SAS’ vision of transforming ‘a world of data into a world of intelligence’ seems more relevant than ever with the incredible and seemingly endless amount of data we are producing with the Internet of Things.
For those who aren’t familiar with SAS, they are one of the largest software companies today, providing a suite of analytics solutions, of which 92% if the Fortune 100 companies use to help access, manage, analyze and report on data to aid in decision-making. This week on the Mr. Beacon Podcast, we are working from home with Saurabh Mishra, who heads up Product Management at SAS for their Internet of Things offerings. In this episode, we talk all things SAS: their company culture, what offerings drive their business, the evolution that brought about the Internet of Things division, and the real life use cases they are solving today in transportation, manufacturing, retail, and supply chain.
2:14 What is SAS?
5:20 Company Culture
9:32 Revenue Model
13:39 Analytics Delivery Methods
16:34 IoT Analytics
21:35 Distributing Analytics in an IoT Environment
25:50 IoT in Manufacturing
32:18 Event Stream Processing Platform
41:30 IoT in Supply Chain Logistics
43:01 IoT in Retail
47:25 Closing Thoughts
The Mr. Beacon Podcast is sponsored by Wiliot, scaling IoT with battery free Bluetooth.
Steve Statler 00:17
Welcome to another episode of The Mr. beacon podcast. This week, we are talking to Saurabh Mishra, who heads up Product Management at SAS for the Internet of Things offerings. So, Saurabh, thanks very much for joining us on the program.
Saurabh Mishra 00:38
Thanks, Steve. Thanks for having me. I'm delighted to be here on this podcast with you.
Steve Statler 00:44
SAS is is an industry behemoth giant. But you're also a private company. fiscally speaking, you're not, I couldn't buy shares in SAS, even though I would like to because you're very successful. But I think you've got a very interesting position in the market. I imagine pretty much every large fortune 500 companies using your software somewhere in a variety of analytical functions. And I always kind of thought of you in terms of the hardcore analytics work. That has been part of our IT business for decades now. So I think it's really interesting to talk to you because you're focused on the world that we live in, which is the Internet of Things. And I'm really looking forward to getting to know what the use cases are that you're seeing, and how SAS has responded with the tools and how they're different to what we might have been using 10 years ago, when the Internet of Things was just a glimmer in a few people's eyes. Before we get into that, let's just explain to people a bit more about sass. How many of you are a little bit about the company? Can you give us a quick intro?
Saurabh Mishra 02:14
Yep, absolutely. So in fact, sass might be the largest, privately held software company around today. So as you said, We are pretty focused on analytics. So that's all we do. So that's kind of the focus, since the sensor date was founded. So we've been around for for 43 years. So SAS was founded in 1976, and still being led by the same founder of the good night. So if you've had, as you can imagine, pretty consistent leadership from him throughout this time. And, yeah, analytics is really what we do, right. So so there's a gamut of capabilities that we can talk about from that perspective that we have. But we are worldwide, a global company, so about 14,000 employees roughly, globally, think we have presence in about 58 countries right now. Six r&d centers across the world, the headquarters is where I am based in so Raleigh area, North Carolina, beautiful campus there. Ah, as you can imagine, right now, I've not been able to go to for the past several weeks. But that's where that's the headquarters. That's where the CCO r&d group is present. And we have a, you know, we are a privately held company, but we do release the revenue and all of that. So it's just a little over $3.2 billion. And we've had consistent growth for a long time. So So yeah, really, really fortunate from that perspective. And I think Dr. Goodnight is, is very focused on making sure that there is a there's a culture of innovation that continues to exist. And I continue data points from that perspective that I think around. I think that if you if you analyze the amount of reinvestment, some of these large corporations put back into their r&d, SAS is probably one of leaders. It's the last number that I remember was, we put about 25% of that back into r&d. So that's been his his focus from from long time. I think that is shown over the time, right? I mean, we've kind of withstood a number of changes, a number of disruptions in the industry. We're still pretty relevant today.
