Tuesday, December 1, 2009

Geo-Location APIs: What Are The Differences?

One topic from the RealTime CrunchUp hosted by TechCrunch (on which I was a panelist representing the Mixer Labs/TownMe GeoAPI.com) was the various location-focused APIs that have launched recently. Given the large spate of recent "Location API" announcements and roll-outs, and how different these APIs all are from each other, I thought it would be useful to break out the major components of APIs that have geo-information and to compare their uses. This helps clarify which API to use when, as well as what APIs are live versus longer term roadmaps.

Goals of a Location API

The purpose of a fully formed Location API is to allow developers to build a robust set of location services using the API as the data source and data store. Many recent Location API announcements are really applications releasing incomplete databases of crowd sourced business listings generated by their users rather than the broader offerings defined below:

The 5 Core Components of Location Based APIs

Much of the confusion around Location APIs is that although all of the APIs that have launched recently have different aspects of location involved, they all differ pretty dramatically in the feature set offered and the types of information that can be accessed for each.

The core components are:

  1. Forward Geocoder
    • What it is: Take an address or business name and returns the lat/lon coordinates associated with it. E.g. "46 West 3rd St, NY NY" converted to 32.343, -123.454".
  2. Reverse Geocoder
    • What it is: Take a coordinate lat/lon and converts it to an intersection, neighborhood city, state, country. E.g. "show me which neighborhood contains the point '37.7810, -122.4050'" and get back "SOMA, San Francisco"
  3. Database of Places.
    • What it is: A database of locations, business listings, and points of interest.
  4. Writable Layers
    • What it is: Ability for developer to store information about places on the API itself in their own private layer, and then do queries against it. E.g. "show me all the places within half a mile that this user has checked-in on my app". This allows for client-side development of apps with no back-end servers built out for the Geo-components of their app.
  5. Media Layers
    • What it is: Let the users query other sources of geotagged information via the API. E.g. "Show me all the Flickr photos within the boundaries (i.e. polygon) of Dolores Park". Basically, the API provides the geo-query/geo-search infrastructure by which complex queries can be made against other media sources such as Twitter, Flickr, etc.

Additional Non-Core Components

In addition to the 5 core Location API components listed above, there are two additional items provided by companies which they call “Location APIs” but which don’t really enable developers to build fully robust geo services in their own right:

  1. Monetization
    • What it is: Ways for the app developer to make money via the API. Usually this means syndication of ads or offers to third party developers using the API with a 3 way rev share between ads provider, the API provider and the app developer.
  2. Company specific or app specific data
    • What it is: Some app developers expose their own user data or user generated data via an API. This is often meant more as an add on for people to develop apps specifically for their service, rather than as stand-alone apps unrelated to their service which can use geo data. These APIs will often have an incomplete database of places annotated with their users’ data (e.g. tips from FourSquare users). Over time, services such as FourSquare may end up with a very robust database of places.

Things That Don't Matter:

  • "Enabled for RealTime". This should be a given. It is sort of like saying "The car I just sold you comes equipped with tires".

Who Has What?:

I have tried to compile a view of what is available with each service as mapped against the 7 components above. A check mark indicates it is fully live, a dash indicates there is some data but it is not yet comprehensive.

Disclaimer: for some of these services information is scattered or not always easy to find. If there are errors in e.g. the table below, please email me and I will correct it.

(Aside: SimpleGeo, another location API, is not covered as the product is currently in “private beta”. No documentation, or demo keys are live. and no developers I have spoken with have been given a key. So it is unclear which parts of the service are live versus forward looking roadmap.)

The Takeways:

  • There are 5 major components to a fully featured Location API - forward geocoding, reverse geocoding, POI database, writable layer, and media layers.
  • Some companies who have recently launched "Location APIs" are basically providing information to the data their users have crowdsourced, rather than comprehensive database or tools for developers to generate location services.
  • For "street address to lat/lon" forward geo-coding or International coverage, the Google API is currently your best bet. GeoAPI.com will be adding international coverage shortly as well.
  • For Twitter focused apps, tying directly into the Twitter API to get basic geo-information is the easiest approach as it is part of the API call. The Twitter API only applies to Twitter data (i.e. it will tell you a Tweets lat/lon coordinates), so this is not a broadbased service for other developers to use. Twitter suggests you use Yahoo! Where on Earth for basic reverse geocoding. So if you want to do something more complicated (e.g. build a FourSquare style app on top of Twitter) you will need to use GeoAPI.com to access a more comprehensive set of services.
  • Apps such as FourSquare and NextStop do not currently have a comprehensive database of locations. However, given their momentum this may only be a matter of time.
  • GeoAPI.com provides comprehensive coverage in all areas of the core GeoAPI components except for forward geocoding (e.g. POI database, writable layer, media layers, reverse geocoder). Simple forward geocoding (built in a way to make overall use of the GeoAPI easier) will come shortly.

