What seemed like nothing more than an interesting concept only a relatively short time ago, the Internet-of-Things (IoT) has become a leading hot topic in IT and business circles.And no wonder, given the predictions of near-term growth that have been aired.
Numbers like these, however, shouldn’t surprise when one considers the productivity and other benefits the IoT potentially offers. Marketers see opportunity to be had by tracking shoppers’ behaviour while they’re are out and about; logistics companies can direct rerouting and schedule maintenance based on real-time feedback from vehicles in transit; farmers are remotely managing the progress of their crops; and manufacturers gain from being able to address unexpected plant issues before production stalls. Organisations of every type are beginning to see new revenue channel opportunities by tapping into data at the edge of the business structure.
And while success is obviously reliant on a strong communications infrastructure, organisations need to exploit another critical IoT element for maximum gain – implementing a strategy built on being able to fully leverage the masses of streaming data the IoT generates. This is where analytics comes into play.
Listening to the data
The data volumes data generated by the IoT are obviously only valuable if they are captured in their entirety in an organised way, and properly understood. Relevant data emanates from a variety of sources – from different devices in and beyond the network including customer interaction, social media, in-house production, historical records and more.
Effective listening calls for deploying advanced analytics and machine learning algorithms to the data and this can be done either when it’s in situ or – preferably – when it’s still in the form of streaming data on its way through the network. In-stream analytics keeps pace with the speed of the IoT and detects patterns of relevance in mid-activity. This enables instant programed or manual reaction to address anomalies on the one hand or quickly exploit opportunities in the other.
Understanding and acting on what is heard
The usefulness of IoT data for business needs is no different to that of any other data. It has to be the right data for what is sought to be achieved and it has to be clean. Additionally, its usefulness depends on being able to manage it effectively. This calls for a strategy to handle large volumes of very fast-paced data flows with a complete range of advanced predictive analytics capabilities including ingesting and visualisation. The analytics tools applied to IoT data must be able to test hypotheses and manage developed models into and after deployment.
The strategy should be to establish an integrated system that will cater for the needs of every type of user – from those who will only want simple descriptive analyses, to those who can work with advanced predictive models delivering insights. It should also include an automated response capability for things like alternative consumer offers on the fly, and alerts for follow up actions.
There are many different technologies that can be built into an IoT management infrastructure and each step in the process can be addressed separately. However, it is far better to have one comprehensive platform for selectively absorbing and evaluating IoT data for usefulness – leading to analysis, understanding and action. All data is potentially valuable but IoT data is also fast so it makes sense to integrate the handling of it in a fast-flowing single stream.
But the ability to listen and act effectively on what the IoT can tell us is not just about deploying a modern analytics platform. Organisations will need to re-tune existing data management and analytics processes for the IoT age and encourage a culture in which jobs themselves will also become more interconnected than ever before.
Geoff Beynon is the General Manager of SAS New Zealand.