say the quality was also higher there too.

But, that’s just one example, and there are plenty of counter-examples. I wrote another post last week about the relationship between Buzz and Friendfeed. The quantity of interaction for that post was:

• Tumblr: 14 Likes and 5 Reblogs.
• Disqus: 1 Comment.
• Buzz: 0 Comments.

So, pretty much all the activity was at Tumblr in this case.

I can’t figure out why interaction around content takes off in certain places, but not others.  The only pattern I can detect is that if a conversation sparks, the conversation generally stays in the channel in which it is sparked.  In other words, if two people start replying back-and-forth about a post of mine in Tumblr reblogs, that conversation rarely jumps over to my Disqus thread, and vice versa.

But, the key question (and one that someone could build a business around) is why does interaction get sparked in the first place, and what content is best suited to spark is specific channels over other channels?

I noticed a simple trend in social services that is obvious in retrospect, but often not well articulated.

In Twitter, I have 2089 total lifetime Tweets and 2367 Followers.  In Tumblr I have 268 total posts and 262 Followers.  In both cases, this is essentially a 1:1 ratio of activity:followers.

Now, this is just a correlation, and does not necessarily imply causation in either direction. But, there is definitely a relationship between organic interaction with a service and the followers you acquire in the service.

The key to this relationship is that the activity has to be genuine and organic.  For example if I just start spamming my Twitter account with the monotonous details of every sandwich I eat, I’ll quickly dilute my activity/followers ratio.

What’s your activity:follow ratio for various services?

Seth Goldstein defined the term iDroid, and in doing so, articulates a big concern of mine as I observe mobile product development in USV’s portfolio companies.  First the definition:

iDroid (n.) - developing for iPhone and Android platforms simultaneously, as in “let’s make sure we have iDroid versions up and running before we launch the new web service.”

I’m concerned that mobile development on multiple platforms is the new browser compatibility battle of this generation.  I’m certainly not the first person to have this concern. But, instead of complaining, I want to know what developers are doing for a solution?

Is everyone just maintaining separate presentation layers for their iPhone and Android sites?  Or, is there a common subset of HTML5 and Javascript calls that developers can use safely for cross-browser support? Or, best of all, is there an abstraction layer (like the ActiveRecord Pattern equivalent for Mobile UI) that developers can use to make sure that the visual layer looks good on all mobile broswers?

FriendFeed, the real-time web aggregation startup founded by Paul Buchheit, was acquired by Facebook in August 2009. The timing of the acquisition is suddenly very interesting to me.

A number of Google employees, including Sergey himself, have said that Google Buzz has been live and actively used internally inside Google for the past 6 months, which means it launched in August 2009, the same month as the FriendFeed acquisition.

Paul Buccheit, as the founder of Gmail and friend of many current Google employees, likely knew about Google Buzz when it launched internally, despite the fact that he was not a Google employee then.  Many bloggers see very strong parallels between Google Buzz and Friendfeed… Best of all, here’s Paul’s own response when Buzz when public yesterday, “This seems vaguely familiar…”:

So, did Paul sell FriendFeed at least in part because he knew Google Buzz was coming soon? Paul might have thought that Buzz would be directly competitive with FriendFeed, but with better placement in users existing workflow, and so August was the best time to sell.

Or, it could just be a coincidence.

Many people seem underwhelmed by the arrival of Google Buzz.  I think SAI’s headline sums up the overall sentiment of the blogosphere nicely: “It’s Late, Boring, and Lame.”

However, I think this kneejerk reaction, especially since it comes from many bloggers that can’t even log into the service yet, underestimates one VERY important feature of Google Buzz.  It’s the old Brick-and-Mortar business adage: “Location, Location, Location.” See related pic:

Google Buzz is baked directly in Gmail.  I suspect many GChat users don’t know this, but GChat was originally a completely independent product from GMail.  GChat was a stand-alone client, just like AIM or Y!Messenger.  It also received a somewhat underwhelming welcome from the market, but once it was baked directly into GMail, and logged you in as soon as you opened your email, it became a killer feature.  The only difference between the original incarnation of GChat and the GChat of today is the location of interaction, being baked directly in GMail matters.

