Hyperledger Foundation Hyperledger Global Forum 2020, 9 Mar 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: End-to-end Pork Meat Traceability with Unilever - A Live Enterprise Production Use... Ricardo Garcia

Description

End-to-end Pork Meat Traceability with Unilever - A Live Enterprise Production Use Case on Hyperledger Sawtooth - Ricardo Garcia, ScanTrust SA

Ricardo will present a live implementation of a project of ScanTrust and Unilever.

The presentation covers the following aspects:
- Why was the project started and what was the motivation
- What technology was used and what's the architecture
- How did the operational implementation work
- Live demonstration of the product
- Challenges faced in project

This project is a real implementation that goes beyond a proof-of-concept and is in production with real users. It's a great showcase to discuss, how we can bring more enterprise blockchain use cases into production to increase the adoption of blockchain solutions in general.

----
Hyperledger – Open Source Blockchain Technologies

Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use

Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory

Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter

Follow-us on Twitter @Hyperledger

Learn about Hyperledger Membership: https://www.hyperledger.org/about/join

#Hyperledger #HyperledgerGlobalForum2020 #Blockchain

A

All right, good afternoon, everybody thanks for coming out, despite all the corona panic so happy to be here today. I'll talk about one of our projects with Unilever, which has been in production for a few months and is a nice example of an actual use case and project in production using blockchain in Asia s, who we just had the session about blockchain adoption in Asia. This project is in Vietnam and is a nice example, tangible example for a real application.

A

So I'll talk about a few points, so, first of all, quick intro about who we are scan, trust. What what do we do, then I'll move over to the business background of the project. So what was the motivation for the project to start and I'll give an overview of the project in general? What was the use case and how do we implement the blockchain solution? How do we integrate the project into the operations and what did we learn from this project so quickly about us scan trust?

A

So we are a company headquartered in Switzerland, with offices in Shanghai, Amsterdam, Kiev and a few other outposts, and we've been active in the market now for five years and we are focused on digitizing supply chains. That's at the core. What we do giving each physical product a digital identity, to be able to track it in supply chain.

A

We have two main customer groups: one is food and beverage in the FMCG space and the other one are industrial good companies such as chemicals, automotive and precision industry and at the core there's three things that we focus on: first, one: establishing end-to-end transparency in a supply chain. Second, enabling touch points based on on QR codes, for example, with consumers at the point-of-sale and third is preventing counterfeits, so protecting your products from being faked and the company originally started out around one specific invention.

A

One patent that we have developed, which is was developed at EPFL, the Technical University in Lausanne, and what we have invented is a secure and unique QR code that cannot be copied. So that's a unique identifier with a special graphic in the middle that makes this QR code copy proof. So once you have printed this QR code, there's no way you can fake it because you can detect fakes with an image processing algorithm in in our back-end and around this patent and technology that we have developed.

A

We have also built a whole what we call smart packaging platform, because if you are looking at at counterfeits and traceability in supply chain, you also and unique QR codes. You also need a platform to manage these unique identifiers and that's the second element of our offering it's a platform that allows you to manage and generate unique QR codes.

A

So you have all the different tools that you need to do that integration with printers software to generate QR codes and tools to track when and where these QR codes have been scanned in what region different analytics, such as alerts. If a QR code is scanned in a market where it's not supposed to be scanned and and more, and we have two scientists system in a very open way such that it can also be integrated with earpiece or with blockchain solutions.

A

So at the core, we're not a blockchain company, we're an anti counterfeit and digitization company. But two years ago we started getting more and more inquiries from our customers in the supply chain, space for blockchain applications and that's why scan trust also started positioning itself and investing into the blockchain space. So we see blockchain as a separate tool that can be integrated with our core unique identifiers and QR code platform.

A

So a few few example applications of where our codes are on from seeds, vegetable seeds to water filters to spare parts to baby milk to pork, meat soups, which what we will be talking about today. So one of the projects that we implemented in the blockchain space was with Unilever in Vietnam, specifically with the brand called ignore that maybe some of you know ignore is one of the oldest and largest brands of Unilever, which makes all types of soups and one of those soups is pork soup.

A

So I'll talk about why this project was started and what the business rationale of Unilever was and then a bit more in detail about the actual implementation from a technical and operational point of view. So in terms of the business background, how did this project even come alive and what was the rationale for it? So in Vietnam, Unilever is the market leader for pork meat soups with noir, so they have consistently had the highest market share, but they have one competitor that is quite tight to them.