Steve Statler 04:35
Well, I think, you know, my perception of your company from the outside is there's a lot of loyal people that because SAS treats its employees, while you've always done really well in these best places to work and I think maybe because your capital is so tied around the brainpower in the company, that That's kind of a natural reason for having that kind of culture. But it kind of predates the excesses of the, the, you know, the Googles, and the the whole Silicon Valley thing where all of that stuff went crazy, but it is. Tell me a bit about how being a private company and impacts the way people are treated and just talk a bit about the culture. From that perspective.
Saurabh Mishra 05:31
Yeah, sure. I mean, so, I mean, it is a lot of focus on making sure that as a workplace we are, we're definitely, you know, as a top workplace, right. So, you know, you talk about these surveys. So I think this is something that, you know, even before a lot of these tech companies have all that have the sprawling campuses and have a lot of facilities on the campus for employers, I think SAS has probably been doing this for a long time, even before this thing became a thing. Right? So the culture is, I mean, it's a technology company at the core, right? So there is a, you will find a lot of just just very creative, innovative people. And I think the, that's the whole aspect of people tend to stay here is a function of many things, right? I think one of the core aspects or drive that is, is this just culture of innovation and creativity that exists, right? I mean, there is just such a strong, you know, kind of peer group that you have here, in talking to people who have specialized degrees, specialized experiences, it's a very rich, interactive experience when you start to work with them in let's say, a project or any kind of initiative. So I mean, that personally, for me, I think is a big reason, right? Well, why why I like being here. It is a win, win. There's obviously things that everybody cares about, but no kind of opinion put everything on the on the on the table. But that to me is one of the strongest reasons why why I think people stick around here, when you kind of add other things. And that just kind of becomes a big a big setup, why people tend to stick around here, right? I mean, there is there's a good balance from from a work life perspective, you're working on cutting edge products, right. I mean, SAS has been doing analytics before analytics become became a hot thing. So we're still working on I mean, the market is kind of pretty hyped around machine learning and AI, when at some level, this is what we've been doing for a very long time. So being able to just kind of work on that innovation is been has been a critical factor in driving, making sure that employees stay here. And being a private company, I think there is obviously some level of flexibility, right? Because you're not not answerable to state, you know, stockholders or, you know, the stock market in general. Right. So there is there is flexibility that that you have, we are obviously, you know, in corporate and you're looking for growth and all of that stuff, but there is certain flexibility that the leadership is cognizant of, and they put that to the best use to make sure that we stay ahead of the curve. At some level, we invest in areas which may be, which may not be immediate, short term, that's a revenue areas, but something that we see going forward with the waters. So I think it's kind of a package state, at some level, there's just a, just a number of these factors that of course, go a long way in making us what we are from a from a good employer perspective.
