14 sept. 2023 By David van Driessche
I'm better than this … machine!
Hands if you have said this more than once over the last week.
In all the projects we undertake, we aim to attain Connected Automation. More specifically, we try to replace tasks that demand strenuous human effort with Connected Automation. This is seldom challenged when we’re looking for a workaround for routine and straightforward tasks. Humans aren't particularly thrilled about sending individual confirmation emails for each and every incoming order. We gladly yield these thankless tasks to software.
However. When the task gets hard and is handled by "experts," the scenario changes dramatically. Experts take immense pride in their proficiency and informing them that software will now manage (for fear of writing “take over” here) something they've painstakingly mastered is a more challenging matter. That’s the moment when the thought that "we can outperform this (insert software brand name)" comes to the fore. An understandable reaction, for sure. But in reality, it often limits automation possibilities and, in the long run, hampers genuine progress, time savings, and improvements in production quality.
Rest assured, this phenomenon isn't confined to the graphics industry; similar debates arise in discussions about Wikipedia, AI, self-driving cars, and more. Let’s delve into a few examples and see if you can relate.
The Unbeatable Expert
The dilemma can be aptly demonstrated using software—let's consider InSoft Imp for this discourse. Imp is an ingenious program that takes a collection of jobs and generates efficiently imposed sheets, enabling these jobs to be optimally printed. It accomplishes this by incorporating data about the jobs, your preferences, and your equipment. It produces an optimized solution in the blink of an eye. So fast.
And yet, almost every single time, as soon as you delve into Imp's output, you can bet there is someone in the room who will jump in and go "Oh, this sheet could be improved! See, if I just rotate this and position this here, it's much more efficient than what Imp has calculated."
And, truth be told, that individual is most likely correct. Humans are adept at solving problems.
So, where lies the fallacy?
The fallacy lies in humans excelling at problem-solving in certain circumstances, particularly when dealing with small-scale problems. We’re brilliant at that. However, what we’re not good at is maintaining consistency, and our brain's processing capability easily gets overwhelmed as the complexity of a problem grows. And here’s why.
Firstly, it's been said that a goldfish possesses an attention span of 9 seconds. Research over the past two decades suggests that humans are performing even worse in this regard, largely due to the digital revolution and its myriad distractions. (By the way, if you're still engaged in reading this, congratulations— you're surpassing the average.)
In all fairness, these specific findings are subject to debate and might not have been based on the most robust research. Nonetheless, the general sentiment remains that our attention spans haven't notably expanded over the years. While we can refocus and concentrate on tasks, there's a definitive limit to our consistency in executing repetitive tasks.
Secondly, imagine I provide you with three jobs and ask you to arrange them on a printing sheet, factoring in their dimensions, colors, finishes, and the desired print quantities for each job. I have no doubt you’ll come up with a solution asap. And if you're experienced in this field, I’m convinced it will be an excellent solution. However, if I presented you with three hundred jobs and requested a solution within the next hour, what do you think the outcome will be?
This predicament is especially pertinent in our industry, where the prevailing trend over the past three decades involves printing more jobs with lower print runs per job. As a result, there are now many more different jobs in different volumes that need to be managed. How are we dealing with this?
The Law of Averages
Imagine introducing a fully self-driving car. After testing, you discover that, on average, our car causes 10 fatalities due to accidents per year for every 1 million cars on the road. Instinctively, one might assume that improving software quality and focusing on quality control are imperative before releasing this potentially hazardous technology on the streets.
However, in the EU, in 2021, the annual average number of traffic-related deaths per million people (not even per million cars!) stood at 44. Given the EU's population of approximately 253 million in 2021, our car would have conservatively saved around 8,349 lives.
A parallel example is evident in how we react to airplane accidents. Despite the extensive media coverage such incidents receive, statistical analysis unequivocally demonstrates that air travel is an extremely safe mode of transportation. In fact, it's notably safer than driving to the airport in your car (even if it’s one of our self-driving cars!).
This tendency to focus on exceptions and mishaps is also evident in our printing industry. Truth time – don’t you spend more time dealing with outliers rather than enhancing your overall process, including jobs that function seamlessly? (I thought so.)
Beating the Average
When contemplating Connected Automation, it's crucial to bear in mind that it doesn't need to outperform your top-performing employee when they're fresh, focused, undistracted, in optimal health, and eager to demonstrate their superiority over software. You can’t automate authenticity and fantasy is a foundation.
The true yardstick lies in whether your automation surpasses the average performance of all individuals performing a specific task. This includes efficiency and error rates. By using this metric to calculate ROI, you can figure out which improvements to make.
It's worth noting that I mentioned "calculate ROI." Given our tendency to fixate on anomalies and mishaps, keeping robust analytics about your business is vital. Rely on measurements rather than instinct when making decisions.
Good analysis shows where things are stuck, helping you to focus better. It also indicates the efficiency and effectiveness of your improvements.
Elon Musk once shared in an interview that they attempted to employ robots for covering battery packs with protective blankets during Tesla's production. They assumed that robots would be faster and more dependable than humans. Yet, upon reviewing the entire process, they discovered that the robots were not optimizing the production process, they were hindering it! Needless to say, they switched back to human involvement after that.
My point here is, don't shy away from critically analyzing the data you are gathering. Regular evaluation and assessment will assure you that your operational methods remain optimal.
Leveraging Expertise When and Where needed
Undoubtedly, your organization can rely on some bona fide experts. These individuals are pivotal to your success, so make sure they get to apply their expertise where it matters the most. Automate routine and simple tasks, free up your experts to concentrate on stimulating endeavors. Depending on your sector, this could range from managing intricate, bespoke jobs or offering exceptional customer service that differentiates you from competitors. Liberate them from monotonous tasks, enable them to truly shine, optimize your business bottom line!
It’s about finding that delicate balance between human and automated efforts. If you can get those working in sync, you are in the sweet spot of true business efficiency.
Just to prove that I put my money where my mouth is. The content of this blog is mine, I wrote the first draft. Some parts were enhanced by ChatGPT (can you spot them?), and our marketing team did a final proofreading check to make sure the overall tone of voice still feels like Four Pees. ChatGPT really did a good job, but it does not convey the expertise and the nuances we wanted to put out there. Its assistance saved me 4 hours of copy work, time I gladly outsourced to our proofreaders. That too is Connected Automation.