Thursday, February 28, 2013

New York, New York, it's a Hell of a town

Over the past 5 years, I have been continually testing my thesis that the best way to make money investing in technology companies is to invest in companies which have the potential to be market leaders; in a market worth caring about. Around ten years ago the folk at Morgan Stanley (at least that's where I recall they were from) did a great piece of research showing that the overwhelming majority of any market segment's capitalization rests with a scant few, mostly public firms. The message was compelling, if you are not in the top three market share spots, an investors risk/reward ratio goes so sky high that you would be better off playing the lottery (the risk increases, while the reward simultaneously decreases). 

Though you may make the right call to invest early in a burgeoning market, unless you execute towards a leadership position, it's going to be a problematic investment. Choosing the right market is a necessary, though not sufficient ingredient for success. It's essential that the management team, supported by investors with sufficient capital and drive have a common objective. 

To paint with a broad brush (exceptions abound) I think it's fair to say that there is a distinct difference between East coast and West coast investors and management teams. I believe that West coast investors and teams have been far more market share driven than their right coast sibling and believe that the preponderance of technology  market share leaders being in the Valley is a direct result of this culture.

The Boston to NY corridor does have its share of companies which are showing great signs of success. Some look really great and have the potential to be market leaders, such as Tumblr, Payoneer, Etsy or 10Gen. But it's also fair to say that since the heyday of Doubleclick and AOL, we may not have a critical mass of companies which are hell bent on being market leaders, in markets which matter.

Friday, February 22, 2013

Mobile company building statistics

Despite having more than 1 billion mobile devices in use and tens of thousands of companies selling and giving away mobile based applications, the number of  successful 'mobile first' companies is still suprisingly small.  Here's some statistics from an end of the year report by Flurry which should help set performance benchmarks for companies to compare their performance and plans against.

For those unfamiliar with the Gartner Magic Quadrants, the upper right corner is where you want to be; clients with high retention and high frequency of use. These can be really fast growers as the churn is so low. The next best place to be is the upper left corner, with intensive use, but high(er) churn. Tough to build a subscription or successful freemium business here, but advertising may work quite well.

 QuadrantChart EngagementRetentionStats ByCategory resized 600

Table EngagementRetentionStats ByCategory resized 600

Social Media explained (image)

From my friends at Tracx:

Monday, February 4, 2013


The other day I realized that we have too many cable connections in the house. Our TV's are just not being used the same way they were in the past and using systems such as AppleTV and Tversity, coupled with a new generation of HD indoor antenna (I purchased a Leaf Plus for $69), we can revert back to an enhanced free TV model where we have access to all the basic channels, libraries of movies on demand, and streaming from our PC's.

Sure, we are keeping a couple of 'cabled-up' sets for viewing, but it's becoming the exception, not the norm in our house.

Big Data

For the past couple of years the explosion of data has opened a great promise for companies to mine data to  identify and better serve customers, identify new opportunities and gain great(er) efficiencies. The research firm IDC estimates the Big Data market is growing 40% per year and is now over $2B. While there's been some great headway with early imnplementations, I think many corporations today are drowning in the holy grail of Big Data. It's not so much that the data is not there for the mining, instead, there's just so much of it that the issue is not the data, it's identifying the proper question. Having access to data is necessary, but not sufficient, to getting value from it.

Knowing what to ask, and being organized to act on the information gleaned from the data are moving to the fore of our industry. Though data has been around for eons, the huge data sets (5 Petabytes and more) available by combining internally generated data with purchased data from social networks, blogs, etc fundamentally is changing the landscape. Traditional relational databases and business intelligence software were not designed to handle inflow of this magnitude, and to do so in real time.

Assume that you are involved with one of the dozens of companies with more than 1B page views per month and more than 20mm unique visitors per month. The amount of data generated around  user behavior is close to infinite. You can slice it by demographics, on-site relationships, borwser, location, OS, engagment, etc. Moreover, you can correlate it with external events like the weather and press releases. Finally, 'special' runs which enable targeting for advertisers or data customers add myriad layers of complexity.

The same way that operating systems have evolved to hide the complexity of computing, it seems to me that the next generation's successes in 'big data' will combine the traits of masking complexity, with a layer of intelligence which highlights the critical data which contains really useful information.  We are seeing a nice wave of this nascent trend through the successes of companies in the IT space such as Splunk, Solar Winds and Puppet Labs.