Analytics: How to avoid the top hindrance to (re)building your business

Manufacturers and others working through the COVID-19 crisis and planning to regain their competitiveness will rely on data analytics more than ever. But they’ll need a smarter, more strategic approach to prevent the most common cause of failure. Fortunately, the problem is more human than technical in nature, and it’s fixable.

The Fourth Industrial Revolution or Industry 4.0 brings among other things IoT data visibility, fueling more and better data modeling, machine learning, deep learning and Big Data. For purposes of this discussion, I’m committing a transgression here, lumping these together as “analytics” despite the technical distinctions. In reality, things just keep converging so that instead of plain-old IoT, more attention is going to the newly-coined AIoT, or artificial intelligence of things.

When deployed in tandem, artificial intelligence (AI) and the internet of things (IoT) can bring powerful new capabilities and competitive advantages—a net effect that’s greater than the sum of its constituent parts. How much more powerful AIoT is than vanilla IoT at tackling organizational improvements? Take a look at the percentages inside the little orange circles:

AIoT is 39% to 45% more effective than IoT, per a survey of 450 business leaders. (SOURCE: SAS, Deloitte, Intel and IDC)

Those surveyed said that among the benefits sought for their IoT efforts, increased revenue topped the list, regardless of geography, industry or company size. AI turbocharges that effort. (More details in the full article.)

AI, analytics: old news

Since my early-1990s coverage of early analytics and AI in such applications as statistical process control and machine vision, solutions have evolved with compute power and cloud services. The practice can now reach data anywhere, including networks bridging sites within single sites, across multiple sites or throughout enterprises and supply chains.

Ceci n'est pas une cobra
Magritte, meet pachyderm: “Ceci n’est pas une cobra”

In seeking improvements, top officers of the company need to instill a pervasive culture of improvement. True, improvements can be made in the silos of R&D, engineering, production, sales, finance or other functions, but still should be guided by market signals and customer demand, not just the blind pursuit of efficiency or cost-cutting. Without that holistic approach, line managers down can fall victim to the parable of the vision-impaired persons and the elephant. In short, things aren’t always as they seem.

If such talk seems a generality, it’s also very real. Plant managers, for instance, can achieve operational excellence, but so can their counterparts in competing organizations — the technology is a known commodity available to all. This diagram from ARC Advisory Group analytics specialist and VP Mike Guilfoyle helps illustrate the scenario — which gets more play in a Smart Industry story (hitting print and online channels at month’s end):

Transformational improvement with analytics can come when market signals lead internal improvements. (Source: ARC VP Mike Guilfoyle)

Where are the gaps in your apps?

The market’s flooded with analytic sellers and solutions. Here are for deceptively simple considerations that might help uncover how to find growth opportunities; where to find the greatest needs; how to approach solutions and who will support a cycle and culture of improvement:

Imagining I were in your shoes, I came up with a thought exercise with four very basic questions that might be a good first step closing the opportunity gap in your application of analytics:

1. In every sector and at every level, organizations generate massive amounts of presumably (or potentially) visible IoT dataso where are most data generated throughout your entire digital universe that you can, or should be measuring?

2. For every area that can be measured, analytics solutions (or platforms) can be bought, adapted or created to yield greater insightso where are there gaps in your management and automation systems across your value chain(s) that if filled, will yield the greatest visibility for tracking and potential improvement/innovation?

3. Every business and operational system in the IT/automation marketplace must include some provision for the use of analyticsso how capable are your systems to perform analytics via native functions, third-party partnerships or via a suitable level of standard integration capability?)

4. Every application of analytics requires a sustainable ecosystem of training and support for better decision-making and ongoing innovationso do you have a sustainable culture of improvement, with suitable lifecycle support structures in place via in-house and/or reliable external resources?


How are you approaching analytics?

Can you help me support the cause of progress? Share your thoughts and anything allowable…I’ll work with you and your organization to see how we can get the word out.

The eye in the sky, now on the edge

Discussing digital transformation with a brilliant mind in the field of digital transformation recently, I asked: “What emerging technology do you believe will be most transformational in the industrial sector?” His answer: Video.

His focus on video stems from the growth in IT bandwidth and the ability to support video with AI-powered analytics to more effectively discern patterns when data is deployed across IoT networks. AI and the Internet of things (IoT), or the combined AIoT, is deemed critically important to digital transformation initiatives, industry leaders say. Where AI analytics are deployed depends upon the challenge to be addressed.

As an AI-focused executive with a leading firm connected to these converging technologies, the gentleman I was speaking with was fascinated by a demo he saw at a recent trade show. It depicted a factory with thousands of cameras installed, one above each employee workstation. The goal, he said, was “to effectively use a camera to ‘sensorize’ a person [and] figure out if folks are deviating from the agreed processes.”

Businesses of all types — from small machinery suppliers to large manufacturers (like Foxconn) have for decades used cameras and machine vision software for applications such as quality assurance, to recognize patterns and detect and reject off-spec products. It’s really no surprise, but a logical extension of what Henry Ford did in automaking; what W. Edwards Deming did in post-WWII Japan; and what Ray “McDonald’s” Kroc did in fast food.

What may be surprising how real-time video analytics can transcend the mechanistic aspects of how people act to discern how they feel. In video and audio applications, machine learning has already been applied to customer service call centers and doctor-patient telemedicine apps to identify such things as anger and frustration. The data inputs for A/V analytics include voice analysis, facial expressions, movements, and posture. Industrial applications are no doubt coming, if not already here.

The edge of progress

As market demand fuels technical capability, video analytics solutions will continue raining down from longer-term remote cloud applications to people-watching applications based in real-time edge computing networks on or near the production floor. Intel’s Chet Hullum and his team improved semiconductor manufacturing at the edge because addressing cloud latency issues “would be far too expensive,” as I reported for SmartIndustry. (Intel Tweet and LinkedIn post (below) give more background:

Logroll on the blogroll: Please to have Intel ‘social’ my writing.

With advances in connectivity and pervasive analytics, the eye in the sky really isn’t the limit anymore.


help me help you!