[00:00:03] Good data science team that is well versed in prescriptive analytics can do things like provide us a resiliency test to your supply chain to say where are the weak points? Is it diverse enough that if something were to happen in any region of the world, can your system recover or manage that? Or how quickly can you do that? [00:00:26][23.0]

[00:00:28] Good morning, good afternoon or good evening wherever you are in the world. This is the Digital Supply-Chain podcast. The number one podcast focussing on the digitisation of supply chain. And I'm your host. Global vice president of SAP, Tom Raftery. Hey, everyone, welcome to the Digital Supply Chain podcast. My name is Tom Raftery with SAP and with me on the show today, I have Heather. Heather, would you like to introduce yourself? [00:00:54][25.8]

[00:00:55] Yeah, hi. Thank you for having me. I am Dr. Heather Krieger. I am a principal data scientist and the lead of the data science team at Savvis. Technology Savvy is a supply chain technology company. We provide software and hardware solutions to our customers, helping them solve in-transit visibility and asset tracking. My primary role there is designing and implementing machine learning algorithms to estimate the times of arrival and things like that, as well as design and run and report on all of our prescriptive analytics. [00:01:34][38.7]

[00:01:35] Okay, Espers. Superb, superb. And this is very timely because, Heather, you're unaware of this. [00:01:42][7.1]

[00:01:42] We're recording this on Thursday. But tomorrow, Friday, I will. And we'll be publishing this next Monday. But tomorrow, Friday the 17th, I will be publishing a podcast on analytics abroad, one on the topic of analytics. So having you come on to talk about prescriptive analytics. The following podcast episode on next Monday is really handy. So thank you for that. But what is what is prescriptive analytics? [00:02:07][24.8]

[00:02:10] Yeah, prescriptive analytics, from my perspective, are any insights that can solve or prevent problems. Most importantly, they allow you to make actionable changes in your supply chain system. So we tend to call them decisions that are backed by data. [00:02:25][15.8]

[00:02:27] They allow you to work with your system and within your system to improve its overall quality or to prevent loss or do things like that. [00:02:36][9.1]

[00:02:37] OK. Can you give me some kind of examples of where companies have used prescriptive analytics and to what end? [00:02:45][7.9]

[00:02:46] Yeah. So I think there are kind of two main ways that prescriptive analytics are often used. One is to solve current problems that are being experienced. For example, one of our customers is a glass manufacturer and they were experiencing breakage during transport. They were shipping predominantly on truck and rail and product was arriving broken. And that's obviously the least ideal situation. And so they started shipping by air, figuring that air travel had less kind of bumps and jar's in it. But using environmental sensors and our analytics team determined that actually the glass breakage occurred on loading and unloading, it did not matter which mode of transportation they used to actually transport the product. So paying extra to fly it on air was not actually getting them less broken product. And so that allowed them to then change the way they load and unload, allowing them to save money on lost products. Another example, solving current problems. A lot of in-transit visibility uses. EDTA is in order to determine when things will arrive. Well, we had a chemical manufacturer, a company that uses manufacturer chemicals that have to remain in a certain environmental kind of acceptable range for the chemicals to be viable. One of their trucks broke down and they they obviously were trying to fix that as rapidly as they could. But when it was determined that they could not fix the track, our predicted estimated time of arrival allowed them to make the decision to take that product back to its origin and put it on a different track rather than try to haul that truck to its destination because it would not have arrived and allowed for the environmental cold chain system to have kept it where it needed to be, they would have lost the entire million dollar worth of chemicals. So, yeah. So it can obviously solve kind of current, really specific case examples. We've also done some work with carrier compliance. A lot of companies subcontract out to carriers to deliver goods, and so they have certain rules about when those carriers should deliver the product by or whether they should drive on the weekends or participate in certain behaviours. And so carriers often report when they are non-compliant. But it's always good to have maybe a third party look over that and say, are they really doing what they say that they're doing? So those are all things that kind of prescriptive analysts can solve current supply chain issues. I think you're prescriptive analytics are most cost effective, as in preventing problems from occurring. And that's a broader look at the supply chain system for inefficiencies. What are best practises, those kind of things? [00:05:43][177.0]

[00:05:44] OK, superb. Are there any kind of particular industries or any parts of supply chain in particular that benefit from prescriptive analytics? [00:05:55][11.2]

[00:05:56] I think that almost any is any type of company or vertical, as marketing people often call them, could use could use prescriptive analytics. We often work with operations and logistics teams within companies, and that can be within an electrics, an electronics manufacturer, or it could be within. We work with a pharmaceutical company delivering products. We've worked with many different types of companies. We've also worked with security teams looking at specifically risk areas where things being stolen. What are stops? Are the stops that are being made in transit secure or safe stops or not? And then we often work with planners, people who are responsible for when things get to their final destinations. [00:06:49][52.3]

