The Challenge
Therefore the challenge, and the biggest hurdle to data centricity, is the fact that data is still stored in siloes. We need to make sure that instead of different business departments creating the same data–set, the data can be shared between these departments. For this to happen, the siloed applications must be broken open through shared access to data and analytics that offer real-time visibility resulting in a horizontal data platform.
The Solution
In such a horizontal data platform, every business unit/use case will store its data in a central layer, and build its application on top of this central storage layer. The benefit of this is that they can easily and quickly include data that they didn’t possess or were not aware of. By having this central data layer, they can become a data–driven organization and take a QuickStart in building additional uses cases, enabling them to explore relationships between different data sets which they may not have discovered in the past.
Additionally, the horizontal layer could allow them to create a new businesses model. This would make the data and insights available to customers or suppliers who could also benefit from the data, by building a business model for sharing the data through a product.
Benefits
It is therefore in the interest of companies to open their data siloes, such as ERP and sales software. Once this happens, they can start creating a central data platform where they can store their newly created IoT data (streaming data) and legacy data sources.
They should also make sure they select IoT solutions which don’t lock them into a vertical silo. We still regularly see companies building end-to-end IoT solutions where they can’t get access to the data directly, or they need to pay an additional fee for it. Through this solution, customers could also use third-party data from other suppliers, as well as publicly available data on the weather, electricity prices, etc.
If this central data layer is built, users will also be able to start using AI and Machine Learning, as both services require these kinds of data sets. The layer will also allow them to use new services to build applications in a cheaper way (e.g., a data warehouse on top of a data lake, where the data warehouse is very expensive and difficult to set up) and enable the standardization of API serving, through the API layer.
Potential Issues
The horizontal data layer will need to be built in a cloud environment with unlimited storage and a pay–as–you–use service model, which also has enough power to find the insights and build the applications. The argument that it wouldn’t be safe enough to store all of this data, or be able to transfer it to the cloud, is just a matter of properly setting up the cloud infrastructure and making the secured on–premise connections. This will not only ensure network security but a good data governance structure for your data.
Examples
Breaking down these data siloes will allow an intelligent and efficient collaboration across departments from start to finish.
If you are working in manufacturing, it will be beneficial to capture your machine data and combine it with supplier data, machine manufacturers’ data and production quality, to improve your overall production efficiency (OEE). Looking for these relationships could bring you benefits for predictive maintenance, enable automatic quality improvement, improve supply chain, and provide additional product traceability, etc.
Putting this into practice in retail may mean you benefit from collecting different information streams, such as customer data, supply chain data, environment data (such as traffic data, crowd control data). All of these data streams can help you improve aspects such as your sales strategy, stocks, supply chain and marketing, as well as produce new revenue streams.
For building owners, breaking down your data siloes means you will be able to collect data from sensors and installed equipment in the building (such as HVAC, entrance scanners), as well as environmental data (weather, traffic) and crowd control data. All of this data could help you to improve the occupation of the building, air quality control (automatic steering of ventilation), and reduce energy costs.
Conclusion
Every industry could benefit from building a central data platform, but this should only be done from a well-constructed base infrastructure. This involves not only copying data to a cloud storage account, but also managing the network and data security, data governance, access control and more. Here at Codit, we can help you to implement a well-architected data platform that can bring value to your business and customers.
Subscribe to our RSS feed