Steve Statler 08:36
Yeah, I can think of public companies that, you know, when the pressure is gone on, activists, activist investors join the board, and suddenly they look at all these benefits the customers get, I'm sure you can, sorry, employees get, I'm sure we could improve our profitability if we economize a bit on the health care plan, and all that sort of thing. And you guys are somewhat immune to that? Well, okay. I don't want to wallow in that too much. Otherwise, everyone's gonna get jealous and annoyed. So let's talk a bit more about the business. And, you know, where does this revenue come from? I've always thought of you guys as doing the kind of the really deep analytics as opposed to, you know, the visualize simple slicing and dicing visualization tools, the I'm sure you a broad set of things. Where is where does most of the money come from? If you look at if you were to slice and dice sasses business, how would you split it up in terms of the revenue that's being driven by the different functional tools? And how do you look at it in different industry areas? Are there any patterns there that are of no gift
Saurabh Mishra 09:50
yet? I can show you so to share with you some patterns, right? I mean, there's multiple dimensions that we can look at right? So if I if I look at it from there, say See a product stack prescriptive. So so I don't have a breakdown on how it gets allocated. But I'll tell you that at a very high level, I think of our product stack as two large buckets, right. One is kind of the the analytical platform, which is essentially horizontal in nature. And you write that no, we we do a fair bit of the advanced predictive type analytics, right? I mean, analytics over the year has become a kind of a generalized term, everybody has some level of analytics that they're able to talk about. So we obviously do that. I mean, there is fair bit of just being able to do visualization reporting, that's, that's bread and butter. But we go beyond that with predictive analytics, optimization, data mining, AI, machine learning, streaming analytics, that is an area that I think both of us hear about, and we'll probably talk about that more. So there's just a set of these horizontal platform type capabilities that we have, that we offer to the market. And then what I think of as a second bucket are more more industry specific solution type capabilities that we have. So some examples are going to be like, we have solutions for predictive maintenance, solution for production quality, customer intelligence, fraud detection is a healthcare fraud, anti money laundering. In fact, if I heard an anecdotal example, that if you swipe a credit card somewhere in us, and if you get detected that there was a fraudulent in swipe, then chances are that that was detected by a SAS lock. So we have a number of those industry type solutions, that can also contribute a lot. So I think that the revenue is kind of based on both of these pillars. We have customers who come to us saying that, you know, we have a strong analytical skill set of our own, what we need the tools and the platforms to equip them. So they would be the customers that would be that would get positioned the platform. But we have other customers who come who are more line of business type customers who like okay, so I have a manufacturing scenario here. And I'm looking to streamline my production quality process. So do you have a solution for that, so that that's how we kind of tend to think about from a product perspective. So that's, that's one dimension. dimension is just look at our our revenue. So historically, I think, at some of these large industries that you can imagine like banking, manufacturing, health care, government. So these are all big contributors to that revenue. So we have, I think, a fairly good split across the board in terms of this revenue. You write about, like the fortune 500. Right. And so SAS tends to it is a move from an enterprise scale perspective is a recognized leader, right? So if, if there is a need for a large enterprise company to think about, you know, war, where would I get scalable software, which would kind of be something that we can grow with, as opposed to like a small PLC that we start and we are done with, they will turn to SAS. So we have a large stack of these fortune 500 customers. In fact, one of the stars I remember is that over 92% of the Fortune 100 companies are SAS customers. So that's that's kind of how I would distribute. It'd be idling right now.
Steve Statler 13:39
And how is the delivery of those analytics changed? Over the last few years? The clouds obviously become a much bigger factor here. Do you? Are there any? Is there any preference or predominance in terms of how you monetize your software? Are you simply licensing software that people can run on premise or in the cloud? Or how is delivered these days?
Saurabh Mishra 14:07
Yeah, so I think that definitely is an area which is gone through a lot of shift over the last several years, right? I mean, so we, we are agnostic of the underlying infrastructure, right? So we don't have any requirements that you have to be funding this versus that. But if you just look at trends, historically, a lot of our software is deployed on prem. Right. So these are large organizations that run their own data center, their own IT staff, and they would, they would license the software, and they will deploy it on prem. And that is still a pretty common pattern for us. A number of our large customers still manage their own data centers, although the underlying deployment technologies might be changing, but that's where they're declining software. But as you can imagine, obviously, a lot of our customer bases are turning towards cloud and a lot of new deployments are just starting with the Load, which is fine, because again, like I said, we are that agnostic of the underlying infrastructure. So we can deliver software in a variety of ways we can we can deploy software on prem, or customers can deploy a software in their cloud, which can be a private cloud or a public cloud. We can host the software for them, right? So if they're just looking to solve the problem, and not necessarily looking to get into, how do I deploy it, how do I configure it. So we have a notion of sex cloud, where we can actually host the software for them. And they can just access the software. Or if they want to completely abstracted, he can also position almost like reserves as a service for them. So let's say let's say you are a retailer, and you're trying to do this effort, profiling of sizes, right. And so you know, how many sizes of a particular skew Should I carry throughout the season, when there's probably a problem your school's required us all, maybe once in a season type, right? So if you want, give us data and have us create that profile, like a one time thing, so that could be a model where we just do it preserves the service delivery. So there is a lot of flexibility in terms of how we deploy our software and how we, let's say have a kind of a engagement of from a licensing perspective.