[00:06:50] So we work with a really broad range of of. Departments and companies. [00:06:57][6.7]

[00:06:58] OK, and just how widely used are prescriptive analytics? You know, would it be one percent of supply chain organisations? Ten. Fifty. You know, where are we roughly in in deployment of these kind of solutions? [00:07:14][15.2]

[00:07:15] I think as a data scientist, I would say not enough. But I it's it's definitely picking up steam. I think the current pandemic situation and the disruptions in supply chain that everyone has been experiencing as a result of that has really highlighted in-transit visibility and the value of that, as well as taking that a step further and saying, I mean, a good data science team that is well versed in prescriptive analytics can do things like provide us a resiliency test to your supply chain to say where are the weak points? Is it diverse enough that if something were to happen in any region of the world, can your system recover or manage that or how quickly can you do that? So I would say that not enough companies are using it, but it is slowly being adopted. And I think I think it will continue just as in-transit visibility has continued to grow. [00:08:16][60.7]

[00:08:16] So a prescriptive analytics. [00:08:17][0.8]

[00:08:18] OK. And hey, are there any regions that are further along the the the deployment route than others? Or is it is it pretty standard across the board? [00:08:27][8.2]

[00:08:28] I would definitely say customers who are transporting really high value goods have definitely leaned in more. But the cost of the loss of any one track or any one container for them can be multi billions of dollars. And so they have a a a thirty or a hundred and twenty dollars sensor to stick on. That container is a very small price to pay floor for that. So I definitely think that there is a higher level of adoption in some of those higher cost good transports. But I do see it. It's slowly trickling down. And I think prescriptive analytics can also play a bigger role in some of those smaller maybe like consumer goods where the value of any one container is not very high. But when you have thousands of containers that are being transported every day, knowing whether there's an inefficiency in that system can actually cost you days or weeks. And that becomes a significant amount of money. [00:09:32][63.8]

[00:09:33] Okay. Interesting. [00:09:35][1.0]

[00:09:36] You mentioned sensors that those are obviously important for the deployment of a prescriptive analytics solution. What else is needed? So if if I am a supply chain organisation somewhere listening to this and I want to deploy a prescriptive analytics. So I think this sounds awesome. I should have one of these. Where do I go to get one or what do I need to do? What's the pathway to ruling out prescriptive analytics and my supply chain? [00:10:03][27.0]

[00:10:03] That's a great question. I think the key component that is a data science team or access to a data science team. Ideally, one that has supply chain experience. There are a lot of providers of in-transit visibility out there, and they all offer some form of prescriptive analytics to varying degrees. And some of those prescriptive analytics are kind of basic insights or charts that tell you how often things are on time, what specific origin and destinations are taking longer than normal. So you can kind of get it starts to see trends using those kind of graphs in your supply chain system. [00:10:44][40.6]

[00:10:45] Obviously, there's the opposite end of the spectrum, like Sauvie, where I work, where we we work very closely with customers to understand their needs and understand what they what their concerns are so that we can specifically address those. [00:11:05][19.9]

[00:11:06] Okay. Okay. And what what kinds of challenges face organisations who want to deploy these kind of things? [00:11:14][8.2]

[00:11:15] Because, you know, I don't imagine there, for example, that many data scientists working on flavour yet challenges are are often related to buy in because in order to to get the analytics and you need a lot of data to do that. And so you have to have kind of the buy in to start collecting the data to have that filter through a system where it's cleaned and it's processed and it's looked at by by a team of data scientists. [00:11:48][33.1]

[00:11:49] But then to really use prescriptive analytics, you have to have the full kind of. He should say a full range of buying. But you have to have people at all multiple levels of the company. Be like we're going to use this data to make decisions and start changing the process. That usually involves someone in upper management. Someone in middle management. And then also, obviously, the people who are actually on the ground, who are who are everyday in and out working with these shipments in order to say, OK, we should no longer use this carrier because they're less efficient. Right. That requires sales and and logistics and everybody kind of no longer contracting that carrier, for example. So it requires. [00:12:33][43.4]

[00:12:34] To apply prescriptive analytics requires a lot of buy in. [00:12:40][5.4]

[00:12:41] OK. And I guess, I mean, feel free to correct me on this, but it did sounds like this is something that would be used by either larger carriers or ones. As to your point earlier, ones who are transporting high value goods. So, you know, maybe smaller or midsize logistics operations or supply chain organisations wouldn't have access to this or wouldn't be able to afford it or dot, dot, dot. [00:13:13][32.0]