    Questions? Comments? Corrections? Please DM me on Twitter @eladgil

Wednesday, October 28, 2009

The 3 Types of Embedded Social Networks That Drive User Generated Content

ALL USER GENERATED CONTENT SITES HAVE AN EMBEDDED SOCIAL NETWORK
One thing that people often miss about user generated content (UGC) sites is that the vast majority of them have an embedded social network that drives all the content generation on the site. These embedded social networks have the following characteristics:

  • Only a fraction of users coming to the site (e.g. 1% or less) participate in this social network. (E.g. a few thousand people created the bulk of early wikipedia even when the site had millions of users streaming through every month.)
  • The features for the social network differs from the features the casual user of the site sees. (E.g. Users hitting Yelp via SEO usually don't notice that the reviewer is a "Yelp Elite" if they just want to know hours of operation).
  • Each type of UGC site has a social network with different characteristics (badges, friending or not, leaderboards, etc.) depending on the type of site it is
  • The users in the social network have very different motivations and drivers then the average user on the site. (Don't assume you can convert a random user from SEO traffic into a content generator - you need to find people who have more free time on their hands then average)
THE THREE CLASSES OF EMBEDDED SOCIAL NETWORKS
I have tried to map some of the basic social gaming mechanics against types of UGC sites.

Takeaways:
a) Some things you would assume are important for social dynamics are not. E.g. "friending" (the ability to add other users as friends) is not needed for Advice/Q&A type sites. People are more motivated by reputation and public knowledge display then by who their friends are.

b) There are clear patterns of features associated with each type of site (e.g. Q&A site vs wiki). It is unclear if this is due to sites copying features from one another, or whether there is a single canonical set of features needed to prime the right social mechanics for a type of site.

c) Wikis have had advanced social mechanics for a long time. Sites like Stack Overflow or FourSquare are heralded for their gaming mechanics (and I like both products quite a bit), but wikis are really the unsung pioneers/heros of this stuff.


Ping me if you have other type of UGC sites with embedded social networks I dont have here. Let me know what you think!

Oh, and you can follow me on Twitter here :)

Tuesday, August 4, 2009

We are hiring - please send engineering resumes! :)

We are hiring 1-2 generalist engineers. http://www.townme.com/jobs.html

Please send resumes to jobs@mixerlabs.com.

Basic description:
-Generalist engineer who can work on a core piece of infrastructure one day and front end the next day.
-We do everything in python/django.
-Nice, smart, "good peoples".

Tuesday, June 30, 2009

The VC Cleanse - Taking Advantage of the Downturn to Walk Away


VCs are cleaning out their portfolios right now. It seems like there are 3 drivers and 3 methods to doing so. Implications are discussed.