So, while I’d love to write-off Google Buzz as DOA like most bloggers, I’ll be curious to see how being baked into the Gmail interface will affect my own adoption of the product, along with the rest of Gmail users.

VCs have a wide variety of styles, and I don’t know if all of them are obvious to entrepreneurs or not, so I’ll enumerate some of the differences in style here.

When a VC firm goes out and raises a fund, one of the docs that is a minimum requirement is a Private Placement Memorandum (PPM). The PPM is what you distributed to potential Limited Partners to explain to them (amongst other things) your flavor of fund. One element of the PPM is always a description of the fund’s focus.  How could this differ?  Well…

1. Geographic: Some VCs only make investments in a specific geography. For example, the DFJ Network of funds are a group of VC Firms, many of which invest specifically within a defined geography.
2. Stage: Some VCs limit their investments to a specific stage. Certain funds will only invest in company with annual revenue greater than $1MM. Other funds, will only invest in seed-early stage opportunities. For example, Accel offers a growth-fund and an early stage fund separately, which differ primarily in stage.
3. Sector: Some funds will invest only in specific industries: such as Information Technology, Biotech, Cleantech, etc… Other funds are sector-agnostic, and use partner specializations to focus on specific sectors within the same fund.
4. Opportunistic: Some funds deliberately market themselves in their PPM as leveraging a proprietary network of deal flow, from which the investment team can be opportunistic. For example, LPs who invested in the Founders Fund likely gave good thought to the appeal of the PayPal Mafia’s network and proprietary deal flow.
5. Quantitative: Some stylistic choices can be very specifically quantified by common financial metrics.  For example, I talked to an investor a few years back that was taking a deep dive into the Ad Network sector, and he said, “we won’t invest in any Ad Networks that can’t reliably produce 40% gross margins.”
6. Thesis-Driven: Some funds will apply a intellectual framework or thesis to all their prospective investments. A thesis is typically used as a filter; a great team generating high-margin revenue in an attractive market might not meet an investor’s investment criteria if the company is not a fit for the investor’s thesis regarding overall market trends. Here’s an example of a thesis we commonly apply at Union Square Ventures.
7. Other LP-related Limitations: Some funds can’t invest in certain industries, such as gambling or alcohol-related revenue because of restrictions in their agreements with their limited partners.  This is most common with Sovereign money limited partners, such as Saudi or Chinese sovereign wealth LPs.

The takeaway here if you’re an entrepreneur is you can use this list as a set of questions to ask a prospective VC to see if they are a good fit with your company.  Sometimes you don’t even need to ask a VC… their current and past portfolio will speak for itself about what they actually do. Their actions define their style.

Building web service startups is not a zero-sum game.

There isn’t a fixed amount of resources: bandwidth, servers, HR, ad-inventory.  And, there isn’t a fix amount of demand: more people are coming online everyday and the usage of existing internet users is increasing. More money is being spent online every year. It’s all increasing and some metrics are even accelerating. Everyone’s payoffs in this “game” are different… what is a valuable user or interaction to one startup is trivial (or lead-gen) for another.

And yet, so often in the startup scene, the news and the gossip about companies is how they are all brutally competing, as if there can only be one winner.  Almost any article on competition between two companies is titled something along the lines of  “X feature is a Y-Killer”… where X is some tiny product design of a non-revenue generating startup and Y is a Fortune 500 company. I don’t fault bloggers for sensationalism (I’m guilty of it too), but I wish their posts didn’t repeatedly come baked with this zero-sum game mentality.

I know that the market for web services feels very competitive right now, but it’s important to remember that it is definitely not a zero-sum game. There is a set of services for which the market would feel more zero-sum: me-too companies which David Shen wrote extensively and lucidly about. Try to avoid fighting over the exact same slice of the pie as your neighbor… instead, work to build a bigger pie together.

