Many analyst firms that estimate mobile device market share rely on "shipments" data. Those include IDC, Gartner, NPD and Strategy Analytics, among others. Few of these firms rely on internet traffic and actual usage or sales data to fuel their estimates and forecasts.
The reason for this is simple: shipments data are easier to get than sales and usage data. While these estimates can be "directionally" correct they're often wildly inaccurate as a practical matter. Indeed, they often misrepresent what's really happening "on the ground." Even actual sales data often don't present an accurate picture of the marketplace.
Consumer survey data, such as used by comScore, Nielsen and Kantar for market-share projections, are in most cases better and more reliable than shipments data (and in some cases sales data). Best of all is actual usage or traffic data. A very clear case-in-point is the tablet market.
IDC released an updated tablet-forecast earlier this month. It shows Android tablets with a global market share of 61% and the iPad with a 35% share.
Looking only at these estimates, one gets the clear sense that the iPad has lost momentum and its preeminent place in the tablet market -- in the way that the iPhone ceded market share to Android handsets. The only problem with this shipments-centric forecast is that it bears almost no relationship to the reality "on the ground."
Checking actual tablet-generated traffic on a global basis we see that the iPad has a 74% share vs. 23% for Android. In Europe the story is much the same. In the US the iPad has a 79% share of tablet traffic. Other traffic-data sources show a comparable if slightly lower figure.
The aforementioned numbers are from StatCounter. Net Marketshare data similarly show the Safari browser as the dominant browser (56%) among mobile devices (smartphones + tablets) vs. 25% for Android (or 33% if Chrome is included; however Chrome is used on iOS devices too). Data from ad network Chitika also show that the iPad's share of tablet-based web traffic in North America is around 80%.
What are we to make of the massive discrepancy between the IDC 2013 estimates and these three traffic sources? Perhaps it's not important to try and reconcile these figures. Rather we should be asking which data "matter"? The answer is: usage is what matters.
Usage matters to developers, publishers and marketers trying to allocate budget and resources. What if millions of Kindles were given as gifts but later sat on bedside tablets, used only occasionally or in very limited ways? The simple notion that they're "out there" is irrelevant if they're not being used.
Collectively we should reject device "shipments" (and even sales) as a definitive market-share metric. Instead the industry should look at more concrete metrics such as traffic and other usage-based data that show what's actually happening in the world.
In the end this is what really matters to everyone, including investors. Even reliable consumer survey data about device possession and usage are better than shipments figures. Actual traffic data are less susceptible to misinterpretation (or manipulation).
Usage doesn't work as a forecasting tool for obvious reasons. Here projections will need to be based on some mix of assumptions and sales trends and other data. But the degree that the media and tech industry simply pick up and run with these regular shipments numbers without comparing them to actual traffic or other usage data -- I've been guilty myself -- is sloppy and misleading.