WHY ARE VCS CLEANING OUT THEIR PORTFOLIOS?
  1. Some companies should get the plug pulled on them. Silicon Valley is built off of a fail fast mentality and the constant birth and death of startups. There are some investments that have underperformed, misexecuted, or been hit hard by the current environment and can not raise cash and are getting shut down.
  2. The downturn is an opportunity to write off companies with a long time horizon and large capital needs. Some VCs are using the recession and removal of companies that "should die" to also force exits or write offs of investments that may do OK longer term, but are either (a) not on a ridiculous growth curve or (b) require large amounts of additional capital to keep going. I think the thinking is: (i) we can free up partner time to invest in all the high growth startups on a ridiculous growth curve that we can get into cheaply today - so why back a company with only so-so prospects for the forseeable future? and (ii) with the broader cleanse of our portfolio, we can use the recession as "air cover" with our LPs for shutting down or getting out of this company, so it is an easy exit for us and frees up time as per (i) so we can find something that is more of a "sure bet".
  3. They need to return their funds and please LPs, so they will force exits. The last 2 weeks have seen a slate of exits. Exits are driven by (a) inability by the company to raise capital (so an exit is a better alternative) (b) VCs want to show strength or return capital to LPs so they may force exists (especially if they are raising money or negotiating terms with their LPs for their fund) (c) larger companies may be opportunistics about low startup valuations (especially if their stock is doing well) and buy out startups who think a safe home may be a better alternative then the extra hard work of navigating a recession or (d) traditional M&A logic - sum greater then the parts, the acquirier is giving a good offer, etc. etc.
HOW ARE VCS EXITING?
  1. Shutting down / liquidating a company.
  2. M&A - selling (often) at a lower price then the last round
  3. Wiping the cap table. Having a company go through a down round without the original VC participating, allowing the VC to leave the board although with loss of any future ownership in the company.
SOME RECENT EXAMPLES:
(I have zero inside knowledge of any of these, so my classification of these companies is PURELY 100% speculative and in some cases baded on rumors or e.g. TechCrunch or other blogs):
-Mochi Media (rumors), Zappos: recent press suggested entrepreneurs and investors may have disagreed on exit paths where investors may have been more eager to sell.
-SearchMe: company significantly downsizing/repositioning due to large amount of add-on capital needed.
-iMeem: Cap table wiped, new money raised.
-Xoopit, Zappos (again): Exit into synergistic company with good product fit.

IMPLICATIONS?
  1. A lot of startups are going to shut down in the next 6 months. As VCs clean out their portfolios further, a lot of companies are going to go out of business.
  2. VCs are going to have a lot more time to do new deals in Q1/Q2 2010. With less partner time going to struggling companies, VCs will be able to buy into all the cheap deals they have been seeing lately. This means getting funded for new startups may be easier in early 2010.
  3. There may be cheap assets to buy. Following the last downturn, there were a number of assets that people bought on the cheap and turned in businesses worth 10s of millions or more. For example, eHow was a site that generated a large amount of content during the last bubble which was then sold for peanuts and reborn as a new version of the original site and eventually sold to Demand Media. You can read about it on wikipedia.
So what do you think? What other companies will be affected by the VC psychology above?

Like this? You should follow me on Twitter here or subscribe.

Tuesday, June 23, 2009

8 Digital Media Companies I would Invest In (If I was a VC)

If I was a VC I would be taking advantage of the sparse funding available for later stage deals and depressed valuations to chase down a handful of deals that I think will end up being substantial companies with a reasonable probability of a good exit.

From what I understand, this is basically what IVP did with Twitter, with Benchmark then being pulled into the fray once Twitter was in play for an investment (until IVP pursued Twitter, the Twitter founders kept claiming they didn't need to raise any more money)

