In a recent LinkedIn Live session, Dr. Darin Detwiler (Dr D) hosted an insightful conversation with Cronan McNamara, founder and CEO of Creme Global. Cronan shared his vision for a world where better decisions are made through real data and scientific models. With a background in physics, computing, and mathematics, Cronan has been at the forefront of data science, particularly in the realm of food safety.
Listen to the interview here via Spotify:
Cronan McNamara’s Journey:
Cronan McNamara’s journey into food safety and data science was discussed during the interview. It began over two decades ago at Trinity College, Dublin. His involvement in an EU risk project called Monte Carlo, which applied advanced mathematical methods to food safety, laid the foundation for what would become Creme Global. Founded in 2005, Creme Global emerged as a spin-out from Trinity College, offering data services and models for food safety, risk, and exposure to governments and industries worldwide.
The Role of Creme Global:
The conversation then moved on to the role of Creme Global. Cronan explained that Creme Global is a data science company with a strong emphasis on scientific rigor. They have developed numerous models deployed on their computing platform, focusing on exposure to chemicals in food, nutritional intake modeling, and predictive models around food safety. More recently, the company has been helping organizations gather and aggregate data through what they call a ‘Data Trust.’ This approach allows for structured data sharing among producers, exporters, importers, regulators, and consumers, facilitating better insights and transparency.
The Journey of Data in Food Safety:
Dr. Darin and Cronan discussed the parallels between food’s journey from farm to table and the journey of data. Cronan emphasized the importance of data sharing across the food safety ecosystem, where the data mirrors the food’s journey (a digital twin for the food product, as such). By digitizing and structuring such information, Creme Global makes food safety processes more transparent, safer, and robust. He highlighted that comprehensive data collection and sharing are essential for understanding risks and trends in these systems.
Predictive Analytics and the Value of Data:
One of the critical discussions was around the value of data at different points in the food safety journey. Cronan pointed out that traditional food safety testing methods, like testing at the port of entry, are insufficient. Instead, gathering data from the early stages of the production process can provide more valuable insights. Predictive analytics, akin to having a Google Maps-like system for food safety, can help foresee risks and mitigate them before they become issues.
Challenges and Solutions in Data Utilization:
A significant part of the discussion focused on the challenges of getting organizational buy-in for data analytics in food safety. Cronan outlined three levels of maturity in data utilization: retrospective data gathering and analysis, real-time data visualization, and predictive analytics. He emphasized that building trust in predictive models is crucial, and organizations must not blindly follow machine learning algorithms without factoring in their team’s critical thinking skills.
Leadership, Culture, and a Touch of Hollywood:
The conversation also took an amusing turn when Dr. Darin and Cronan discussed the movie Armageddon. Darin likened the need for a multidisciplinary team in food safety to Bruce Willis and his team of oil drillers working with NASA scientists to save the world. Just as the characters in the movie had to blend their expertise and common sense, Cronan stressed the importance of a culture that values both scientific rigor and practical experience. This analogy highlighted the need for collaboration between data scientists and industry experts to truly unlock the potential of data in food safety.
Conclusion:
Cronan McNamara’s insights underscored the evolving food safety landscape, driven by data science and predictive analytics. As technology becomes more accessible, the food industry must prioritize data-driven approaches to enhance food safety and quality. By fostering a culture that embraces data, organizations can better protect consumers and ensure a safer food supply chain.
For more information on Creme Global’s innovative solutions in food safety, visit our home page.
Full automated transcript:
Embracing the Future of Food Safety: Insights from Creme Global CEO Cronan McNamara
The elements of the data journey in food safety
[00:04:58] Dr. Darin: We talk about things in food safety, like food’s journey from farm to table, but we can even talk about the idea of food safety’s journey and like the last mile of food safety to some degree, that idea of.
Of food’s journey. There is also this idea of the data’s journey, right? We collect data, we store data, we secure data, we access data, we analyze data, and we communicate data. What am I missing here regarding the various elements of the data’s journey?
[00:05:34] Cronan: That’s a great point. When you’re thinking about it, you can think of the imports and exports of food and the data that needs to travel with the food.
Everybody collaborates in this global food system, which then becomes a global food safety system in which we export, import, produce, and consume.
