September 24, 2019 Manufacturing Must Be Data-Driven

Supply chains have been slow-moving in terms of innovation. Manufacturers can only become data-driven after a systematic transformation.

A data-driven approach significantly improves throughput and adds to the bottom line.

September 23, 2019

One of the most common misconceptions about manufacturing is that it’s data-driven. People assume that because supply chains are complex and precise, they must be the products of extensive, ongoing analysis. In reality, supply chains have been historically slow-moving, especially in terms of innovation.

That fact is borne out by Deloitte’s Industry 4.0 survey asking decision makers what business functions are driving digital transformation. Only 34% identified supply chains, far less than those who identified operations, production, or IT.

Reluctance around supply chain innovation is unfortunate. The goal of any supply chain is to maximize throughput, yet past and present approaches cannot track performance in real time. These issues will only be exacerbated as supply chains become more interconnected. Thankfully, they can all be mitigated with data.

Why Data Is a Unique Asset for Manufacturers

Data is one of the most important assets for all companies. It’s clear that a data-driven approach doesn’t just produce vague “improvements” — it significantly improves throughput and adds to the bottom line.

PepsiCo discovered this after implementing analytics software to help it track how it distributed soda flavors. Data revealed how to fix the problem of shipping so much product that it ended up expired before it was used. After adjusting the balance, PepsiCo reduced shipments and, therefore, waste.

Amazon is another compelling case for what data can do for supply chains. The company has invested heavily in automation in order to turn its fulfillment centers into cohesive ecosystems. Thanks to the data-driven innovations, Amazon has cut supply chain costs and achieved annual profit growth.

With those numbers, one would assume manufacturers would want to become fluent with data. However, digital transformation often fails because of a lack of buy-in at the top. Leadership often sees advanced technology as a risk. There’s also fear that newer technologies will replace workers. Therefore, only 31% of companies consider themselves truly data-driven, which is down from 37% two years ago.

These anxieties are unfounded. When factories use data to improve processes and artificial intelligence to accelerate efficiency, those factories become more competitive. Research shows that manufacturers have committed $907 billion (5% of revenues) to improve connectivity and that 72% expect to be “digitally advanced” by 2020. Therefore, embracing data is essential for remaining relevant.

Manufacturers will need to overcome several hurdles along the way. First and foremost, the company needs the right technology stack and the staff to manage it. Beyond that, it needs to improve access to data so decision makers (not just IT) have the most accurate information. Finally, whatever technologies are put in place need to be dynamic enough to incorporate new types of data and analysis as necessary.

5 Steps to Becoming More Data-Driven

Manufacturers only become data-driven after undergoing a systematic transformation. Follow these steps:

1. Investigate time bottlenecks.

Lead time, more than any other variable, reveals where true problems exist in the supply chain. Look for bottlenecks, then investigate how and why they happen and what impact they have. Optimizing manufacturing requires many things to be fine-tuned, but eliminating persistent delays and lost time is the priority.

2. Draw on existing data.

Even if manufacturers need to collect more data from more sources to gain true insights, they already have data they can begin analyzing. It could be financial, operational, or physical — all of it contains insights that might be relevant to process engineers and continuous improvement experts. Working with available data helps companies cultivate their capabilities for the “big” data coming later.

3. Use AI to search for insights.

Collecting data is the first challenge; finding the insights within that data is the second. AI can aid this effort because it’s smarter and faster than humans. Analytics driven by AI has been shown to improve order-to-delivery cycle times by 425% and supply chain efficiency by 260%. Compared to the alternatives, AI makes it easy to begin leveraging analytics effectively.

4. Expose the unknowns.

The majority of details related to operations are unknown, even at the world’s leading factories. Data should be collected from sources that can illuminate these unknowns. Installing connected sensors is an ideal way to learn about previously opaque processes.

5. Keep things in perspective.

As manufacturers become more fluent with data, it’s tempting to become as tech-driven as possible. However, fully automated manufacturing is only an asset for some companies, namely those with predictable demand. In companies where demand is dynamic, automation is less of an asset. Every technology should be evaluated based on whether it delivers actual business value rather than just advanced capabilities.

People mistakenly think manufacturing is data-driven because logically it absolutely should be. Decision makers are discovering this at the exact same time that technology is finally making it possible. It’s an incredible opportunity, but soon it will become an industrywide obligation.

Bhaskar Ballapragada

Bhaskar Ballapragada is chief architect at ThroughPut Inc. Bhaskar leads product- and technology-related initiatives and helps companies detect, prioritize, and alleviate dynamic operational bottlenecks by applying machine learning algorithms.

Subscribe to Industry Today