Within the manufacturing sector, data has proven to be a driver of innovation — from deciding the next generation of automobiles to the evolution of smart medical devices.
For this reason, enterprises have embraced technologies that harness the power of data, be it IoT, automation, or AI.
Nonetheless, there are still many possibilities that manufacturers have yet to unlock. To this end, senior IT leaders from manufacturing gathered to explore such opportunities during the “Unlocking the Potential of Data in Manufacturing” roundtable, organised by Jicara Media, and hosted by Rackspace Technology and AWS.
A couple of years back, digital transformation used to be just a buzzword, the title of a project, or capital requests that were taken to the board of directors, recalled Hemanta Banerjee, Vice President, Public Cloud Data Services, Rackspace Technology.
“Now, digital transformation has gone from being a project to being something inherent in how businesses work,” he said.
The executive then outlined three factors that impact the success of an enterprise, in terms of transformation:
- Public cloud
- People
- Data
“The primary among these is data because it lets you understand what is happening in your business process, and lets you optimise that. When you find that out, you get to have a positive feedback loop around how to improve. (The challenge is) how do we use data to improve your manufacturing processes, your supply chain challenges, and so on,” Banerjee noted.
Data challenges
While many manufacturers have already leveraged what data has to offer, still there are a few more roadblocks that are keeping executives up at night.
Common among the manufacturing industry, it would seem, are the challenges posed by data residency, and centralisation versus decentralisation.
“Previously, everything’s about consolidation, so we can make use of things. Now, we are looking at how we decentralise, yet are able to achieve targeted outcomes. Due to data residency challenges, we have to be mindful. Questions around ‘How do we bring data back?’, ‘How do we organise across teams?’, and ‘How do we aggregate?’ are answers we seek,” said Isaac Tan, Director for IT, APAC, Hologic.
“The challenge that we currently face is collating the data from various sources,” shared Azuhar Mohammed, Head of Global Solution Centre Expertise & Innovation, Sanofi. “We have a lot of systems in different countries, but we are trying to deploy our solutions to standardise the data. The next step is gathering our data from various sources such as sales, marketing, finance, and HR, taking it to one place like a data lake, and then using that data with intelligent analytics for the business to make a meaningful decision.”
Meanwhile, over at a Singapore-based medical equipment maker, data integration is among the current areas of focus, noted one of its senior IT decision-makers.
“I feel it’s important to get the integration of data because, on this journey of transformation, we learned that we have quite a number of siloed departments and siloed data. How to integrate, from finance to operations, to supply chains, that is one of the real challenges that we have,” the executive said.
Transformation journey
Despite the setbacks, many manufacturers have already embarked on their digital transformation journey, making the most of data opportunities offered by new technology.
Sanofi, for example, has already explored RPA as early as 2017.
“Now, we are using ML and AI to build an intelligent decision-making process. The AI is being applied in modelling and pattern recognition using a large amount of data to enhance our R&D productivity,” Sanofi’s Azuhar Mohammed said.
For Isaac Tan, while the scale is a challenge, it has not deterred Hologic from its IoT push.
“We are very focused on getting all the devices that we place in our customer sites connected. So, from our connectivity standpoint, it is critical to have more data points, more sensors in place,” said Isaac.
However, the medical device manufacturer didn’t always have its data strategy all figured out in the past.
Isaac elaborated: “A couple of years back, our approach was ‘Just get all the data in, then we’ll make some sense out of it.’ Over time, we realised this was not working. The IT team will never be able to meet the diverse set of demands coming from stakeholders worldwide. How do we enable everyone rather than just give what they ask? That’s how we handled it. We thought about decentralised or distributed data products. We are treating them as separate, but we are also putting in a framework and certain core technology foundational components that will then allow them to build faster.”
Isaac also recommends having a common data model put in place, which can be used as a standard that people can develop from and can be used when somebody needs it.
Indeed, digital transformation projects are not a one-off undertaking, usually taking place over several years, Rackspace Technology’s Hemanta Banerjee observed. “It’s a long journey. It’s not a project, but rather a way of life,” he said
Data optimisation strategy
Although enterprises are at different stages in their data journey, Rackspace Technology has identified data silos as the usual origin of customers’ headaches.
Shwetank Sheel, Director, Data Services Sales – APJ, Rackspace Technology, noted that the first step in breaking down silos is identifying and bringing them to the surface.
“Others have spoken about having the combination of centralised and decentralised data, but in the context of curated data lake along with decentralised analytics, one of the things that I’ve seen a lot of customers talk about — from a governance perspective — is the idea of a catalogue or a stream of data products. Therefore, when you’re trying to drive innovation, which is either across countries or across business units, and you want to have an index of what already exists, it’s stepping up on what’s already there, as opposed to starting from scratch,” he advised.
In terms of determining the ROI, Sheel said that organisations must first confirm that the data that they need exists internally.
“Focusing on use cases or outcomes, (we usually) work backwards with users to help them understand what’s possible with the data, and then try to put into place a framework which helps them and the organisation identify what is the value of a particular use case, versus the cost of doing it, versus whether the data even exists in the organisation, so working through the data, all the way through to be able to do advanced analytics, AI and ML,” Sheel explained.
Ultimately, what has proven to be successful thus far, is the distribution of analytics to business units, noted Banerjee.
“What we’ve seen is, we’re moving away from centralising everything, having a data team, and the data team doing stuff and people queuing up to get the data. What we see is that those organisations that are successful, actually distribute that function out onto the businesses, and the central team provides frameworks, guidelines, and infrastructure for the businesses to be successful,” he concluded.