You often hear the phrase "the tail wagging the dog" and anyone who owns a Labrador will see it in action on a daily basis! In IT we come across cases of this on a regular basis as a new project ends up 'taking over' when it should be supporting the existing processes. In extreme cases this has been known to completely destabilise companies and damage their competitiveness. An example of this being Nokia, whose MeeGo smart phone operating system development project became so out of control it caused Nokia to lose its leading market position and eventually end with it being sold to Microsoft. Internet of Things (IoT) with its promise of cheap and ubiquitous sensing technology does carry with it the risk of becoming a very big tail that could wag the corporate dog.
Whilst writing this blog I came across a paper Microsoft recently published of some research they had done with North Carolina State University (NCSU) into the use of telemetry with the Microsoft organisation. Microsoft is one of the true hyperscale companies, with over 1 Million servers running Azure. They provide innumerable services to support their business , which produce telemetry data that people within Microsoft want to see and use. This is 'big data' and although of a larger scale than even the data volumes that are envisioned for IoT solutions, the lessons learned here can help everyone trying to build IoT solutions.
For me two of the findings in the report stand out. Firstly, Microsoft has the same problems as most companies - a lack of 'data scientists' to produce all the insights that people want. The second is that there is some distrust of insights where people cannot see the source data and how the insight was generated. These are problems in any organisation from the smallest up to the largest.
So let’s look at these findings, I’ll start with the data processing problems. Consider for example smart meters. The energy supplier might start with a meter reading every 6 months and move to a meter reading every 15 minutes. That's a 17,000 fold increase in data. Clearly that sort of increase in data flow in many organisations would be too much to handle in its raw form. The electricity utilities know that they need to work at this granularity to support demand side management as this is a vital component of smart grids, so accept that they have to transform their business or face huge problems down the line. For a company embarking on a digital transformation journey and embracing IoT, they have to remain competitive during the transformation and sinking under the weight of new data without a way of processing it effectively does not aid that.
As for trust, this is as much about the 'soft skills' as the technology, but it does show that analytics must be auditable and traceable so people can understand where the answers have come from. In any organisation there will be a mistrust of a new technology from some, and part of the project should be to get everyone 'on board' to the solution as it’s being developed.
So how does a company make sure that the tail does not wag the dog with IoT? At the simplest level they have to start from the business outcomes and work back. Collecting huge amounts of new data with no real idea of what to do with it is a sure fire way to incur large costs without realizing any significant benefits. The process must be top down and the value proposition must be understood before any significant work is done.
In my next blogs I will explore more aspects of this.
In the meantime, learn more about IoT, download a free copy of our IoT for Dummies Guide.