We believe that data has to travel on that same journey, and sharing data amongst producers, exporters, importers, regulators, and even consumers is really important. Information should be shared and accompany the food traveling across that same journey.
So, everything we do as a company is to try to digitize that information, structure it, and make it useful to help streamline those processes and make them safer and more robust. As the food travels on that journey, the data can also. Does that make sense?
The Biggest Value of Data
[00:06:40] Dr. Darin: It does.
And I want to go back to something you said about the idea of the value of the data. Is the value of the data the same at all those different points, or are there some places or ways in which it could be more valuable at different points than at other points?
[00:06:56] Cronan: Excellent. And that’s a great question because I think the traditional method of food safety testing and monitoring often happens at the port of entry. And I think that’s like trying to find a needle in the haystack. That was spoken about at GFSI, which is this needle in the haystack of food safety testing. I believe more comprehensive data collection and sharing has to happen to understand the risks and trends in the production system right from the early stages in the grower or the producer, the processor, and the shippers.
We can’t just rely on a small number of inspection tests at the port to compensate for the lack of transparency or data sharing that’s not happening along the earlier part of the chain. That’s where the Western Growers Greenlink project is a real trailblazer, in which we capture the raw data and analyze it and the results anonymously and confidentially.
But it’s all there. And that’s going right back to the farm or the ranch. And I think that’s important. Then, you can look at what’s happening and relate that to the weather, different things that are going on in the environment, and even commercial commodity information, such as prices, etc. Are those affecting production, commerce, and trade? All of these things play a role. Gathering that data early is hugely valuable.
[00:08:16] Dr. Darin: Where is the biggest value in data and food safety?
Different categories, whether we’re talking about quality, safety, defense, or even authenticity. And I imagine that even data can be looked at that way, right? You have to look at the quality and authenticity of the data, and we have to secure that data. I don’t want to say that those are obstacles. I think those are just part of the job of collecting data. But I want to get to the big question here: where do you see some of the biggest value?
It could be from your personal experience or the companies and industry you work with. Where do you think the biggest value is seen regarding data and food safety?
[00:09:01] Cronan: I think what we aim to develop is having the wisdom of the crowd and sharing that wisdom amongst a big industry group.
So you mentioned authenticity in the fiin project in the UK. The project is headquartered in the UK but has members from other regions. They’re sharing insights on authenticity concerns they detect in their supply chain.
This is quite sensitive information that companies wouldn’t readily want to share, but fiin has come up with a mechanism using a law firm, legal privilege, and secure anonymization systems to protect the data. Now, they can confidently share this information amongst the whole group of around 80 companies and learn from all of that.
Therefore, they can target their resources into those areas where they’re seeing more authenticity risk in the supply chain. Now, that’s not necessarily food safety, but it can also be analogous to food safety. If there are certain risks in the supply chain regarding authenticity, there may also be risks regarding food safety.
On the food safety side, I believe we can predict better when risks are higher, and we can have a kind of Waze or Google Maps-like intelligent system where we can share data and get guidance to make better decisions depending on our exact circumstances or environment.
For example, growers could be informed if certain risks increase because of the weather or potential weather coming down the line. We advise you to put these mitigations in place, as in, in an analogous way, you wouldn’t use that road in Google Maps or Waze because we know there’s traffic on it.
By understanding the environment, the end users can adjust their strategy and mitigate risk. So, I believe there’s a lot of potential to combine all that data that we’re working on, whether with environmental information, what’s happening in the market, what’s happening in the growers regarding safety data, and all of that.
They can share that information and have a better predictive system to help them mitigate risk and even make recommendations about what to do in their environment.
Predicting Food Safety with Data
[00:11:03] Dr. Darin: I’m really glad you brought up predictability and predictive analytics. I believe when I meet with industry folks and food industry folks talking about data collection and digital solutions, we talk about three phases. They all hold value, but I think some hold more value than others.
- 1) You collect data but don’t look at it until after an event has occurred. There’s definitely some value in there. What took place? After the incident debriefing, and looking at root cause analysis, things of that nature.
