The next revolution in grading is not about rejection, but about insight

‘For years, the industry invested millions in removing bad products. The next challenge is understanding the good ones.’

Over the past ten years, the onion and potato industry has made enormous strides in automation. Driven by rising labor costs, a structural shortage of staff and increasing quality demands, packers worldwide invested in optical sorting technology. In particular, the rise of relatively easy-to-implement pre-grading systems meant that manual quality control at many sites could largely be replaced by cameras and software. For many companies this marked an important step forward: more consistent product quality, lower labor costs and higher processing capacity.

Now that this first wave of automation is maturing, a new trend is emerging within the sector. It is no longer so much about how well a machine can reject products, but about how much data, and therefore knowledge, a grading line actually generates. Because although modern systems keep getting better at detecting defects, it turns out that many processors ultimately gain only limited insight into the quality of the millions of onions and potatoes that flow through their installations each year. They know how many products are rejected, but often not exactly why. They know the end result, but not always the underlying causes.

That is remarkable, because every rejected onion or potato in fact represents a data point. Behind every defect lies information about growing conditions, storage quality, logistics, supplier performance or seasonal influences. When that information is not captured, an important part of the potential value of automation is lost. The installation may sort the product, but it does not collect the knowledge. It is precisely here that an important shift now appears to be taking shape. More and more companies realize that the economic value of a grading line is determined not only by how much labor it saves, but also by how much insight it delivers. In other words, until now many companies have underestimated the potential of using data from their grading process intelligently.

This development is reinforced by the rise of technologies that can fully analyze individual products. When every onion or potato is assessed separately on weight, size, shape, external quality and, where possible, even internal quality, a fundamentally different picture of the product flow emerges. The grading line thereby changes from a machine that separates products into a system that measures quality. A new quality system, then, that inspects 100% of the onions and potatoes non-destructively and with high reliability, instead of a quality system based on sampling. This makes it possible to guarantee the quality of the entire harvest and to prevent or minimize waste. It offers new ways to reveal differences between growers, plots, storage facilities and sales markets. On top of that, it makes quality development throughout the season measurable rather than merely presumed.

In the area of internal quality measurement in particular, many experts expect a major breakthrough in the coming years. Historically, the inside of an onion or potato remained largely hidden until the product was cut open or consumed. As a result, quality problems often only became visible after economic damage had already occurred. Technologies that detect internal defects earlier make it possible to shift quality management from reactive to predictive. This creates not only more control over product quality, but also over yield, customer satisfaction and risk management.

The question increasingly being asked in boardrooms is therefore slowly changing in character. Where investment decisions were traditionally judged on labor savings, attention is shifting to a different KPI: how much better do we understand our product after this investment? That may seem a subtle difference, but its effect is significant. Labor savings ultimately have a natural limit. Insight does not. Companies that understand their product flows better can steer more precisely on quality, storage, logistics, market segmentation and yield.

The agri-industry thus finds itself at an interesting turning point. The past few years have been dominated by automation. The coming years appear to be mainly about using data from the grading process intelligently, and with it building fundamental knowledge. Not sorting faster, but steering smarter. Not rejecting more, but understanding better.

‘Perhaps the most important question for the coming years is therefore not how many tons of product a grading line can process, but rather how much knowledge it has gathered once that ton has been processed. The winners of tomorrow are not the organizations that process the most products, but the companies that learn the most from every product they process.’