google

...now browsing by tag

 
 

Are You Building Something That Will Move the Ball Forward?

Saturday, November 27th, 2010

My co-founders (David and Matt) and me are embarking on building a new company called Gigwalk.  The experience got me thinking about why we start companies.  I was going to write a long post about this but came across a Tim O’Reilly interview where he captures it perfectly while commenting on some of the TechCrunch50 companies “I don’t really care if they succeed or not, they are not really moving the ball forward in any meaningful way”

Ultimately life is too short to work on something where I could make a bit of money but really didn’t care for.   If my only goal was to make money I would be working on Wall Street (as a few of college professors encouraged me to do so). Wall Street is a great place to make tons of money working on stuff that doesn’t matter much.

If you take a much longer perspective – say 100 years – companies that move the ball forward succeed in changing what people do and how they think.  They change the course of history. Each of these companies dramatically changed billions of people’s lives over the past fifteen years.

  • Microsoft and Google changed how people work.
  • Facebook and Yahoo changed how people communicate.
  • eBay and Amazon changed how people buy and sell things.
  • LinkedIn is changing how people manage their career.
  • PayPal changed how we pay for stuff.

As Tim says these companies clearly created more economic and social value than they captured – the founders and investors did pretty well along the way.  I have no idea whether we will create a company on the same level as some of these — only lots of hard work and a bit of luck will determine that; yet this interview got me thinking on why we should start companies.

Why Talent is Limiting Supply in the Smartphone OS Market

Tuesday, August 17th, 2010

How many smartphone mobile operating systems can the market support? Wrong question. In a market of billions of potential devices with limited network effects [1] and hyper-growth demand the market could support lots of smartphone mobile operating systems – where lots is defined as greater than six and less than twenty. The limiting factor is not demand but rather supply. How many companies can build a competitive mobile OS? The fire sale of Palm and the launch of Blackberry Torch makes it clear that the list of companies that have the skills and resources to build, launch, and sustain a competitive mobile OS is getting smaller each month. Watching this playout feels like watching the top riders in the Tour de France fight their way to the top of Col du Tourmalet — each and every year only the most talented riders keep pace as the ride becomes absolutely grueling.

Companies are starting to fall by the wayside as the stacked teams of Apple, Google, and soon Microsoft take firm control of the race.
Lets look at a checklist of capabilities one needs to launch a competitive mobile OS. Half way down this list you will realize that very few companies have assembled a deep and wide enough talent pool to execute a smartphone mobile os on a global scale.

What you need What it gets you
User Experience
Multi-touch interface Parity
Visual appeal Potential differentiation
Multi-tasking Parity
Apps
Games, Games, and Games (1) Acquisition (great games sell devices)
(2) Lock-In (spend creates switching costs)
20,000 Quality Apps Parity
Discovery & Merchandising Easy & trust worthy
Media
Music Lock-In (spend creates switching costs)
Movie rentals Engagement
TV show rentals Engagement
Browser
HTML5 + CSS3 Compliant Browser Parity
Information Finding
Search Utility
Carrier financial incentive via revenue share
Voice Search Cool demo
Maps (limited number of suppliers) Utility + parity
Navigation (limited number of suppliers) Utility + parity
Communication Cloud
Mail Utility + parity
Voice Transcription Diffentriated
PIM Services Utility + parity
Payments
One Click Buy (micro transactions) Parity + Frictionless commerce
Global (90+ markets) Utility + parity
Fraud Mgmt Utility + parity
Launch Marketing
$250M consumer marketing Consumer awareness.
Percieved momentum for developers
$50M App developer launch Studio compensation to seed app library and app competitions
Devices & Distribution
#1 and/or #2 Carrier in top 20 markets (first yr) Market coverage
3 to 4 OEM’s building 8 million devices (first yr) Consumer choice of mid to high-end devices
Developers, Developers, Developers!
Popular IDE App quality
Time to market
Developer happiness
Deep SDK Device access (e.g. camera, acceleramator)
UI (e.g. animation, controls)
Service access (e.g. maps, contacts, mails)
Enterprise
Security Parity
Employee device & app provisioning Parity

[1] Limited Network Effects: During the PC war application developers propelled Windows PC to its monopoly position – users adopted the OS with the widest range of applications and developers adopted the platform with the most users. In the mobile smartphone OS war the top 20,000 apps that matter will be replicated to the top five mobile os platforms in each market or region muting the possible network effects because the development costs are relatively low.