And, if you ever feel like your competing over a fixed amount of user attention, make sure you’re looking at a long enough time horizon.  I’ll close this post with a nice 10-year chart from the Pew Research Center that can help ground your sense of prospective. The pie is getting bigger everyday.

It appears as though Google SocialCircle was first announced in late October, but I’m just using it now for the first time.  This page (if you’re logged into Google) should contain your entire social graph along with all the services that everyone in your graph uses. And all this data is now being used in search results to source relevant links written/recommended by your social graph.

This feels quite big to me.  First of all, I’m impressed by the accuracy of the social graph data… how did Google know that THIS Andrew Parker (aka Me) and not some other Andrew Parker is the one that is Tweeting at my Twitter account and blogging at Tumblr? And, they managed to tie it back to my Google account… and they did this magic trick for everyone in my SocialCircle.

This level of disambiguation is a hard problem to solve, and while I’m sure they didn’t totally crack this nut, it seems like a remarkable feat for an initial launch.

I bumped into this SocialCircle feature because links from my SocialCircle are now being included in my search results.  I was searching for some information on syncing calendars, and at the bottom of the SERP, Eric Friedman’s face popped up next to a post he wrote on the subject. Here’s a screenshot.*

Eric’s post wasn’t a perfect match for my search, but it definitely caught my attention because I will always trust anything that comes from my network more than just a random page.

Some implications of this new feature I predict:

• This is a death knell for the *Single Definitive SERP*. Google is going farther and farther towards customer tailoring all search results based on individuals’ data, and SocialCircle is a big step in that inevitable march towards completely personalized SERPs. As a result, the effect on the SEO market will be large.
• Businesses’ FB and Twitter accounts are going to start mattering.  A LOT.  If your business is linked tightly into the social graph, you will appear more often on SERPs.  That’s free share of voice and free traffic.  Expect a deluge of friend requests from marketers and other entities that are currently spending money on SEO and SEM.
• No more scrolling Tweets from random Twitter users in my SERPs.  Instead, all social media sources will be heavily weighted by my social graph.  If I’m interested in the ski conditions in Tahoe this weekend, I will only see Twitter-sourced results from people within 3 degrees of social graph connection to me.  Anything further out than 3 degrees is assumed to be noise because the source is too untrusted for social media.

I look forward to seeing what APIs Google opens up around this SocialCircle feature… this is ripe for mashups.

* Hat tip to Eric for giving me a tour of the ins-and-outs of this new feature Thursday morning.

Alley Insider reports that Google’s dialing up their M&A activity again. There are plenty of startups that have failed to find a sustainable business model over the past couple years of recession, and for their sake, I’m glad to hear that the M&A market is warming up again.  HR-acquisitions can be a great result for a startup that can’t otherwise remain independent.

But, I can’t help but wonder what “could have been” for many of the companies Google acquired in the past.

For example, what if Dodgeball had not sold to Google… could it today be the equivalent of foursquare with 2 additional years of growth and product development?

Or how about AdMob? The leading mobile ad-network is now owned by a company that also owns a mobile operating system… it feels like a case of the fox guarding the hen house to me… what could have AdMob have become as mobile usage explodes going forward if they had remained independent?

Here’s a complete list of Google acquisitions throughout the life of the company, according to Wikipedia. There’s a lot of value for Google in that list, and there are a number of key strategic products that could not have been built without an acquisition. For example, Google tried to compete with YouTube by building out Google Video, but ultimately had to buy YouTube in order to be a dominant player in the online video market.

So, Google ramping M&A activity back up is a double-edged sword. It’s a great source of talent and new products, and it’s a good way for businesses to end that could not otherwise be independent.  But, I don’t see increased Google M&A activity as a cause for celebration, and it could ultimately distract and impede some of the brightest opportunities in internet innovation.

[As a disclosure, Google acquired FeedBurner, a former Union Square Ventures portfolio company.]