A

The it's name is ajinomoto, it's a Japanese soup brand and they are kind of racing each other in in the market, for who holds the top position and kannur is more of a premium brand, whereas ajinomoto is a little more.

A

Let's say: budget option for a similar type of soup and noir has always been the leader in the cities, but recently ajinomoto started catching up a lot of market share in the rural area and they started to lose market share, so noir started looking for ways of how can they differentiate themselves in the market and how can they increase the trust and perception of the rural population to purchase your brand their brand instead of the product of their competitor, so they started brainstorming around different ideas.

A

How can we, how can we differentiate our brand and make a stronger case for the rural population to buy our product and so to ask themselves? What is our next big thing?

A

How can we differentiate ourselves from from our competitor and that's how they came up with the project that I'll be talking about now, so the solution that they came up with was they looked at their their own product, the kannur product, and they analyzed what are like the things that people really care about and the things that can set them apart from their competitor and one thing, is they have a very strong claim in the market, which is that this Kannur brand is made from carefully selected clean pork farms, and this claim of our pork meat is clean, is something that their competitor cannot claim, because they don't have enough evidence that all the pork meat in their soups is actually coming from clean certified farms, and so they realized that this could be one of the arguments that they could use to differentiate themselves against their competitor.

A

At the same time, what they also realized is that this claim was not perceived that strongly by the rural population. It was received well by people in the cities who like appealed to this concept of clean pork and who also indicated that they believe that the soups have better taste because of that, but less so for the rural population. That, for some reason, didn't trust this claim enough.

A

So they realized that they have an opportunity to build trust and transparency around this clean pork standards and develop a messaging around food safety and around clean pork farms, and that's what they then decided to focus on. So they launched a whole campaign which was focused on this real, honest and authentic messaging, and this campaign was built on three pillars.

A

So the first one is they developed a new marketing and communication campaign where they communicated the superior quality aspects of their pork meat suits to their consumers and, as part of that, they also developed a whole TV commercial or checking gobut.

B

Menken of the continuity thing on port and dissected it yet who detention of battle quicken our github jam user come on and then doing.

A

B

A

Someone is having struggles here in the back. The second component of their campaign was smart packaging, so they brainstormed around ideas.

A

How can we use our packaging itself rather than commercials to communicate with the customers, and that's where the idea came up to use the packaging to communicate the provenance of the pork meat to the end consumers and show them that the pork meat is indeed coming from clean and certified pork farms in Vietnam, something that their competitor cannot say and the third element was they also improved some quality aspects of their product, and it is this second pillar here in the middle, where we as scan trust, were brought on as a partner for this project.

A

So I'll explain a bit how this solution was developed. That today enables consumers to scan a QR code in Vietnam on their nor suits and check all the different steps along the supply chain of the pork meat ingredient in the nor soups.

A

So the project was launched on 22nd July last year and has been live now since more than six months in Vietnam and there's approximately thirty million packs in the market, which have an individual QR code from scan trust linked to a blockchain based solution which shows where the pork meat in this soup is coming from. So it's not a POC.

A

In that sense, it's an actual in the mind of Unilever. It's still a pilot project because it's limited amount of time in one region, but it's definitely production scale in a sense that it's 30 million units that have the code in the market and there's real users scanning and using the solution on a daily base. So what we did is the first nut we had to crack is to actually map out the supply chain to get with Unilever, so there's three parties that are involved.

A

First of all, there's two pork meat farms in Vietnam that produce the porks.

A

Second, there are two meat processors, so these are facilities that take the pork meat and then slaughter the pork and process it and third, there's the Unilever soup production facility, which takes the pork meat processes it into powder and then packages it and chips it so before, even starting to develop anything.

A

The first exercise we had to do together with Unilever is map out the supply chain, understand who are the suppliers and what is the data that we are able to get and that we then also want to show to the consumer as a proof of claim that the meat that you are eating in this soup is coming from from clean pork farms.

A

So the the number of parameters that we identified were the foreign details and certifications that the farm has then the lot number and shipment details from the meat processors and the batch production batch ID of the soup powder that is produced in Unilever.

A

So the way the provenance as such is established is that there is a batch of pork. So a thousand pigs that get shipped to the meat processing facility and they get butchered on a certain date. So then there is a date and there is a lot number from this processed meat. Then this processed meat gets shipped to the powdering facility, which takes it and then processes it into soups and also has a date and and the batch number.

