Disrupting rather than innovating

Robots, image data and new digital supply chains are only three examples of developments that will also determine the market in the new year. Harrij Schmeitz of the Dutch Fresh Informationmanagement Center, GroentenFruit Huis’s knowledge centre, emphasises it’s not about innovations, but about disruptions. The biggest challenge isn’t renewing existing processes, but developing new processes. “We have the techniques, but how do we apply them?” Using examples, he outlines the factors that will change the market.

The difference between innovating and disrupting is how the market is changing. An innovation improves an existing model, a disruption is a new approach, or a new model, Harrij explains. “We’re dealing with a new reality,” he says. “You have to compare it to the impact of Uber on the taxi world, despite the fact that Uber has no taxis of itself. Airbnb is now the largest hotel in the world, but doesn’t even own one room. And Dutch company Thuisbezorgt delivers the most meals to people’s homes, but doesn’t have its own kitchen. ”Supermarkets-without-stock, such as Picnic, and the rise of meal boxes are examples of disruptive developments closer to the sector.

“A meal box isn’t just a new sales channel, it’s also a completely different approach of the supply chain,” Harrij says. He exemplifies: the supplier of a meal box wants leek that meets strict requirements during a certain week. A leek that’s too long doesn’t fit in the box, a leek that’s too heavy doesn’t comply with ratios in the recipe. “One week you can supply 50,000 leeks, and then none for weeks, because the meal box doesn’t need it.” This requires a different approach from the supply chain.

A similar image can be seen on the online market. “Online sales are gradually growing, but we have to make sure we don’t lose the market,” Harrij warns. Although the number of online sales is increasing, it doesn’t mean the consumer chooses fresh produce when making online purchases. Besides, packaging has to be adjusted for the products of this online sales channel. “I think strawberries in a traditional paper punnet are the best example; you can’t sell these online. If the packaging doesn’t have a lid, the strawberries will be all over the box once they’re delivered.”

“We have to go along with that trend, and that also requires data quality,” he continues. “It isn’t just the availability of data, but data also has to be the right quality.” Another example: the ‘Kies Ik Gezond’ app (Am I Choosing Healthy) launched by the Dutch Voedingscentrum (the Dutch national dietary centre) earlier this month. “It’s about the quality of the data in that app: if your article isn’t listed in the app, it says ‘article not found.’” Supermarkets therefore dedicate themselves to availability of this data. The sector could be paying more attention to this, according to Harrij. “The sector isn’t quite aware of it yet. Labelling often occurs somewhere in a warehouse, and they print what is asked by the customer. If the label says ‘store somewhere cold’ and the app says ‘store in refrigerator’ there’s a difference in data.”

He calls on companies to stop the discussion about ownership of data and the price tag data should come with, and to put more energy into the ways in which that data can have an added value. “When you say ‘the data is mine’ you must be dreaming, other people are already earning money off your data,” Harrij says. “That train has left the station.” It’s more important to decide what sort of added value the data can have, but first experience has to be gained. One example is the production data of growers who work with a certain variety, which can be important to agriculturalists, but also to the grower. So then the question is whether a price tag can even be put on that. Another example is the data collected by sorting machines. “I’m certain not one grower or trader has contractually established he owns this data. The data is often used by the machine builder to optimise the technique, but much more can be done with it. This is a difference between data ownership and data availability.”

For production, image recognition will also become a major disruptor, Harrij predicts. “Image recognition is taking flight. This development is going incredibly quickly, take Google’s face recognition software, for example.” For growers, a comparable technique, however, can have just as much impact. What if the camera system can recognise certain things in crops? What if cameras could determine the perfect time for harvesting? Or if they could inform growers about the presence of flies in part of the plot, for example? And taking it one step further, you’d find a camera that has learned to distinguish harmful fruit flies from harmless ones, and is therefore capable of raising the alarm in time. “The question is whether we’re using the opportunities in this field,” Harrij continues. “We have a good knowledge of production, but how are we going to use this knowledge for technological disruption?”

Another persistent discussion is robotisation. The asparagus robot, the apple harvesting robot, the tomato leaf picking robot and the strawberry robot will probably all be marketed in 2018 or 2019. “It’s an interesting thought that the robot could lower labour costs, but perhaps a grower is also capable of giving a better harvesting prognosis because of the robot, or it might become more appealing to grow a different variety with a higher value added. Another entry in the books would suddenly increase.” The question is whether current production methods are suitable for robots. “The way of thinking about robots has to change. The first question growers always ask is how much can the robot pick? A robot can’t look past the leaves, but it could still be interesting if part of the staff were to become redundant. We have to do things completely differently to discover new types of added value,” Harrij says. “Perhaps you have to grow a different variety, or plant the plants further apart.”

A robot works differently, and that can have far-reaching results. Take the asparagus robot, for instance, which drives around the field three times per day to pick asparagus. “It’s possible a robot leaves an asparagus, because the robot knows it’ll drive the same way in four hours. The asparagus is then a bit longer when it’s harvested, but won’t have a green tip yet, and it’s therefore harvested later than if a person were to harvest it.” Another example is the cucumber. “That product grows so quickly it can grow from a small size to a large one within a day,” Harrij explains. “A robot coming by three times per day can harvest the cucumber when it’s the perfect size.”

A short detour to ornamental plant cultivation illustrates this even more. The harvesting moment of roses requires much precision, and it costs growers much money to train staff to determine the right time. Besides, roses are always harvested too early or too late, which results in loss. The sector is therefore experimenting with image recognition, applied in glasses, that tell the harvesting staff exactly which rose can be harvested and which has to stay on the bush for a while longer. “We need fewer green thumbs and more systems. The vertical farm is another example of this.”

“Unmanned fork-lift trucks aren’t standard in our commercial enterprises yet, but why not?” Harrij wonders. The techniques for this are applied in other sectors, they just have to be translated into the fresh produce sector. “We have to look outside more to see what is happening, and then translate that for our sector. Why wouldn’t a robot be capable of stacking pallets if it is capable of sorting empty boxes?” Harrij refers to developments at Google and Amazon, which are frontrunners in the development of unmanned DCs.

That also requires a change. “With automatic fork-lift trucks you don’t need drivers, but an operator,” Harrij continues. He thinks the sector is too hesitant in appealing to IT workers. “These developments can’t be outsourced to technology companies, you personally need more knowledge. The IT department is often located somewhere behind the cold store, but it should be the CEO’s responsibility.”

More information:
Fresh Informationmanagement Center
Harrij Schmeitz