[00:13:14] I definitely think there's more incentive for for larger companies or companies of who who transport high value goods. But I think with the Internet of Things and sensors becoming much less expensive as as their adoption and availability continues to increase that contracting. A company like Sabby who comes with the data science team and the prescriptive analytics is not it's not out of reach for many companies. [00:13:43][28.9]

[00:13:48] And I mean, you've mentioned a couple of times now that the sensors that are needed to collect the data, what's typically what kind of sensors are required to get the data to have something like this? [00:14:01][12.7]

[00:14:01] Sensors are definitely really important from a data perspective. There are high fidelity data source. I can trust what comes out of a sensor versus waiting for a carrier, milestones to be sent that says, oh, yes, we did pick it up. I can then tell you if it was or was not, because I know exactly where it is. So they are important, but the the use of sensors ranges based on what cut what customers are or what's actually being transported. So the chemical manufacturer, for example, they need cold chain sensors, they need environmental as they need to know the temperature and humidity and the light situation is, whereas someone who is transporting other goods really just needs to know where it is. And so then that's just a G.P.S. cellular sensor that's pinging its location as often as we want. So it really is based on the needs. And and sensors now range from small disposable sensors to the kinds of sensors that will give you a G.P.S. fix anywhere in the world, regardless of kind of what's happening. [00:15:12][71.0]

[00:15:13] Sure, sure. Sure, sure, sure. And I mean, where where is this all going? What's the kind of where we're at early stages of, you know, prescriptive analytics now, as you were saying. But where are we going ultimately where prescriptive. Take us to in five, 10 years? [00:15:31][17.0]

[00:15:32] I think that they will help us get to a place where the entire supply chain system is more efficient. I can see a time where we are routeing goods to a factory in real time based on or to a plant based on on needs, so that there isn't necessarily a person who's saying, oh, we need to deliver us to this location, that there is an entire prescriptive analytics system that that's reading the current inventory and all of the warehouses is saying here's the predicted amount of volume that they're going to need in the next three weeks. And so this truck that just got there just picks up cargo needs to go to plant a circuit. It can kind of remove some of the oversight that is currently being done by people. I will caveat that and say that I think the best prescriptive analytics. Do you need a human oversight? Machines will only get us so far. And anyone who's worked in data knows that sometimes you get some really strange answers that have to be back that holds. But I see prescriptive analytics as a way to take a supply chain system that right now has on average, I would say, like two or three weeks of just kind of play time in it where it might be late. And so we're planning a month ahead of time. We can shorten that down so that we have less inventory sitting and waiting because we've made the system more precise. [00:17:08][95.8]

[00:17:09] OK, OK, that makes sense. So it's increasing automation basically all the way along. [00:17:13][4.1]

[00:17:15] Yeah. [00:17:15][0.0]

[00:17:16] Yeah, some degree. OK. We're it. We're coming up towards the end of the podcast at this point. [00:17:20][4.3]

[00:17:20] Heather, I was just wondering, is there anything I haven't asked you that you think I should have? Is any points that we haven't talked to that you think would be important for people to be aware of? [00:17:30][9.5]

[00:17:31] I think we've covered a lot of the. Lights. I do think that the real I would like to emphasise that I think the real value of prescriptive analytics comes into kind of the preventative aspect and looking at things like we can we can be identifying the the battery, health and optimisation in different locations. We can look at fleet utilisation. We can find high risk areas and discourage travelling to those routes. So we can really we can really identify how the system is is flowing and ways to make that best practise instead of just less and efficient. [00:18:09][37.6]

[00:18:12] Okay. Okay, that's great. [00:18:13][1.2]

[00:18:14] Heather, if people wanted to know more about Heather or about prescriptive analytics or anything that we've talked about today, where would you have me direct them? [00:18:22][7.9]

[00:18:23] Yeah, people against ethically reach out to me, send me messages on LinkedIn. If you want to know more details about prescriptive analytics or what savvy does. Please cheque out our Savvis website. Sauvie dot com. And then people can feel free to email me at H. K.R. I e.g. e r at Savill. [00:18:46][22.8]

[00:18:47] Superb, superb header. That's been great. Thanks a million for coming on the show today. [00:18:51][3.5]

[00:18:52] Thank you for having me. It's been a pleasure. [00:18:53][1.4]

[00:18:54] OK, we've come to the end of the show. Thanks everyone for listening. If you'd like to know more about digital supply chains, head on over to SAP Dot com slash digital supply chain or simply drop me an email to Tom Dot Raftery at SAP dot com if you'd like to show. Please don't forget to subscribe to it on your podcast application of choice to get new episodes as soon as they're published. Also, please don't forget to rate and review the podcast. It really does help new people to find the show. Thanks. Catch you all next time. [00:18:54][0.0]

[1076.6]