A

And then there are X amount of QR codes that get linked to this batch of produced soups and we link these different ids together. So the batch number, with the lot number with the farm ID and the QR codes, and that's then, in the end, when you scan a QR code, it retrieves this record from the blockchain that was previously uploaded.

A

So after we had done this exercise, we had mapped out a whole supply chain and then actually we had all the information available that we needed to develop the new solution. So why do we need plug chain here as an intermediary step and so to critically also assess that there were two arguments why blockchain was proposed as a solution here? The first one is a data governance question, so the supply chain in Vietnam is a very short one and there were two farms and two meat processors involved, and so maybe potentially like.

A

There is not such a huge trust mistrust issue, because these suppliers know each other. But if you want to make this scalable for the Kannur brand globally in in other countries, there's going to be thousands of suppliers and and farms that are involved, and so they chose to use blockchain as a technology to be sure that you have a solution that can scale to multiple suppliers from the start, without having a data, governance and and trust issue. And the second argument was a consumer trust argument.

A

So if you are using blockchain to prove a claim to a consumer, you would want this data to be immutable so that the consumer can trust that it hasn't been manipulated and the information he sees is authentic.

A

So technical implementation, how that is actually work, and how did we develop it? So there were two technologies that we used.

A

On the one hand, it was our own platform from scan trust, which was mainly used for the unique ID management, so every Canora, every pack of this 30 million su packs, has a unique QR code printed on it, which is an identity that lives in our system, and you can track and and manage in terms of you can see when it was scanned and where and you can adjust the content that gets displayed to the consumer and the second part is as a blockchain platform.

A

We used hyper ledger Sawtooth, mainly because it seems to be the hyper ledger protocol that is used for supply chain use cases. It's also embedded in the hyper ledger grid initiative. It promises to be scalable, and since we were dealing with 30 million products without scalability is something that is important and it has a global community designed for enterprise. So we chose this as a protocol one and a half years ago when we started with the project in terms of the actual technical implementation. This is a highly simplified architecture diagram.

A

So we the the technical solution as such, consists of from a very high level of five components. So, first there is an application interface. So this is an online portal web web-based solution that allows the actors in the supply chains of the meat processors and the production facility to upload the information about the produced pork meat. So the batch numbers lot, numbers shipment dates and, and so on, I'll show what that looks like in in a second. The second component is the consumer interface.

A

So this is the mobile web page that the consumer gets after he scanned the QR code and then the backend is based on our on the scantron system, where all the QR codes live and a private blockchain network based on hyper ledger. So it's actually a little bit more complex than here, because there's also a database involved, since, if you have 30 million QR codes, you cannot do queries to the blockchain in real time.

A

So you need a buffer in between so it the actual architecture is a bit more complex than what you see here, but from high-level that that's what it is and then there's also an integration to Unilever ERP system to pull some some of the information.

A

So this is the what we call the blockchain connector, so the portal that the actors can use to upload information here you see a couple of screenshots of what it what it looks like. So you have an interface to register farms and upload a new farm that is part of the system and manages identity.

A

You have an interface for the meat processors to record a new shipment from where, from which farm this came and which date it was, what production a lot number it was, and then there is also an interface for the a super production facility that can then upload the batch of soup 10 that was produced.

A

So that's like the interface for the actors in the supply chain and then the consumer interface I'll show a quick demo of this as well. This is the response that the consumer gets after. He has scanned the QR code, so I'll just this is now an example, QR code that was scanned.

B

A

You scan it and then this makes a request to our system, and this looks up the information corresponding to this specific pack of soup and loads it from from the blockchain, and it shows you a number of things so first it shows you the product as such, and then here you see the actual track and trace history of where the pork meat is coming from, on the one hand, on a map with dates and geolocation.

A

So here you see how the meat went from the farm to the processing facility to then the location where the product was scanned. A.

A

Little bit, oh, we just leave.

A

So this is the map based view and then there is also the actual steps in the supply chain. Again, you see it's in Vietnamese, so we don't actually see what it means, but this is the step from the farm. Second step is the processor third step is the production facility for step is, is the scan then.

A

Additionally, you see the details of the farm and the certifications that the farm has, so you can retrieve the actual certificate that shows that the farm has the required certifications for clean pork regarding the water that was used and the fact that no hormones were put into the pork and then, after that, this is like the provenance view that shows the history of the pork meat and after that, Unilever also uses.