- 2) Live in real-time, like, where are we right now? I had a conversation yesterday with some folks about knowing where you are between the guardrails. Are you too high? Are you too low? Or whatever you’re measuring. And again, there’s value there, maybe more so than the after-the-fact (1 above), because you have the ability to insert some element of control before it’s too late.
- 3) Predictive analytics. It’s the idea of being able to predict, and you can’t predict everything, but you can’t predict anything if you’re not in that mode if you’re not working that way. And the idea of mitigating—to me, I see it as where the greatest value can be found because, ideally, you’re preventing problems before they happen.
“you’re preventing problems before they happen“
How do you get organization buy-in?
Dr. Darin: You have no liabilities if you’re doing it right? Yeah. You have no losses of product. You have no consumers being harmed, you have no recall, and there you have financial and reputation loss, even in some cases, so where would you say if someone says I only see value in one of those phases?
What would you say if you wanted to convince someone that no, there’s value in all phases, it’s just that they’re different in terms of how you do it.
[00:13:14] Cronan: That’s a good question. And I think these are three levels of maturity, Number one, you talked about they have some data collection capability, but tend to only review it in hindsight when there has been an issue.
There is value in that. As you mentioned, we are trying to move many of those projects mentioned before into more real-time data collection. Even when you visualize that real-time data, it provides great insight.
Food industry practitioners are already very knowledgeable about their own industry and what they can do, so once they unlock some of that data that they can aggregate and share at an industry level, they can add great additional insights in real-time.
Okay, so let’s move on to predictive. This is the Holy Grail, the dream. We want to have that predictive Google Maps kind of app helping us. That is the icing on the cake that we’re trying to get to, and we’ll probably talk about this a little bit later regarding the poll you shared around leadership buy-in. Are they skeptical that you can predict more than they already know about their industry, being experts in their field already?
Some people may be skeptical about predictive analytics, which can certainly be seen as a black box. It takes time for them to trust the black box, that model, or the predictive outputs of a machine.
But the second phase—bringing their data to life in real-time, visualizing it, starting to work on those predictive elements, and showing that building trust in a system’s predictive capabilities is where you need to get to. We need to back up our predictions with visualizations to show that mitigating or eliminating certain risks that would or could have happened in the future is a valuable capability.
So, those are certainly three interesting levels of data maturity in food safety. And I’d say a lot of organizations are still at the first level. They have the data, but they only use it in a kind of retrospective analysis, as you say. But we see that organisations are trying to get to real-time data with plans to move forward to the predictive maturity level.
[00:15:33] Dr. Darin: Correct me if I’m wrong, but in a way, it builds upon itself, right? I mentioned the idea of looking at the data after an incident. It’s after the fact, but doesn’t the act of doing that after an event help in terms of when you’re trying to be predictive and look at the idea of, wow, is there a pattern, or is there a trend that we can, we now see?
And we can say that if we see this pattern again, here’s what history has told us.
[00:16:03] Cronan: That’s machine learning in a nutshell! So, a machine learning system constantly reviews what’s happening, different input factors, and the actual outputs. That’s what we need to remember about a machine learning system – it’s constantly learning from the data. So it’s constantly looking back and checking its predictions against what actually happens every single time you iterate through your your data. That is exactly why machine learning models can learn well – if you have enough data.
Now, there’s still value in humans reviewing the data (or graphs). I often say that you can’t blindly follow these machine learning algorithms. You have to use your intuition, look at the data yourself, and double-check that you’re not going to just blindly follow a machine’s predictions, right?
That’s why this second step of visualizing our data in real-time can really help us build trust between our understanding of our systems and the recommendations the models make for us.
Practical ways machine learning will actually help.
How can you trust the data?
[00:17:17] Dr. Darin: I am so glad you brought up machine learning because I was recently at Future Food Tech in San Francisco and someone made a very, I think it was an incredibly powerful point that some leaders don’t understand the role of machine learning, and even if they have a rudimentary understanding, they’ll think of it as machine learning will do the job for you.
But in reality, it won’t do the job for you. You need to treat machine learning like a lab assistant or an assistant to you who has some very good skills and can accelerate the speed at which you’re doing your work. But it’s still only an assistant to the people who are actually humans making decisions and making analyses.
How do you relate to that comment?
