Is this machine the end of quality control issues?
Almond growers, processors and exporters get paid based on their product quality. Until now product quality is determined by the hands and eyes of quality inspectors – at different stages in the supply chain with different people, different views and different results. California-based Qcify is working to take out the human factor in quality assessment for almonds. Both the ABC and AlmondCo are testing it. If all goes well the machine could replace manual assessment and become an industry standard. We talked to Raf Peeters, Founder and one of his clients, Brenton Woolston, Brenton J. Woolston, Managing Director at Almondco Australia Limited.
The Clipper: What are the benefits of automated sample assessment?
Raf Peeters: Consistency. If we have 5 people looking at a sample all of them might get different results. And the results might even be different with the same people on a different day. The machine does not change. It’s an objective way of generating data. There will always be a margin of error with human assessment unless we start becoming robots. Manual assessment also has limitations because of increasing labor costs. And the production volumes are increasing. If you want to increase your processing volume from 10,000 pounds per hour to 20,000 pounds per hour you have to double your staff for sample assessment. The same holds for the government entities like USDA. They have to have a massive workforce for the sampling. We hope that USDA will use the technology for their reports in the future. When they understand that the technology will help them to save on staff and speed up the whole sampling process they might become customers. Taking out the human factor of assessment is one of the most important aspects of the technology.
The Clipper: How does assessment work at AlmondCo today?
Brenton J. Woolston: We are very focused on quality. We do a lot of testing in the sorting and packing stage of the supply chain. And then we check the finished product. Every customer delivery gets checked. We do not have a USDA check so we present the report to the customer. Every delivery is QC-assessed or we call it ‘certificate analysis’. The quality team is currently the biggest department in our company – about 20 people. All the sample product has to be bagged, labeled – and put into a retention sample library. If there is any problem with the delivery you can go back to the library and have another look at the sample. But the samples are in storage for almost a year. With this technology you can digitally store and send to the customer to verify the quality that was shipped.
The Clipper: Would you reduce your quality control staff if all works well?
Brenton J. Woolston: Until now the whole incoming and outgoing quality is manually assessed. And there are always discussions. I think the machine can eliminate that. We would definitely reduce the staff because with the growing workforce you become more inefficient, not more efficient. And the labor costs are high. With the technology you take samples from the processing line and put it into the machine. Within minutes you get a QC report.
Raf Peeters: In California there are third party entities like USDA and DFA of California who do quality assessment. And the client request that kind of assessment because they want to have a ‘neutral’ position. They do not always trust the supplier because they could tweak the numbers. Qcify offers a ‘third party inhouse’ if we can establish a system in which the machine is seen as a third party validator. And perhaps USDA or DFA can certify our system and have it used by their staff, which will ultimately save money for them and the processors.
The Clipper: Would you use automatic sampling to warn the grower in case the quality is decreasing?
Brenton J. Woolston: We do an economic analysis for the grower. We are trying to have as much product in the required quality range of the receiver. If you are able to lower your insect damage from 1 percent to a half percent that might give you another 100,000 US$ of value.
Raf Peeters: The data of every single sample is available. It can be linked to every grower. Processors can decide what information they want to share. If somebody has a claim or a request you can send them all data for review. Some companies might not want to share information. The machine stores reports, it stores all imagery and QC data and the customer can decide which information to share to single or multiple users. The information can go upstream to the growers but it can also go downstream to the end customers. Some end customers have also installed the machine. So instead of providing a written report or a USDA report processors might just send the Qcify report – not all the images but the cover page with the results.
The Clipper: If there are conflicting claims about quality: How would the process look like?
Sometimes end customers try to argue that the delivery did not meet the specs. But it might just be an attempt to knock off a few cents of the established price. Now you can say: Let’s have a look at it. Here is the report with the data. You can follow a link to the data of the sample. And then the end customer gets access to a 1,000 to 1,500 images. It will be much harder to make false claims then. We see it as a data collection system but also to a certain extent an arbitration tool that can go upstream to the growers or downstream to the end customers. That was missing in the past. Processors used to store the samples for a year or longer in cold storage – and it used to be a great effort to find the sample in case of a claim or a request. Also the almonds in the sample degrade over time – and look very different. Sometimes the samples are thrown away or put back into the line. With the QC report and the images you do not need to keep the samples anymore. Everything is in the database.
The Clipper: Talking about arbitration. Would it make sense to accept machines like this as a basis of contracting? Like a standard?
Raf Peeters: 2017 has been the first year for the machines to be installed in Australia and California. People have been doing QC in the same way for decades. It takes a certain amount of time to get to a stage where they start trusting the technology sufficiently. And there is a need for a structured way to share data that has to be met.