Friday, December 12, 2014

Welcome to The Show(1) HockeyApp

In what can be described as the latest Snipe(2) in the Barn Burner(3) game we call application analytics, Microsoft announced its acquisition of HockeyApp. Most of the early commentary I’ve read seems to focus on the fact that Microsoft has invested in a native iOS/Android API, but to me that’s not the most interesting nugget… what’s most interesting to me is that HockeyApp has been built to be a hardcore Stay-at-home defenseman (4) (that’s the last hockey pun, I swear, see definitions below). 
(1) The Show (noun): the NHL, used in the context of “making it to The Show”.
(2) Snipe (noun): a powerful or well-placed shot that results in a pretty goal.
(3) Barn Burner (noun): used to describe a game that is high scoring, fast paced, and exciting to watch.
(4) Stay-at-home defenseman: A defenseman who plays very defensively. He doesn't skate with the puck toward the offensive zone very often but will look to pass first. Usually the last player to leave his defensive zone.

 Maybe I’m too close to this space, but hasn't Microsoft been consistent and clear in their communications that it was always the plan to have Application Insights support native iOS and Android apps? – so they bought some technology and talent to accelerate the process …that’s neither novel nor controversial – seems like a smart move.

What DOES strike me as interesting is HockeyApp’s focus on analytics for testing versus production – and for enterprise use versus consumer-facing.

Microsoft is in an all-out sprint to deliver a comprehensive and fully integrated dev-devops-ops ecosystem where the distinctions between enterprise, b2b, and consumer categories dissolve – and, with HockeyApp, they appear to be killing two birds with one stone; native iOS and Android APIs AND an analytics framework optimized for test and other (relatively) small and well-defined user communities (such as some enterprise scenarios) – two areas where Microsoft has traditionally been quite successful.

Support for side loading, user-by-user index buckets, and a privately managed feaux marketplace all work well in these scenarios, but (I would suggest) the very same implementations will struggle under the strain of high volume 24x7 operations - but that's OK, that's not the intent.

It’s clear that, as application analytics matures as a category, we should expect to see increasing specialization and segmentation – software built to track shopping cart behavior and user loyalty (or generate system logs) is not going to be able to cover these increasingly well-defined use cases. 

Here's my latest chart comparing the various "categories" of application analytics solutions... (all errors and omissions are obviously mine and mine alone)

As always, checkout the latest on PreEmptive Analytics at