[00:18:06] Cronan: I totally agree. It’s a tool in your arsenal of technologies that you want to bring to bear. And I see a few interesting comments here from Drew. It’s saying, assuming we’re collecting the right data and that it’s authentic data.
If you can trust the data you’re gathering, it’s the right data in a timely fashion and in a structured way. There are not too many errors in the data. My view on what to collect is that you should try to be as broad as possible with data collection. So you’re capturing as much of the inputs that you think are, or could be, important regarding what will affect food safety.
Anything you can digitize and put as an input into your system that you think is reasonably important can be useful. Then, you can train a model, and if it converges, you can have confidence in that model. The models can report back their level of confidence in their predictions. Then you can review those with a critical scientific head on your shoulders. And never forget that it’s your job at the end of the day to review and critically analyze those models and results and make decisions based on those.
[00:19:12] Dr. Darin: Yeah. In the chat, Jennifer is reemphasizing that getting people on your team who can understand what the data is saying and how important that is. It’s only as strong as how you’re able to interpret it.
When working with data, there is a need for knowledge of the food sector and science.
And I think it’s funny, too, because sometimes I’ll hear people say, oh, we’ll get our team to do this, and it’s no. No, I think you want to have people beyond your IT team, people who actually understand the product, they understand the process, they understand the manufacturing floor or the, the production line, they understand this.
You can’t do this from a remote office where you’re just crunching numbers. You have to understand the real world. Three dimensional, real time, what is going on in the plant in order to be able to truly interpret that data.
[00:20:06] Cronan: And I think that’s one of the important things that Creme Global brings to these projects in that we have almost 20 years experience in the food sector and we have strong scientists, data scientists and IT people/engineers working together.
So we have a multidisciplinary team that can understand and solve problems. If we don’t know your manufacturing environment at the start, we learn as quickly as we can, and we talk the experts on the ground who know what’s important in the manufacturing environment or production environment, and we try to get their input on what’s important.
Then we look at what data can you capture to represent that factor? How do we interpret that data? If it was a sample in the field, how many grams was it? And how many acres it was? All of those important concepts that we then can build into the model, rather than making a naive mistake on leaving something important out of the data.
So we try to truly understand the data from the end user and from a scientific point of view, so that the models are not misrepresenting it or, making a mistake in how they’re using the data. So I think that’s a key point.
Yeah, I’m bringing in that last point on the poll. The point about having qualified staff came up on the poll—but it’s not number one. We should come back to that.
The Armageddon factor! The Importance of a Multidisciplinary Team.
[00:21:25] Dr. Darin: Yeah, by the way, we’re going to get to the poll here and really quick. You were talking about multidisciplinary team. Call me crazy, but I had this flashback to that movie Armageddon, where Bruce Willis and his team of jack-hammering rock-exploder-people had to work with NASA scientists, to blow up the asteroid. And there’s always that question: should they have taught the astronauts how to blow up the rocks, or did they teach the rock experts how to become astronauts? And obviously as the movie played out, it took both. Yeah, for them to be able to pull this off. So do you teach the, do you teach the data folks about the product or do you teach the product experts about the data or is it a little bit of both?
[00:22:09] Cronan: It’s a bit of both. What we try to do is hire really good scientists who have a good grounding in science and also good intuition for data. They could be in the physical sciences or chemistry, where you do a lot of data modeling as part of your curriculum. Still, you also need an understanding of science and natural or physical processes.
You can build from that foundation into understanding more about the customer’s challenge, what they’re doing in the ranch or the production environment. They have the core concepts pretty well understood and they’re curious to learn more about that. I believe that if don’t start from a good scientific base, it’s hard to learn data science. Actually, it’s hard to be a good data science without a good scientific background.
The other thing, the funny thing about that movie, I think was that it turns out that guys who like to explode rocks had a lot more common sense than the NASA scientists, right? So you need a good amount of common sense as well as a good scientific, theoretical, and conceptual understanding of what might work, so a strong dose of common sense thrown in is very good.
[00:23:17] Dr. Darin: I had no idea going into this conversation we were going to talk about a Bruce Willis movie, but it makes sense. Since you have to have the multidisciplinary team, you have to have the cross training, you have to have the strong science background, but it also takes the common sense and the courage to really be able to pull this all off.
