Internet Trends from Mary Meeker

Mary Meeker, a partner at Kleiner Perkins Caulfield and Byers, published her '2012 Internet Trends Year-End Update'



The full presentation is on Slideshare, first seen on Venturebeat.

Here are some of the figures that caught my eye:

• 2.4 billion Internet users with 8% year on year growth. 500 million of these (ie 20%) are in China - this only represents 40% of penetration in China.
• Let's contrast China's figures with the US: 250 million users with only 3% annual growth with a penetration of 78%.
• 29% of adults in the US now own either a tablet or an e-reader.
• Mobile devices now account for 13% of worldwide internet traffic - 4% in 2010. In India, mobile Internet traffic now exceeds desktop Internet traffic.
• The graph above shows the dominance of Windows and Intel in the 90s and early 2000s - and the explosion in Android devices.
• The number of landlines peaked in 2006 with 1.3 billion.
• In 2102, it is estimated that 2 billion bluetooth devices were shipped (up 87 times in 10 years); 1.5 billion Wi-fi devices were shipped (up 5 times in 4 years)
• By Q2 2013 it is esimated that the number of Smartphones + Tablets will exceed the number of PCs in the world.
• If 2,000 zettabytes was created and shared in 2011 then 8,000 zettabytes are estimated to be shared in 2015. (1 zettabyte = 1 trillion zettabytes).
• In 2006, Nokia Symbian operating system was shipped on 65% of smartphones (globally). By 2012, 65% of smartphones were shipped on Android and another 20% were shipped on iOS

Wi-fi predominant connection over mobile network for smartphone users

Deloitte Global Mobile Consumer Survey 2012 shows that 58% of smartphone users and 93% of tablet users connect to the internet over wi-fi rather than over the mobile network. It's faster more reliable and more responsive.

Interesting, smartphone users in other countries (US, Germany, France and Japan) prefer to connect via the mobile network.

• Other stats: 30% would prefer to pay for a fixed amount of data and then pay addition usage charges once they have reached the limit.  
• Half the smartphone owners subscribe to less than 1Gb of data and only a fifth of respondents subscribe to unlimited data package.

Mobile Operators are missing a simple trick - but one that massively demonstrates their value: a simple counter that is constantly displayed on their home page which shows their usage since last billing date / top-up date vs their contract ceiling. It should display call minutes, text minutes and data limits.

What I find irritating about mobile operators is that they wish to charge users a different additional tariff for tethering another device (for me, a laptop or a netbook) to their mobile phone. I tether these device infrequently to my phone, but when I do, it is very useful, but my usage isn't predictable enough to justify the expense - particularly when swinging off someone's wi-fi usually works.

Naturally, mobile operators don't want to offer unlimited data - this resources isn't infinite. Such a policy is madness, as I have blogged a long time ago:

• Orange (at last) has a common sense data plan (April 2007) 
• 3 launches X-Series - Mobile Broadband bundle - when will they learn?? (November 2006)

However, I would like to use what has been contracted without any gotcha clauses! David Halstead, technology, media and telecommunications partner at Deloitte in Cambridge recommends:

Mobile operatios should offer their customers a seamless connectivity experience including Wi-fi hotspots that their customers can use when they are out and about.

Hmm, this concept is known as fixed mobile convergence and has been around for ages!

Article first seen on Cambridge Network.

Data usage on smartphones delivered mostly by Wi-fi

By 2015, Juniper analysis predicts that 63% of data will be delivered to smartphones by wi-fi and femtocells versus over the mobile networks (here's Silicon.com's Cheat Sheet on femtocells).

The quantity of data required will be 14,000 Petabytes (1 Petabyte = 1 million Gigabytes). This is an eye boggling amount of data - equivalent to 18 billion film downloads or 3 trillion music tracks. This is 'chunky' - Apple recently purchased 12 Petabytes of data from Isilon Systems, rumoured for video storage for iTunes.

There are two opposing effects here in mobile data:

• users using wi-fi for browsing / increased wi-fi penetration
• fixed to mobile substitution - ie people pitching their fixed line data connection for all mobile data connection

So the net effect is pretty big data chomping, but not as bad as mobile world feared some time ago.

Operators share 3G masts in UK

3 and T-Mobile intend to share their 3G cell towers. They expect to save GBP2 billion (USD4 billion) over ten years, by decommissioning over 5000 duplicate base stations.

Analysys, in this insightful report, applauds the decision. (Note: for some time after 3G licenses were sold in the UK, sharing of infrastructure was illegal. This requirement has since been relaxed.)

Voice remains the mainstay of ARPU with SMS being the primary source of non-voice ARPU. The demand for heavy data hasn't been apparent enough. AND that requirement is threatened by substitutes such as Wi-fi and Wi-Max, which means that operators need to invest in alternative technologies (examples below), which reduces the investment that they can make in their traditional cellular infrastructure.

• fixed broadband ie triple plays and quadruple plays
• Mobile TV
• Femtocell - indoor base stations

Analysys also sees this as indicative of how operators will make future investments (such HSPA+ & LTE) - a strategy that led to the dominance of GSM.

Analysys calls this a 'major shift by mobile network operators' - and I agree.

With ads .... radio, television, wi-fi, telephone?

Those ingenious chaps at Fon (free wi-fi) have been messing around with an ad-supported service - and I don't mean the dull Google Ads variety. When you connect to a 'Fonera' (ie a Fon hotspot), you'll have to watch an ad to continue surfing - one every 30 minutes, according to Martin Varsavsky's blog.

Any reason why the same business model can't be applied to ordinary ISP, IM or (mobile) telephone service? Not really - apart from market appetite.