Behind Google’s Data Buying Binge

Friday, August 6th, 2010

Google used to be based on a simple premise.  The web is a big place, we help you find the relevant piece of information for your question and direct you there — as quickly as possible.  You don’t consume information on Google, you simply find it.  Users only spend 3.4% of their time on search engines.  This is changing.  Having the best algorithm is no longer enough.  Google is investing heavily to own the data across key vertical categories and slowly becoming a destination experience for consuming this data.  Unless you own and curate rich data-sets there are natural limits to both the search relevancy and experience you can provide.  Google is quickly adapting by buying access to vast and rich data sets. Let’s look at their recent buying binge:

  • Travel: Acquired ITA Software which aggregates flight routes and pricing information and enables advanced search capabilities on travel data.
  • Local: Attempted to purchase Yelp.com, and after that fell through they ramped investment to build their own local data set.  Additionally,  Google has been investing for years in map with StreetView and satellite imagery.
  • Metaweb (Freebase): Structured data of  people, places, and things.

When Google focuses on a category like local this is what happens to the search experience.  For a query like Delfina Pizzeria (an excellent pizza place in San Franciso) rather than linking to the best sources of information like Yelp, Zagat, SF Chronicle, etc. Google first pushes you towards Google Places. It currently includes a mix of their own content and other sources of licensed content.  What happens when they have their own pictures, reviews, and check-in data — do they really need to license all this other content from the likes of  Zagat and SF Chronicle.

I expect Google to complete the buying binge by acquiring companies with rich data sets across other highly monetizable categories:

  • Shopping: Amazon and eBay to a lesser extent are capturing significant percent of query share in a very lucrative area.  If users bypass Google and go directly to Amazon for their product queries this represents a serious threat to their business.  They need to acquire a company with a huge selection of product data — rich and structured product attribute (size, color), inventory availability, pricing and promotions, and user reviews.  I can’t think of one company (beyond Amazon) that has done this well at the scale Google would need.  This may require acquiring multiple companies to create this.
  • Real-Estate: Is somebody like Trulia next?

Why Barring AdMob (aka Google) from the iOS is a Smart Move

Thursday, June 10th, 2010

Two platforms are emerging the mobile operating system platform and nested within it is the mobile app and ad platform.  For over a decade Google played for keeps in its platform (search connects searchers and advertisers), so why shouldn’t Apple do the same? By barring AdMob from the iOS they:

  1. Limits Google’s Ability to Achieve Mobile Ad Dominance: The mobile ad business is a two-sided network. Apple is ripping away one side of that marketplace (or at least a very large chunk of it — 100M eyeballs).  This will limit AdMob’s (i.e. Google) ability to achieve critical mass in the mobile ad space during these formative growth years.
  2. iAd Accelerator: With AdMob out of the picture advertisers will move their mobile ad budgets to other ad networks and clearly  the iAd network will grab some of these dollars accelerating its growth as it attempts to reach critical mass.  Remember, in any market with strong network effects the strong get stronger and the weak get weaker.  Apple is positioning iAd to get stronger.
  3. Keeps Ad Innovation Alive (some): A key hallmark of an open system is the increased rate of innovation. By opening iPhone to select mobile ad networks the iOS platform will still benefit from the ad innovation these providers generate.
  4. Switching Costs for Developers are Still Low: In a mature market developers could incur significant switching costs i.e. if AdMob had 10x better ad relevancy algorithms combined with an advertiser base 20x the combined competition it would be challenging for the others players to match the revenues developers were generating.  But in such a nascent marketplace nobody has these kinds of advantages.
    • Other mobile ad networks (Apple iAd, Millennial, Greystripe, and Mojiva) will quickly pick up the slack and provide comparable monetization rates.
    • The cost of switching to another ad mobile provider is not significant – a new billing relationship, a couple of lines of code, and new analytic tools to monitor your performance.
    • The vast majority of mobile developers making real money (now) are not doing so with AdMob display ads — they are doing it with paid apps.

Finally, do you really believe that app developers would give up access to 100M mobile users who have a credit card on file.

65,000 New Android Devices Ship Each Day. How Much Are They Worth To Google?

Friday, May 14th, 2010

Google recently announced that partners are shipping 65,000 new units of Android each day. How much is that worth to Google – in revenues, not brand equity, rather real hard cash?  Some simple math will provide us the answer.

First, What % of Android Devices are Shipping with Google as Default?

Most users are default users.  They use the email service, search engine, browser, etc. put in front of them.   Of these 65,000 Android devices, how many have Google as the default search engine? Almost all.  Lets assume 95% because there are just a few Android devices shipping with Yahoo! Search as the default.  So, 61,750 Android devices ship each day with the home screen search box or built-in search button hardwired to Google.

How many searches per month does an Android user perform?

