Data Quality Matters a Great Deal in the Internet of Things
There’s plenty of excitement – and an abundance of just plain hype – about the enormous possibilities with the emerging Internet of Things (IoT). IoT offers intelligence and transparency to businesses that seek to better understand the world around them. The market projections are soaring into the billions and billions range – for example, BCG recently issued estimates that B2B spending on IoT technologies, apps and solutions will reach $267 billion in the next three years, and IoT Analytics spending is predicted to generate $21.4 billion at that time.
There’s no doubt that quantity of data is the essence of IoT. However, there also needs to be increased quality with the data generated by IoT connected devices, and subsequently moving through enterprises and to decision makers. It has to be trusted.
Sam Ransbotham, writing in MIT Sloan Management Review, recently explored the implications of the data trust challenge as we move into the IoT era. He and his colleagues are finding, for example, that trust is the necessary glue that will make IoT work.
The good news is that data quality rises as IoT implementations mature. This may reflect a cycle of learning and continuous improvement as an IoT-based network becomes an important part of the business. “We found increased experience with IoT projects is associated with improvements in the timeliness, detail, accuracy, and reliability of data,” Ransbotham relates. “A greater volume of data from IoT devices seems inevitable. But beyond that, organizations improve over time in their ability to get better quality data, not just greater quantities.”
Data timeliness shows the most profound impact on quality perceptions, “About 40% of respondents whose organizations aren’t active with IoT reported that their data has ‘mostly’ or ‘completely’ sufficient timeliness; in contrast, 76% of respondents who have two years or more of IoT experience said their data was sufficiently timely.” It’s important to note that “as systems monitor and transmit data closer to the source, delays associated with data gathering decrease.”
As enterprises rely on IoT data for a range of functions – Ransbotham identified areas such as improved customer experience, early warning on poor equipment performance and failure, and automated alerting for critical systems, the trust in such data needs to be in place.
How can data quality be addressed in fast-growing IoT initiatives? Standardization of data is key, ensuring that data coming from different sources tell the same tales. Also, IoT devices and networks need to be secure, to avoid tampering.
Thomas Davenport, analytics guru with Babson College and MIT, also urges enterprises to hold device manufacturers’ feet to the fire. “Insist on two levels of calibration from your device supplier,” he noted in a post last year. “First, there should be rigorous calibration before the device leaves the factory, and an on-installation calibration routine to ensure that the device works as expected. Second, ongoing calibration is required to make sure the device continues to work properly. Ideally, the on-installation and ongoing calibration routines should built-in and automated.” Ultimately, Davenport cautions, “you should not expect perfection, particularly with new devices. But you must insist on rapid improvement.”
As with everything in enterprise IT, big data and IoT these days, continuous improvement is the key. Especially when it comes to the data we trust to guide our organizations.