Nexperia

Smart Manufacturing, Industry 4.0, digital transformation, AI & ML – probably are the words that one can hear in every semiconductor company board rooms and read in strategy documents in 2022. Sectoral leaders have found value in these already in last decade and some have increasing efforts and focus post 2020 Covid era. Today businesses are increasingly focusing on the elusive and difficult to achieve value on a transformative basis in these themes. Though good value has been already gained from one-off, use case based adaptions so far. These use-case based adaptions already prove the massive potential of these themes in our industry.

This article covers on how to convert this one-off success into a transformative gain focusing on value from data.

Artificial intelligence, machine learning are tools that are available for us to use today. These are disruptive technologies and hence would produce higher value but value is elusive as well. Usual suggested way to look at adaption has been to prepare use cases and trying to solve them. Say predictive revenue or forecast etc. I believe this approach won’t yield transformative value. Since these are disrupting tools – we must look at history and look at truly successful disrupters like Microsoft Excel / PowerPoint. Sorry, this is not a sexy example, but you will know what I mean shortly.

Excel has replaced painstaking collection of data on paper, hand drawing graphs, analyzing business problems completely - across companies, across globe like no other Software ever has. Today you won’t find a company trying to analyze data by collecting data points on paper, nor making old style projector slide shows. If you look at this transformation a massive value creator of the past – at some point in time, companies in their job description would have made knowledge of Excel compulsory across roles from factory engineer to accounting analyst at finance. Our hiring / talent management / training / compensation all have had changes to make sure that excel as a skill is respected and used.

“I believe the moment for data has arrived. It's time to take this seriously and unlock its potential.”

I believe Data analysis as a skill needs to be a foundational skill for everyone in the company if we would like to get the transformational gain. Particularly covering how to define a problem, explorative data analysis, hypothesis building and testing&basic statistics. This is our first challenge- People & skillset to know how to get value from data.

Second challenge is an easier one – to establish a technology framework, set of tools that can truly be Microsoft Excel’s of the future for AI/ML. I say this as easy, because I feel there is a lot of option out there today that can help. Personally, I feel business intelligence is a fist step towards diagnostic models which can further become predictive or even prescriptive models. Hence a key selection is to believe in this process and select a tool platform that can help a person to go from BI to AI. Considering that different part of organizations would have different speeds, this is an essential need of the tool. Selecting tools that are built for Self Service AI/BI, that are open for rapidly developing and changing data tool sets is a must.

Last challenge is probably as big and important as the challenge of people, skill sets and mind seti.e. of Data. Technology partners might trivialize this with promise of data lakes, streaming analytics, bigdata etc. but transformational gain is much more cultural change than anything else. Let me give an example – if as a company we wrote our quality specs in word/excel documents with tables of data – a true BI system can not process the same. (Unless we do a hard work to digitize this unstructured data). Each department in the company need to figure out what’s the most valuable data asset for them and focus on getting that digitized – change processes, tools, mind set etc. Key thing is to also have a proper governance setup for collecting, storing, consuming, purging, archiving, securing data. As this is a massive enterprise, best would be to establish most valued data assets and start journey from there. If data is to be an asset, it needs to be treated as one.

Final part of the puzzle is organizing ourselves for success. I believe that each department must have a analytics team to catalyze the journey and to find out most valuable data, building governance together with core set of talent in IT. Our strategic programs like that of margin improvement, cost reduction, revenue growth, factory OEE/ Yield etc must have embedded data analysis teams to unearth hidden avenues.

Unlike other industries, semiconductor has always been known for its business complexity. We are hence, never early adapters of technology hype cycles. Our leadership don’t give into euphoria, rather focus on something that we can achieve now. I believe the moment for data has arrived. Its time to take this seriously and unlock its potential. Bon voyage for this journey!!!