A

This touch points that now they have with the consumer for all types of other things, which was also one of the aspects that then justified the business case for this project, for example, they use it as a touch point to sign up the consumer into their loyalty community, so they say after you have checked. This also join our loyalty program.

A

Third thing: is they use it to give them recommendations on recipes, how they can use the soups with other ingredients and last they also give them recommendations on other products that they can buy, which are somehow related to the soup that they have, and this is how they are really trying to use. This touch point that they now get with the consumer to the fullest to build up this two-way connection with their consumer.

A

So this was the the technical part like the software development part. But apart from that and I think this was the real challenge to crack in the project. There is the operations and and the launch of the solution, so we from the blockchain community always talk about software development, challenges and scalability and and so on. But when you do an actual supply chain project, there's one part which is also really important, which is how do you integrate a solution into an existing supply chain process?

A

What does it mean to print a unique QR code on each super-pac and how do you ensure that each super-pac has the correct information that it should belong to, and how do you create all these all these links and that's a whole world for for itself? So to give you an example here, traditionally Unilever prints their packaging with an offset machine. So that's an offset printer, it's a printer where you put in a template and then it just prints millions of times the exact same thing.

A

But now, all of a sudden, we have a unique QR code that needs to be embedded into the packaging, and this should not slow down the speed at which they can they can print. So how did we solve this for this project? Is they left the rest of the packaging as it was before, and they added an empty spot for for a unique QR code, so they had a hybrid print process.

A

First, they printed all the packaging, and then they installed a new digital printer line in the production line, which was able to print unique QR codes on on the product. So that was the printing challenge and the other thing is usually the way that production works is that labels for products are pre-printed, so they're pre-printed and then once the meat and the product arrive.

A

Only then is the packaging used and and processed. So this means there were millions of pre-printed Kannur super packs with QR codes already on them, and then somehow we had to create the link between the meat that is arriving and the QR code, where this meat is going to be put into, and this means that in the production line, each QR code had to be scanned and then associated with the meat batch that that it belongs to. And for this also, to give you an example, we had to inline camera.

A

And this was again a process that we learned a lot of testing, ensure that.

A

So this is the operations view and then the third thing that there is really important- and here we were a little bit less involved- is that a QR code based solution is only as effective as the amount of people that know about it and the amount of people that actually scan it and that's.

A

Why then Unilever with a little bit of support from our side, did like a whole launch campaign where they advertised that now they have a QR code on their product packaging, and this was a 360-degree launch which was done in the noir category and sometimes also combined with other categories like drinks that that Unilever has so that was a whole multi-channel effort that they did with TV ads in store ads and and so on.

B

A

So that's the TV ad again. So what did we get out of the project and what were like lessons learned so, first of all, I think it was a strong co-innovation push towards product transparency. It was quite a bold step that Unilever took in this case and quite a big investment that they they made to make. This solution happen not only into like a blockchain solution, but also and in adjusting their operations. So we had a few learnings with regards to how to make such a unique identifier.

A

Project happen, yeah and then learnings that we had is okay. This was now based on 30 million codes. How can we scale this solution globally? That's that's a challenge that we are still working on, because this was now implemented for one particular facility in Vietnam, but it doesn't mean that the processes in that facility work in the same way as they work in Germany or in the US or in Indonesia or or in Thailand. So you might have to repeat the same supply chain.

A

Mapping exercise for all the other places where this suppes is produced, and the last learning was that we we definitely need to improve the architecture and the of it. If you talk about the blockchain based solution, if we want to do a global rollout, so sawtooth was was definitely a good solution, but we also realized the limitations that that it has and the areas where it can still be improved, such as scalability and actual stability of of the network.

A

So if you have a solution that is in the market, you need to make sure that everything runs and and the blockchain network or other things just don't stop all of a sudden, and we had to do quite some work in ensuring that we have an enterprise ready infrastructure that also works in in such an environment.

A

So that's all I wanted to talk about I hope that gave you a good example of like an actual project and what it means to roll out a solution not only as a POC but there's an actual solution at the market and I think we have around 10 minutes left. So we can take some questions. Yeah.

A

That's a good question that it is an option, but one main benefit that Unilever wanted to get out of this project is consumer interactions and we asked aunt Russ. We have a system that allows you to track each individual QR code, so you can see how many times- and where is this individual QR code being scanned? And if you connect this with data from Google Analytics, you can add a sufficiently high number of scans. You can also be arrived like whether unique codes in Ho Chi Minh are scanned by males or females and so on.