[00:23:37] Cronan: Yeah, that’s where the fun is. And Darren, this is an interesting and challenging field with so many moving parts. That’s why you need to have good people who think critically, think scientifically, but also have the data skills then to plug it together and come up with something either visual analytics or even even further into machine learning modeling.
[00:23:59] Dr. Darin: Yeah. Someone just posted here, Jennifer: so many strengths, strategic, analytical and learning.
You want someone who can think through the information and can see the trends. What I also think is. And she brings us some very good points about seeing the trends. It’s also got to be an environment where someone is confident and has the skills and the freedom to communicate the findings and the trends.
The importance of leadership and communication
But communication is only as strong as the idea that there are people who are receptive to listening to this. I remember, oh man, it was actually in Dubai. One of the guys who was speaking said that his company was hired to analyze his data for a company, and they gave him the answer that they were actually looking for to solve their problem.
But the company was like, yeah, we don’t want to do that. And then why did you ask us to crunch all this data? And it just blows me up. It’s mind-blowing to think that there are people who can truly analyze and show you this is the issue. This is the thing you need to do.
And yet there are still other people, leaders who will say, yeah, but I’m not going to do that right now. Cause it’s not my priority, or it goes against what I hold as a reasonable response.
[00:25:17] Cronan: Yes, it comes down to that leadership piece in your poll leadership buy-in, what they want to achieve. Maybe they have different priorities. It’s might be cost saving rather than making things safer, or it could be some other priority that they have.
What do you believe is the biggest obstacle in unlocking the value of data in food safety?
[00:25:40] Dr. Darin: It’s sometimes long-term versus short-term goals, right? But let’s do this. We’ve had we’re up over 108 votes now. So again, the question is. What do you believe is the biggest obstacle in unlocking the value of data in food safety?
We had four options at number four. We’re at 11 percent not having qualified staff. We’ve heard that from our participants already. Are you surprised by that?
[00:26:25] Cronan: I’m not too surprised it came in last place (or number four) to some extent. There is good talent out there and people coming out of university these days off, with a scientific degree or PhD, they’ll often have good data skills. I can see why it’s not a huge barrier. It’s not the top barrier, in the poll.
[00:26:47] Dr. Darin: All right. Number three, there is no funding for tools or training. And that is at 18%.
[00:26:56] Cronan: Yeah. And I think the next three we’ll probably talk about together. They are linked in a lot of ways. In that I mean the top one is leadership buy-in that we’ll come to in a minute, but without that, the organisation will not be willing to make that investment into the right tools or the right team.
So the bottom two – team and tools – could be grouped together as well. You don’t need to spend a fortune on tools these days with the open source technologies that are out there and Amazon’s cheap cloud computing capabilities, but that does take a bit of expertise to put those together.
It is cheaper to invest now. The cost of doing nothing.
[00:27:34] Dr. Darin: Is it safe to say that if you’re not investing into it now, you may be paying more down the road?
[00:27:41] Cronan: Yes, I would say so. I think it’s so important to prioritize. Your data strategy now to help you production, your food safety, your costs, …. It can do a lot of things and ultimately protect your reputation, right?
If something does go wrong, if you’ve got the data to show that you’ve done everything right, it could mitigate a lot of the downside. If the regulators come knocking, for example, after an incident and you can show them all the work you’ve been doing with data to back it up, it can really help you protect your reputation and help you get back to business a lot cleaner and a lot smoother than if you have nothing. So I would say it’s vitally important.
[00:28:21] Dr. Darin: I recently, within the last couple of months, wrote an article. It was in food safety news about the cost of doing nothing is going up.
The idea of it’s not just like the cost of a recall. It’s also, how much is it going to take in terms of restoring your reputation. If you look at one of the most recent federal fines, it was a 40 million to one company because of their failure to act quickly and appropriately and of course, there’s also the cost in terms of impact on consumers.
Health hospital, that kind of thing. So it’s important to get a sense of the idea of, okay, it may cost this much to invest in doing it right, but if we don’t do it, it can really cost this much.
3. Data Analytics not being a priority
So, let’s get to number three – at 26 percent of our participants voting that data analytics is not a priority.
