Why IoT Analytics is more challenging than Traditional Analytics

8 reasons IoT Analytics is challenging

At Vitria we often get asked why IoT Analytics is so much more challenging than traditional analytics. After years of experience on IoT Analytics projects, we have distilled our experience down into a specific set of lessons learned. We have found that there are eight key reasons that analytics in IoT is more difficult than traditional enterprise analytics.

  1. The speed and scale of data in IoT is unprecedented in traditional IT environments. It is routine in IoT to see millions of events per second, which are orders of magnitude bigger than traditional enterprise scenarios and use cases. This type of volume and speed requires new kinds of architectural approaches and technology.

  2. Variety of data – The range of types of data in IoT is also unprecedented. IoT data and use cases can include device data, external data like weather, enterprise data from CRM and other IT systems, and business process data. Putting this wide range of data into context for decision-making is a new level of analytics challenge.

  3. The need for Real-time Analytics. Many IoT use cases are time-sensitive, and the analytics and potential action need to be timely. These speeds are orders of magnitude beyond anything in traditional enterprise environments.

  4. Greater diversity of analytics methods required for IoT use cases – advanced analytics in IoT means using a variety of analytics in combination to yield a recommended next action. This often involves historical, real-time streaming, predictive, and prescriptive analytics. It is rare for traditional enterprise analytics to require this diversity.

  5. Timely Action is Mandatory. The fast pace of IoT data flows means that action to capitalize on the data must be taken quickly, or the value perishes. Given the speed and scale of IoT, the full or partial automation of actions is often required.

  6. IoT Software/Ecosystem environments are very complex. The software components required in an IoT Analytics solution are more complex than previous IT eras. Also, many of the key technologies are early in their maturity cycle. This consequences of this immaturity can be seen in Gartner research showing that 60% of Big Data projects fail to meet their goals.

  7. Highly Technical skillsets are often required. The technical skillsets in Big Data/IoT are often highly specialized and may include the learning of novel programming languages. These highly specialized skills that are in short supply.

  8. Operations analysts and managers are not well-served by traditional enterprise analytics. Operational analytics in IoT demands “heat of the moment” processing and action that is not common for enterprise BI requirements. Operations analysts and managers are the focus in IoT – not necessarily traditional business analysts. Traditional enterprise solutions are simply not designed to provide the real-time analytics required for operations analysts dealing with the demands of IoT use cases.

This set of eight challenges illustrates why IoT Analytics is quantitatively and qualitatively different from traditional enterprise scenarios. They have significant implications for executives and operations managers looking to get traction and value from IoT and analytics. As you consider the software and IT infrastructure required for IoT, it is critical to ask both your prospective vendors and internal IT teams how they will address these issues.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>