Is Bundling Video Analytics a Good Strategy?

JH
John Honovich
•Published Jan 12, 2010 00:00 AM
PUBLIC - This article does not require an IPVM subscription. Feel free to share.

While increasingly adopted, we believe bundling video analytics with cameras and recorders is a reflection of failing initial strategies rather than an inherently strong approach.

Bundling video analytics is increasingly common. The 3 biggest acquisitions of analytic companies have fostered bundling - DVTel/ioimage, American Dynamics/Intellivid and Honeywell/ActiveEye. Beyond that, a number of the most well known analytic companies have evolved into bundled product offerings - 3VR's searchable video recorders, Cernium's managed video offerings and VideoIQ's iCVR [link no longer available].

There are benefits of bundling:

  • Tighter integration can allow for improved performance of the analytics
  • Lower cost of combined product as duplicate costs can be eliminated
  • Greater revenue because more functionalities are being sold

On the other hand, there are downsides:
  • It is harder to be strong at both analytics and cameras/VMS - company can become stuck in the middle - average at both
  • Many customers only want one part of the package and may reject bundled product because of cheaper or simpler alternatives

As a practical matter, we believe the rise of bundling analytics is a result of failure to meet original investor expectations for rapid and sizable revenue growth. Many of these companies received tens of millions in investment, only to generate a few million in revenue after the first few years. This was primarily a result of failure to meet the original performance claims (both in accuracy and ease of use in real deployments). We believe this was a critical factor in the shift to bundling. Bundling will almost certainly increase revenue growth but at the cost of making the bundled provider a weaker alternative to more focused providers (3VR compared to Exacq, Cernium to Axis STS, VideoIQ to Mobotix, etc.).

Ultimately, bundling may be the best option but only given the failure of analytic performance to justify mainstream adoption.