So let’s get to number three at 26 percent of our participants voting about data analytics not being a priority. And you’ve talked about this a little bit already.
[00:29:20] Cronan: Yes, and again this will be related to the top point, and it’s got, a quarter of the people have voted for this one, 26%. People are busy and may be fighting on all fronts in their day job to keep the production running.
It just sinks down the list of priorities. Maybe it seems like a daunting challenge to get started and you’re not sure about how to do it and not sure about what value you’re going to get from it. So perhaps it just drops down the priority list of these companies. But as we discussed they might pay a price for that if they don’t get started.
4. Lack of leadership buy-in.
[00:29:57] Dr. Darin: Very good point. And at number one, 45 percent of 108 votes. Lack of leadership buy in.
[00:30:08] Cronan: Almost half the votes, eh? And that one probably would solve a lot of the issues below – if you could solve that first one, right? Because, if leadership really buys into it, obviously it becomes more of a priority, the funding will become available, and they can then hire the right staff.
Now why would it not be a priority? What would you think, Darren, or what are your thoughts on that? I have a few thoughts of my own.
[00:30:33] Dr. Darin: I think that there are two ways of answering it. One is the idea that we’re doing things the way we’ve always done. So why do something else? something different? Data collection is not the priority.
Data collection is a priority, but not for food safety. Again, like that idea of short-term versus long-term, data collection is about consumer purchase trends or how we can maximize profits, and we’ve seen that some of us have seen this sometimes where, oh, we’re going to use artificial intelligence to improve food safety and customer experience.
What do they use artificial intelligence and machine learning for an AI? How do we maximize profits? How do we minimize workers’ hours, and things like that? The thing is that they didn’t truly prioritize food safety in terms of their use of data collection. So it’s not like binary in it. It’s either you prioritize it all or nothing. It could be that you’re prioritizing it, but not for food safety purposes.
[00:31:42] Cronan: That’s a good point. And maybe there’s not an awful lot of knowledge about what they need to do to prioritize predictive analytics and data for food safety. And it is an emerging field from the newer ideas like simply tracking your inspections, your testing regime around water, around product. You have to do all of that anyway, so why not organise the data?
That is, start gathering that data in a more digital and structured format. Then there are novel things that are coming out now about metagenomics (and swapping is getting cheaper) so sending those off and getting data and learning about the microbiome of your plant is now possible. Perhaps these things are still beyond the understanding of what benefit they can bring, and how you can get started to use that data for food safety? But those tools are maturing, which is the good news, and are becoming adopted by companies and government agencies.
These are coming down the track, and I think companies need to start, planning for that and getting on board with it.
Will data stop issues from occurring?
[00:32:52] Dr. Darin: In the chat, Drew asks a question here I want to respond to. What if a firm is collecting all of their required compliance data and it shows that everything is hunky dory but still has an issue?
Before I ask you your thoughts on that, I want to add that A, no guarantee collecting all the data, having data analytics, and, leveraging the value of food safety data is going to prevent things from happening. Things can still happen. It could be literally, a tree fell in a tornado through the roof.
It could be a catastrophic failure of a piece of equipment that you had no idea had some structural issue, or it could be, things that are out of your control. I think that the whole idea of doing nothing is when leadership had all the information and could have taken action but decided not to the recent one I mentioned about the 40 million fine that was with a company that had report after report of a rodent infection infestation and still decided to not do anything about it until the federal government came in, validated the rodent infestation and said, you knew about this for a long time.
You did not take action. That is the doing nothing that I’m talking about. It’s not so much the company that is taking, the opportunity, they really collecting all this data, and things just still somehow happen. And I think this goes back to the idea of leadership of what Jennifer mentioned earlier the idea of management being open to actually acting on the information you give them. And here she’s back again: “having data and not using it is like buying a smart TV for your break room and never plugging it in”. Why do you have it if you aren’t using it? seeing what, it’s saying there’s a lot of really good points here. What are your thoughts on that?