Companies I would pursue*:
  • Mint.com. Mint has a great model, a market leading position, and access to tons of interesting and unique data that leads to some pretty direct monetization opportunities. They have solid user growth and traction, and nice repeat use. They will have a great exit. Now may be the last chance to get in reasonably cheap.
  • Coupons Inc. Really fast growth in coupon volume, a patent based monopoly on online coupons, a crappy macro econ environment in which people want more coupons. Oh yeah, and as far as I know they have never taken any VC money.
  • Playdom. Zynga gets all the press, but also probably sports a higher valuation per $ revenue or per user. Playdom is a leading player in social gaming and is on a rapid growth (and monetization) curve. I think the danger is that (a) certain short term aspects of social media monetization (i.e. offers) will soon collapse in value and (b) as a social gaming company they constantly need to be iterating on the next hit. But despite these risks I think the company is on a tear and the traditional gaming companies (e.g. EA) are still totally lost when it comes to social gaming. Playdom just recruited John Pleasant, the COO of EA as their CEO, so perhaps this would make them harder to pick off to fund :)
  • Admob. When I was as intern at Cisco many years ago one of my mentors there said "growth covers up for A LOT of mistakes". The mobile web is now exploding due to iPhone et al, and a rising tide raises all ships. Just as Cisco benefited from the rise of the Internet/IP everywhere, AdMob is well positioned to capitalize on mobile web explosion. Further, there are a number of great potential buyers for the company if the company decides not to build to an IPO (Microsoft, AOL, etc. will all need mobile stories, and some of them really want to build out their ad networks). That said AdMob faces potential strategic challenges as the mobile web shifts dramatically from WAP to iPhone and Apple effectively controls the channel and could potentially partner with Google to compete in ads.
  • Cloudera. Alas - in this case Greylock beat me to it! Greylock was smart to make the investment in Cloudera right now. Cloudera is rapidly gaining traction as an industry leader in a market that will continue to grow due to the masses of data being accumulated everywhere (e.g. bioinformatics, digital media, advertising etc.) and hadoop will be increasingly all pervasive. I think the most likely outcome is an exit to IBM, HP or the like, but now is the time to invest.
  • SMSgupshup. OK, maybe the exit here won't be huge given the India market focus / lower value per user, but I think this company is pretty cool.
  • Fixya. Like many SEO-driven businesses, it took Fixya 18+ months to really ramp up and get traction, but they have recently hit the "virtuous cycle ramp" where users are generating content to attract users who generate more content. Fixya has some interesting monetization opportunities it can capitalize on. Not sure they will ever be a consumer brand (and benefit from the value of a brand) but they could be a great SEO driven business.
  • Twitter. I would only invest in Twitter at this point if I was a VC looking to build my brand as a consumer internet investor. Twitter's internal valuation is probably quite high at this point, but at the same time I think they will have a $1 billion+ exit. So the return on Twitter itself might be sucky (i.e. somewhere in the "government bonds to 2X range"), but from a brand building exercise it might be worth it if I wanted to build a consumer internet/digital media franchise for my venture fund. (I will write a future post on brand building for VCs in digital media later)
Feedback?
So what do you think? What companies would you invest in if you were a VC and why?

Subscribe
If you like this - feel free to subscribe to this blog / RSS. I will be posting about VC brand building shortly.

*Disclaimer: I have zero inside knowledge into these companies and don't have access to their internal metrics. So this is me making guesses best on total stabs in the dark on what I think is interesting medium term.

Monday, June 22, 2009

M&A: How to tell if Google/Microsoft/FB is about to buy your social media startup

Whenever a larger company (e.g. Google, Y!, Microsoft, Facebook, or the like) is trying to buy a smaller company (e.g. Twitter, Facebook 2 years ago, Myspace, etc.) a leading indicator of interest is the fact that the company execs all suddenly sign up for accounts on the social media service they are trying to buy.

E.g. when rumors abounded about multiple companies trying to buy Twitter, I suddenly saw a bunch of execs from a not-to-be-named-here company join Twitter.

Indeed, I am tempted to create a list of email address of top execs at different companies I know, and run this through the email scrapers of various services periodically to see where M&A interest (or cloning of functionality) is about to occur.

If anyone wants to automate this and share results, let me know, could be an interesting view into what is the next hot (or soon to be acquired) company in the valley is.... :)

Monday, March 2, 2009

Twitter is to Facebook as Google was to Yahoo!

My friend David King, (former Googler, founder and CEO of Lil Green Patch (# 1 social game on FB) and one of the smartest people I know) came up with an analogy I thought was really interesting to explore. His basic premise:

Twitter is to Facebook as Google was to Yahoo!

Similarities:
  • Google took a core piece of Yahoo! page views (search) and made it a key, stand alone product and brand. I.e. Yahoo! had pageviews from movies, real estate, etc. and so viewed search as an add on. Turns out search was a key product in its own right. Similarly, Facebook pageviews seem to be driven in large part by photos and apps, with status messages a big chunk of page views but not something that has been focused on as a stand alone extensible product. Enter Twitter - with "extensible status" - which has been establishing itself as a standalone product/brand.
  • Additional key insight David had: The piece of Yahoo's traffic that Google took over, was the most monetizable piece. Similarly, Twitter's extensibility of status messages will be more monetizable then the average Facebook page. Why? Because a subset of Twitter pages show direct intent to purchase something, as well as search on Twitter should be monetizable for a subset of terms.
This ties in to something I have been thinking about the last few weeks, namely...
Twitter will be the winner in social search
I know a number of people who have each launched a social recommendation product in the last few months. They are all really smart people who built nice products. However, I think for quick, real time responses Twitter is emerging as the winner in social search. People send out quick tweets to ask which restaurant to go to, what car their friends like, and what gift to give their significant other.

