Optimise your Master Data Management before entering the IoT era
Internet of Things (IoT) can and will change business models, increase productivity and even our way of life. It’s an exciting topic that is getting a lot of attention. Meanwhile products like SAP Leonardo and Microsoft Azure IoT Suite reach enterprise grade maturity levels. But are you ready to exploit the full potential of IoT? Many businesses are thinking about investing in IoT infrastructures. But what many people seem to forget is the importance of Master Data Management (MDM). It’s a critical – albeit not that sexy – enabler for IoT initiatives to succeed. I will explain why.
IoT infrastructures are built to receive, process and analyse massive amounts of data. This requires a process that is highly automated. In these infrastructures, machines spit out large amounts of data. But without the right context, this data is meaningless. When you receive an error message from a machine, you need to know what kind of machine it is. What’s the location of the machine? When was it last serviced? And by whom? What product was the machine producing? And for which customer? That’s where Master Data Management comes in. You need to have accurate Master Data in order to give incoming raw data the context needed to interpret the information and to trigger the right responses. Your Master Data provides a single trusted view to the raw data received from IoT devices.
Master Data Management bridges the gap between the world of IoT (raw analytical data) and the world of business applications and processes (enterprise data). Simply put, MDM translates raw data from IoT devices to useful data that your systems can understand and work with.
The development of IoT sets new challenges and requirements for MDM environments. It changes the dynamics of MDM. Topics like the lead-time of MDM processes, data quality and the availability of master data are crucial components of any IoT strategy.
Real-time data and real-time action require real-time context. When a machine stops functioning, you want the system to send an urgent request to a service engineer, immediately providing the right context by sending him all relevant details. Real-time processing of IoT data is dependent on accurate Master Data, so the creation, maintenance and distribution of master data should happen in real-time as well.
Before Internet of Things existed, a data quality error could be recognized and corrected in time before it would cause any major disruption. But real-time data sent from IoT devices enables action in real-time, without human intervention. Data quality errors will therefore have an immediate impact on operational processes. There is less time, or no time at all, to recognize or correct these errors.
But also with stored data that is used for - for instance - predictive modelling, a data quality error can lead to an incorrect representation of the truth. In contrast to real-time data, the impact on stored data is more hidden as such errors have an impact deep in these models. Some errors might take weeks, months or even years to detect. This is something you want to avoid, because this data is used to define business strategies and to add analytical context to the real-time data.
The knowledge and expertise that your company is built on was formerly stored in the minds of you and your staff. Now, you’ve moved towards a future where that information is stored in machines to better interpret data and to initiate follow-up automatically. But more automation requires more context. This means that not only quality master data is required but also more master data is required.
Additionally, to solve some of the infrastructure challenges of processing vast amounts of data ‘edge analytics’ and ‘distributed analytics’ are gaining popularity. These systems are designed to perform the analytics at the point – or very close to – where the data is collected. Often, this is where action based on insights provided by the data is most needed. Rather than designing centralized systems where all the data is sent back to your data warehouse in a raw state, where it has to be cleaned and analysed before being of any value. But this type of analytics requires accurate master data as well.
Make someone responsible
So, evaluate the maturity of your Master Data Management before starting the era of Internet of Things. Not only on your technical infrastructure but also your MDM processes. Make someone in your company responsible for MDM and involve him into your IoT projects. Invest in MDM and your business will fully profit from Internet of Things.