[00:34:53] Cronan: Yes. Back to Drew’s question, if they’re genuinely gathering the data and storing it and analyzing it. And everything looks good. And they’re, maybe going a little bit deeper than just the basic compliance data and testing and investigating. And then something goes wrong, I think that’s unfortunate, but at least they can say to whoever comes to talk to them about it, be it a customer or a regulator. Listen, we’ve done everything right, look at our data, it goes back five years and every time we see something, we, have a strategy and we mitigate the risk. So I think that’s just unfortunate. And you just have to try to, you have that data there as your backup to say we’re doing as much as we can, if something goes wrong, try and learn from it as well. Like what did we not have in our data that could have predicted this?
Okay. Thanks. Perhaps it’s not predictable like you’re saying; it’s just a freak event, but definitely, I would be of the view that we should go a little bit deeper and try to figure out if there was a potential way of predicting that if we had done something, can we enhance our data strategy to try and figure that one out next time?
Typical responses from companies for food safety issues.
[00:36:02] Dr. Darin: I want to thank everyone for taking part in the poll here. What I find interesting is that. In over three decades of my observing companies, when there are issues and they communicate with the media, they typically say things that kind of reflect what our poll is indicating here the quote-unquote I call it the playbook responses are, we really do value food safety in response to this issue that we could have prevented, but we didn’t.
We’re going to put someone in charge of this, make sure that this is our priority, invest in it, and retrain our people. These are the same bullet points you hear every single time. And yet, when you look at the responses here, they hint at the idea that prioritization and leadership are key issues.
How do we encourage more folks? To look at this deeper, Drew and Timothy also mentioned that data is being used to investigate food safety failures, diagnose them, and use them for predictive purposes. This all comes together with management commitment and a food safety culture.
We, I know we were jokingly talking about the movie Armageddon, but the idea is that it’s not just the tools. It’s not just one set of stakeholders. It is multidisciplinary, and it requires common sense and courage. I want to turn it over to you. Based on the conversation and the poll, what do you recommend to leadership?
[00:37:36] Cronan: I think it’s about impressing upon them that this is a very valuable resource as Jennifer mentioned, that it’s like buying a smart TV, but never plugging it in. This data is at your fingertips. Other companies are using it. The government’s going to be using it. You need to be able to converse in a way with, in your industry on a data-based approach.
If you’re not, you’ll look a bit like you’re lacking in your food safety endeavors, and if and when an incident does occur, you’ll not have anything to back it up. I think it’s vital that they start on this journey as soon as possible. That’s how I would try to explain it to company leadership.
Government and consumer pressure for better food safety. Having better data then everyone else.
[00:38:26] Dr. Darin: You mentioned that there’s a government pressure, right? Here. We look at the idea of FSMA 204 transparency traceability. Do you think that consumers are demanding or expecting that this is going on behind the scenes?
[00:38:47] Cronan: I think consumers are expecting that their food is safe in the first instance. It’s taken for granted almost at the consumer level, but they assume that the industry is doing everything to protect them and that the government is doing everything it can to protect them. And all of those initiatives you just mentioned are certainly going to help. Consumers are data-savvy and they can learn, they can research, they can share information on the web, they can discuss products, they can discuss incidents, so I think there is pressure coming from that as well as from NGOs who are a bit more organized and putting pressure as well and gathering their own data.
So, sometimes the best defense against those those reports is having better data yourself, a company or as an industry. So if something comes up, you could say we’ve got like 200, 000 data points that we’ve been gathering for 10 years, and this is what it shows. Your little study is not as powerful as the data we have, which we’ve collected here in our industry group. So it’s always great to have that confidence and that backup of having better data than anybody else.
Creating a culture that understands and works together
[00:39:58] Dr. Darin: I remember talking with a consumer audience recently, and someone was like, look, I have real-time data.
When I order something online, I have real-time data of, where it is, if it’s been packaged, if it’s been, left the warehouse, whatever. And I should be able to expect, when someone’s cooking my food that they use more than just their eyes. There’s some element of thermal intelligence being used for either real-time monitoring or more precision type monitoring of whatever parameter it is in terms of food safety, especially with things that are being mass produced and manufactured confused consumer packaged goods, ready-to-eat foods, things of that nature.
And I think it’s a good point. We shouldn’t expect technology from 50 years ago when it comes to how our food is produced, when, in other aspects of our lives, we have much more cutting-edge technology.
