Surfair
162724

Main “streaming” of streaming analytics

The next generation of streaming analytics solutions

Technology trends and news by Debjani Deb
February 2, 2017
Short URL: http://vator.tv/n/489f

In the last six months, a good bit has been published about Streaming Analytics and the various use cases that it can power within an enterprise.

These have covered use cases in areas such as fraud detection, marketing and customer experience.

However, most of these have been written from the viewpoint of powering insights rather than action i.e. how does an enterprise draw on multiple sources of real-time (streaming) data and derive insights from them. Also, most of the vendor landscape in streaming analytics has been populated by platform level offerings targeted at the CIOs to do just that – put a range of real-time sources together to derive insights for various functions within the enterprise.

I contend that the next generation of applications of Streaming Analytics, will go one step further and equip enterprises to take real-time actions based on the real-time insights.

Action could be machine-to-user (experience, service, marketing) or machine-to-machine (fraud, BPM workflows).

Products that enable such use cases will be powered by stream processing platforms, but will also enable higher-level components in the stack.

Some examples of such components are activity-based profile management, pattern and trigger management, real-time response management, and individual activity dash boarding. This will equip the business user to enable use cases out of the box, without being dependent on IT.

I believe this will manifest as a whole new category of products powered by stream processing at its core. Such products could be categorized along horizontal lines such as stream processing for security and fraud detection, stream processing for customer service, stream processing for marketing, stream processing to power customer experience, OR, they could categorize themselves along vertical lines such as stream processing for financial services, retail, and so on.

Today, we see Big Data teams and CIOs at enterprises experimenting with stream processing, often with existing open source platforms. This takes significant time and effort.

In fact, what they need is an engineering acceleration to enable their stream processing use cases. I believe, that in the coming years the vendor landscape will rise to this challenge and offer out-of-the-box solutions for streaming analytics. The business users will be in a position to embrace the functionality that these vendors bring without too much dependence on IT.

We will then have moved to the next generation of streaming analytics solutions that are about actions and not just insights.