This real time web of communication tying into a broader use case (of regular, mostly mundane Tweets) is similar to Google vs vertical search engines. People use Google because they can find *anything* on it, so they also enter their commercial queries (which monetize best) into Google as well. People ask their friends for recommendations (which should monetize well) on Twitter because they are constantly broadcasting other stuff (which doesn't monetize so well - such as "I just ate stinky cheese. It made me feel funny".) on Twitter. By establishing a broad use case around real time communication, Twitter has also made itself the potential winner in social recommendations.

Why is Twitter the winner in this and not Facebook?
Facebook still has an opportunity to make a lot of money in this area. However the things that are likely to hold Facebook back are:
  • The need to focus on the broader experience (i.e. how does status tie into everything else?) This prevents it from emerging as a stand alone app with a simple, clearly defined use case.
  • Facebook has a ton of stuff on its plate (monetization, mobile, etc. are probably all sucking up resources). Twitter does one thing and does it really well.
  • Facebook doesn't have real time communication feedback loops built in as tightly as Twitter.

Where the Twitter/Google Facebook/Yahoo analogy breaks down:

Size of outcome (Facebook is adding 5m people a week!)
Google is currently worth>5X Yahoo! by market cap (as of March 02 2009). Facebook is adding 5 million new users a week (i.e. the rough equivalent of a Twitter) and I think will be a very valuable, very large property. I do not think Twitter will be worth 5X more then Facebook in 3-4 years  :)

Technology as key differentiator
One of the reasons people say Google went on to be more valuable then Yahoo! is on its focus technology. I.e. better technology led to lower costs (and higher margins) on the serving side and higher revenue on the ads targeting side (e.g. when to trigger good ads, which ads to show that will monetize best, etc.). In the case of FB vs Twitter, FB is known as the stronger technology company.

To sum up with one last commonlaity in Twitter/Facebook vs Google/Yahoo- both Twitter and Facebook will be enormously successful companies (people seem to forget Yahoo is worth >$15 billion and have 100s of millions of users worldwide).

Thursday, January 8, 2009

Lotto vs Stocks: what performs better? :)

A friend of mine decided to invest in equal amount this coming year in stocks, and in lotto tickets.  At the end of the year she will report on what performed better financially.

Real time updates on her progress here: Lotto Vs Stocks

Tuesday, January 6, 2009

Top 8 Tech Companies Started or Funded During a Recession

Some of the greatest companies in tech were started, or funded, during a recession.  A number of people I know are starting companies today despite the economic times.  Others are holding back, worried that a great company can not get up and running during a recession.  Evidence to the contrary below :-) (in no particular order)
  1. Cisco Systems.  Funded 2 weeks after "Black October" market crash, 1987.  As the global leader in networking gear, Cisco posted 2008 revenue of $39 billion.
  2. Wikipedia was started in January 2001 (during the dot com bust).  Now the #8 most trafficked website on the planet according to Alexa.
  3. HP.  Hewlett Packard was founded in 1939 - right as the US was transitioning out of the Great Depression due to WWII.  HP now operates in most countries worldwide and generated almost $92 billion in revenue in 2006. 
  4. Sun Microsystems.  Founded in February 1982, Sun Microsystems was born in the recession of the early 1980s.  
  5. General Electric.  A cutting edge technology company of its day ("there is this magic thing called 'electricity'") GE was started in 1876 during the "panic of 1873" (which, despite its name was actually a 6 year downturn lasting through 1879).
  6. Amgen.  The world's biggest biotech and maker of EPO was founded during the recession of 1980.
  7. Fairchild Semiconductor.  One of the pioneering semiconductor companies of its day (and a "grand daddy company" of Silicon Valley), Fairchild Semiconductor was founded in 1957, a new company growing into the recession of 1957.  Fairchild's founders included Eugene Kleiner (who went on to found VC Kleiner Perkins) Gordon Moore and Robert Noyce (founders of Intel) 
  8. Logitech.  Started in the recession of the early 1980s by two Stanford grads, Logitech recently shipped its 1 billionth mouse (the computer peripheral, not the animal!)
So what are you waiting for?  Go out and do something great!!! :-) 

Know of any other great tech companies started in a downturn?  Drop me a line :)