I think that’s where some of the pressures are, but I do also think that we have to remember that, the people that work the entire workforce and the leadership teams for companies. They’re also consumers. And we need to make sure that there is this culture. I can’t remember who brought it.
I think it was Jennifer who brought it up. Someone brought it up. The idea that the culture Oh Timothy brought up the culture. The culture has to accept. This partnership between the workforce and machine learning and the artificial intelligence and the data analytics and the predictive analytics has to be a a culture that is understood and works together.
[00:41:40] Cronan: Yes, and the manufacturing environment, as you’re mentioning, you can still see very traditional approaches. But the good news is that sensors and monitoring technologies are getting cheaper and more available so we know companies can use better sensors that are there in real time monitoring temperature, humidity, pH, all of these things. So they can get food safety down to a ‘t’ in these manufacturing environments. That’s what consumers expect.
A further benefit is that they can use that data to save cost or time on production without compromising quality, taste or food safety. These are all multifaceted challenges that food companies are grappling with and they’re doing it by trial and error or experimentation, on the production environment, but data can help inform those from the recipe right through to sensory and food safety results. So that’s a really interesting journey that, that we’ve been working on with a few companies to gather data along that process and try to see if we can predict sensory, safety, physical and chemical parameters of the finished product right from the start and then tweak a few knobs in the production process to save time, save money without compromising on any of that. These technologies are developing and becoming more mature in the manufacturing environment, which is good to see.
Wrap up
[00:43:08] Dr. Darin: Great. Cronan McNamara, Creme Global, thank you very much. I want to thank you for not only helping us prepare this poll that we had today but also helping us break down the poll results.
Cronan, it’s always a pleasure to meet you. And it’s funny because we always meet in different parts of the world, and next month, you’re going to be at IAFP, which is in my backyard. Excellent! And this time, I don’t have to travel.
Thank you for all that you do and for taking the time with us today to discuss this topic and feature these items.
About Cronan
Our guest is Crona McNamara, founder and CEO of Crame Global in Dublin, Ireland. Cronin believes in a world where everyone makes better decisions based on real data and expert models.
Creme Global’s role is to enable organizations to make this a reality through its experience, data science platform, and global network. I’ve had the wonderful opportunity to meet Cronan several times in Ireland and Dubai where we have worked together and talked about food safety data.
While our participants are getting situated Cronan, many of the people watching this might not really have any idea who you are. Tell us a little bit about yourself and also a little bit more about Creme Global. Tell us a little bit more about what you do at Creme Global.
[00:02:20] Cronan: Sure. So my background is Physics, computing and maths, and that’s what I was studying in college, and I really enjoyed the simulation and data science aspect of it. Now, this is more than 20 years ago before it was probably called ‘data science’. I did a master’s at Trinity College, and there I met a professor who was working on an EU food risk project called ‘Monte Carlo’, where we applied advanced mathematical methods to food safety. So I thought that it would be an interesting thing to work on for a few years. So I got involved in food safety just over 20 years ago at Trinity College applying these mathematical methods and developing risk software.
In 2005, just under 20 years ago, I founded Creme Global as a spin-out from Trinity College and we started providing data services and models in the area of food safety, risk, and exposure to governments and industry around the world. So that’s how I got into this space and I’ve been excited to work in it over the years because it’s an ever-changing, challenging environment with lots of interesting data, if you can gather it.
And that’s what we’re going to talk about today.
About Creme Global
Thank you for that. Creme Global is a data science company with a strong emphasis on science. We take a scientific approach to data and analytics and have developed several in-house models that we’ve deployed on our own computing platform on Amazon.
So, we spend a lot of time developing our platform and models. Those models are doing lots of things, from exposure to chemicals in the food supply to nutritional intake modeling and predictive modeling for food safety. Recently, we’ve been working a lot on helping organizations gather. Aggregate data using what we call a ‘Data Trust’ (or a data portal), allowing organizations and members of a broader group to share data in a semi-structured or structured way. Then, we aggregate all that data so the group can gain insights. This is done through visual analytics or more advanced methods like machine learning. We’re working on these kinds of projects with the US FDA, Western Growers in California, and a group in the UK, a retail group called fiin.
So, we have several of these Data Trusts up and running. It is fascinating to see the insights you can gather from the data we’re aggregating across many organizations.