A

That's one reason. So the consumer, demographic and insights part, if you have one code per batch, which may be one batch of meat, generates like a thousand packs you're not able to like you see that one batch code is scanned in thousands of locations, whereas if you have an individual code per pack, you can see that this individual pack was scanned five times. So you know, okay, consumers seem to scan packs more than once or they sin they scan it like at this time of the day and and so on.

A

That was one reason and the other reason is also just for security concerns. If you have a unique QR code on a super-pac, you can also detect, if maybe someone is far as like faking your products. If you have like one soup, that is scanned in a thousand different places, all of a sudden you can derive that. Probably someone is copying, your your soups, so that was the the other reason.

A

A

Yeah so you're talking about the SKU number in it, it's not really necessary to tie the QR code to the the SKU. We have this ability in our system, so a QR code can be tied to an infinite number of.

A

The barcode yeah yeah I mean the QR code is not really used for the point-of-sale, so it yes, so we could, in theory, try to do that. We can try to embed the barcode ID in into the QR code, but to my knowledge, I, don't think that the Vietnamese point-of-sale systems are able to scan QR codes, so there is still a barcode and the QR code. Our dream is, of course, that at some point we only have a QR code on a product and you can use those with point-of-sale systems as well.

A

A

As long as Unilever wants to support it, so from our side right now, like the whole solution is done over a period of 18 months, that's the first trial to see how many people scan it and and so on and after that, we need to reevaluate how yeah how to continue. Yes,.

A

Yes, so, regarding the first question, I mean part of that information is confidential because that's Unilever and for what I can say is they have experienced a sales increase in the last months, and now it's it's hard to judge if this was because of the blockchain solution or because of other factors, but they are experiencing sales increases and there is a sufficiently high number of people that are actually scanning the in the solution.

A

I can't tell you the exact percentage, Kuznets Unilever's info, but I can say that there's a quite a good scan rate of people that are actually scanning, and that's also one advantage that you have in Asia that people QR code adoption is is a lot higher than in Europe and in the u.s.. With regards to the second question, whether consumers care about the different steps of the supply chain, I think what you nearly were found in in terms of research before the project was that the like they they have this.

A

They call it Nielsen social listening indicator, so they they use this to extract themes from social media and from other topics, and they realized that food safety was a very dominant topic among consumers in Vietnam, because there have been scandals in the past and Vietnam is very close to China and there's a lot of food scandals in in that area. So they knew that this topic of food safety is actually is very important. Now, what we're also doing is they have in the next release?

A

Now we have embedded a small survey in in the landing page which pops up after the consumer scans it with a few questions that vary every every month and Unilever can just flexibly adjust and with that they're trying to get some more feedback on consumer preferences and and what they really care about.

A

A

Yeah so right now the yeah they're using CSV file, uploads too, and sometimes yeah integrations to the ERP system to upload information to the blockchain in the future. It should be I OT, based or something in that direction, but for the first base we worked with what we had, which was CSV and ERP integrations.

A

A

So, in the first phase of the project, now there were two farms involved, so it wasn't a whole lot like of farms and what we just did is we registered the farm in the system as like, like as a user or as a trusted farm from where meat is coming from.

A

Then, if you scale this up to thousands of farms and then you might think of implementing a farmer, management system or ID management system, or if you look at individual farmers, for example, you could use some kind of identity system to manage all these different identities. I.

A

Think we have two three more minutes. So if there's any yes.

A

That's a good question, so today it's a it's a certification body, I I, don't know the name to be honest of like the actual entity that issues the certificate. I can look that up. So it's still it's a third party that is certifying that and doing inspections.

A

What is interesting about this whole space of certification is I personally believe that in the future, as we have more and more IOT sensors in place, I think we can move into a real time certification environment, where you don't need someone to walk there and check, but with sensors. You can monitor the like what the parameters that are relevant is: has the the pork received clean water and has the pork being fed with any hormones?

A

And these are things that in theory, you can measure with biometric sensors and with water sensors, so maybe in the future, we're able to to automate that rather than relying on on certificates.

A

A

We've had numerous conversations with certifiers in the past and we are very interested in working together with them. They're, not the parties that are the fastest to move, unfortunately, but they are a very relevant party for for these projects.

A

Alright I think we're done. Thank you.