Last year Google and Stanford published an excellent report on mobile search query behavior comparing the search usage patterns across PC, iPhone, and feature phones.   The report discloses two important numbers: the average number of search sessions (8.06) the average number of queries per session (1.86) users perform on their iPhone over a 35 day period.

Multiplying those two numbers (search sessions by queries per session) produces the average number of queries (14.7) an iPhone user does per 35 days.  Lets adjust this number upwards by 50% for two reasons:

  1. The data-set is from the summer of 2008. Since that time as the underlying devices and networks get faster the number of queries users perform increases.
  2. Android devices ship with a search box sitting on the home screen or a built-in search button. Whereas, the iPhone the web search box is out of sight integrated into the Safari app.

After adjusting the number upwards by 50% an Android user is performing 22 queries per 35 days, or 19 per 30 days to keep our units similar.

65,000 New Android Devices are  Worth: $7,011/month

$7,000 a month — thats it.  Lets see how we get to this number.  The 61,750 Android devices with Google as the default are generating 1.2M queries per month (# of shipped Android devices times monthly searches per Android device).  At an RPM (revenue per thousand) of $20 that is $23,370 a month or $0.36 per device/month.   Now, hold on. Google does not keep all of this revenue.  Google is paying carriers a traffic acquisition cost (commonly referred to as TAC) anywhere between 60% to 80%.  Using a 70% TAC we get our answer:

Google earns $7,011 a month in search revenues from the 65,000 Android devices that ship each day.

Google earns $0.11 a month in search revenue per shipped Android device.

Looking Ahead: What is the Search Lifetime Value of an Android User?

Android users are worth more then just $0.11/month because they usually keep their phones for longer – twenty-four months or the average contract length.  Assuming twenty-four months the lifetime value of an Android user is $8.63 (monthly search revenues per device x 24 months).  Put simply, each Android device shipping is worth $8.63 in search revenues over the lifetime of their Android device.  As Google improves monetization of mobile search queries this number will go up.  Using a TAC of 70%  Google earns $2.6 in search revenues per user over the life of an Android device after paying the carriers.

So, why was Google trying to bypass carriers and sell Android phones directly to consumers?  Simple economics.  By selling directly they earn 3x more on each Android device, it’s the difference between $2.6 and $8.6.

Notes:

  1. If you have better data you can change the assumptions, the spreadsheet is here
  2. I did not calculate the search monetization opportunity from the Maps application.  I believe a significant percentage of local search queries are moving from web search to the Maps application, I am not aware of sufficient data to estimate this revenue.
  3. Search monetization differs by market — I do not account for this.  I assume $20 RPM across all markets.

There Will be No Globally Dominant Mobile OS

Wednesday, March 10th, 2010

All three major players — Apple, Google (Android) , Microsoft –have launched or announced their mobile OS the question the question we are asking is who will dominate the mobile OS ecosystem? The premise of the question is flawed.  There will be no globally dominant mobile OS the way Windows dominates the PC.  Just a few of many reasons:

  • Price: Price-insensitive markets (e.g. most of North America and Europe) view their phones as jewelry. At $99  (iPhone 3G) it is not a very expensive piece of jewelry.  Highly price sensitive markets (Southeast Asia and South America) will see Android coupled with copycat devices.
  • Carrier Distribution: Carriers remain a vital distributor in most markets. They have financial (generously provided by OEM and web companies) and strategic (keep Google honest) incentives to distribute multiple platforms.
  • Open Platforms Win: An overstated arguement.  The 20% of apps that matter will get ported to all platforms with over 10% share and the 3-4 incremental days that it takes to get an App into the iPhone App Store is annoying but mostly inconsequential.  As the mobile browsers continue to expose more device APIs the Android is open argument wears thin.   OK, so you say well I want to go really deep into the mobile OS and create my own layer above the OS.  There are very few companies who have the resources and skill-sets to do this. OEM’s like Motorola are trying but don’t have the right skill-set.  The companies who have the resources and skill-set to this are the exact same companies who have launched their own mobile OS.
  • Data Lock-In: Data you create on your device will not stay on your device the same way that data you generated on your PC 20 years ago remained there.  As the world moves to the cloud in 2010 these client level data lock-in advantages are muted.   More important are the cloud based data lock-in (e.g. Y!Mail, Gmail, etc.).  The device is just access point.
  • Application Lock-In The argument goes something like this.  Users spend lots of money on Apps and music and given these investments will be locked into the platform where they made these investments. To build significant application lock-in users will need to be spending far more then $50 a year on Apps [source].  Recall, to create strong switching costs 10 years ago users spent hundreds of dollars on Microsoft Office, Adobe Creative Suites.  Could it really be anything less?

[For good history and insight into mobile fragmentation see Richard Wong's TechCrunch post and presentation]