IoT Analytics: Evolving Quickly at IoT Evolution

Analytics Evolves at IoTEvolutionExpo 2015

 

At the recent IoT Evolution Expo in Fort Lauderdale, I had the opportunity to participate in a couple of lively panels as well as spend considerable time sharing ideas with colleagues in IoT.

It was a great way to get a pulse on the state of the market.  Among the key takeaways is that IoT Analytics is heating up and continues to gain mindshare.  It seems like almost every session or informal discussion at IoT Evolution had an analytics angle.

From my conversations, I have synthesized four key lessons from successful customer implementations of IoT analytics.

1) Fast Analytics is Crucial to Maximizing Business Value.

Many IoT business problems are extremely time-sensitive, requiring “fast analytics” that is executed continuously and in real-time (in contrast to “slow”, traditional analytics, which is executed hourly or daily).  Examples of time-sensitive IoT problems include:

  • Preventative failure in manufacturing where breakdowns are costly and must be anticipated and prevented when possible, or mitigated when not.
  • Predictive 1:1 marketing where real-time offers to consumers based on current behavior are critical to closing sales.
  • Early detection and mitigation of cybersecurity attacks that could disable critical industrial equipment or utility grids.

2) Fast Analytics over Slow Data is Not an Oxymoron

Fast Analytics should not be confused with the velocity of the data itself.  Fast Analytics can often be just as useful over “slow data,” where devices report in on an hourly, daily or weekly basis.

Even over such slow data, Fast Analytics can still have a huge impact for several reasons. First, there still may be residual time-sensitive value in the data, and Fast Analytics enables the harvesting of that value.

Second, many important IoT use cases involve networks consisting of millions of devices with significant amounts of new data generated every minute. An early indicator of a problem or opportunity may be hidden in the data.  Fast Analytics over slow data can provide value by:

  • Detecting anomalies and leading indicators early
  • Leveraging these leading indicators to predict opportunities and threats early
  • And then triggering action in time to capitalize on opportunities or mitigate threats.

Finally, Fast Analytics over Slow Data has one other major long-term advantage:

  • It can future-proof you against ever faster data cycles

Your IoT network today may have a daily reporting cycle, but next year this could be accelerated to hourly, and the following year to every few minutes.  No matter how fast the data cycle times, Fast Analytics will always be able to keep up.

From Smart Meters to Smart Cities, Fast Analytics will likely add more value than Slow Analytics, even over Slow Data.

3) Fast Innovation and Continuous Learning through Experience beats the “Best” Analytics Models.

Many organizations focus on building the very “best” model initially and then investing heavily in operationalizing this complicated model.

Since the cost and time to operationalize a model is high, the “best” model is infrequently refined and updated. However, “best” models tend not to remain “best” for long in the face of changing IoT and business requirements and the iterative nature of learning.

A different approach is to invest in an IoT Platform that permits one to operationalize new models quickly. In this way, an organization can set up a “Rapid Innovation Cycle” of model à operationalize à observe à and refine. This empowers an organization to continuously learn and operationalize their learning, and to adapt quickly to changing IoT and business needs.

Analogous to the ‘agile’ movement in software development based on rapid software iterations, IoT is driving us towards a similar trend in analytics – Agile Analytics.

4) Action is the Last Critical Mile in the Analytics Value Chain. 

The final lesson takeaway from the IoT Evolution Expo is that analytics by itself is not sufficient to capture the value inherent in IoT use cases.

The final and most critical step in analytics is to trigger actions to capture business value – a topic we have written about in previous blog entries.  Insight-driven action, powered by advanced analytics and performed at the right time, is the key to value creation in IoT.

Final Thoughts

IoT Analytics has come a long way since the IoT Evolution conference last year. Driven by a number of exciting new use cases, IoT Analytics “best practices” are emerging and accelerating value creation – a trend we expect to continue as we look forward to another exciting year for IoT Analytics.

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>