Steve Statler 16:27
That's good. Okay, let's get into the Internet of Things piece, which is your bread and butter on a daily basis? How does IoT analytics differ from everything else that you guys have been doing for a very long time? Why call it out as something different or special?
Saurabh Mishra 16:48
Yeah, that's a that's a great question. Right? So I mean, at some level, I mean, I've been doing IoT analytics for a long time, right? Because if you think about the underlying data that has existed, I can talk about sensor data, some of the textual data, unstructured data that has existed, we really started thinking about IoT a few years back. So I was, I was part of the retail product management group within SAS. So that's kind of my background. Before SAS, I currently work for retail companies. And retail software is one of the things that I've done in the past. So when I came to SAS, I did that I have written solutions that we offer to large retailers, let's say from a price optimization perspective, or in that space. So that's kind of my background. But we started looking at, you know, one of the initiatives that was happening was actually a beacon initiative, where somebody was installing beacons in the building, and they were doing just take a fun app to track, you know, how people are moving around in a building, where are they spending time and things like that? And so I was involved with that. And then we just started thinking about this, like, you know, what's going on with this IoT space? Right? I mean, we obviously have dealt with this from the, say, manufacturing customers have had this data and our energy customers have had this data. And you've dealt with that. But we just saw enough of organization happening within the market that we thought that there's a number of reasons why we need to approach it. But some sort of a focus here, as opposed to just regular SAS. And those reasons were a few I can put in is that there was a strong focus on on that we saw on being able to run on devices, or edge, as we call it, right? Because we quickly saw that all of this data is not going to make its way to to the data center or the cloud, there has to be more of a distributed processing. And that was for us, right? Because we actually, if I step back on the product side, although we are known for the analytics part of it, but we also have a fair bit of data management capabilities in our product stack. And the traditional model has been that you have data as an organization, that data could be in files and databases, message queues or whatever. Before we can do analytics on it, we do data management on it. But the driver was that you create this big data lake or a data warehouse that's analytics ready, and then you analyze that data? Well, guess what, that model was not going to work for IoT, because all of this data was just not going to make its way to this data lake that you had set up. So this is where we needed to kind of understand the problem. Rather than bring the data to the analytics, we had to take analytics to the data. So this was a completely different way of looking at how we've done analytics. So it required a different approach just from a product positioning perspective. So that's that's one big factor. The other was the way we go to market right I mean, so we we go to market You know, the customers large enterprise customers, as we talked about, but from an IoT perspective, what what quickly became apparent was that analytics was one of the things that an IoT use cases required, right, it was an important thing. And we do that well, for to truly solve an analytics use case, when I think of it as a, as a stack that has number of layers, you need to have hardware, you need to have connectivity you need to have so there is data producers. So there's a number of other complimentary layers that need to come together to kind of fit this puzzle. And our go to market should adapt so that it can co exist in this new ecosystem paradigm. So that kind of led us to this notion of partner ecosystems, right. And that is a huge part of what drove the formation of no separate group within SAS focus on IoT. So I would say the two driving factors was just the fact that in the underlying data required a different way of tackling from an analytics perspective. So you know, distributed analytics was a big focus. And then the go to market required a lot of partner ecosystem type drivers around it. So those two kindnesses were the necessary drivers and creating a more separate focus group for IoT. That's how we are thinking about it. And we've been at this thing for for two years now, doing it, this
Steve Statler 21:33
is very interesting. So give us give us some examples of both of those aspects, then, how can you put a bit more meat on that bone of how you might distribute the data gathering in an IoT environment, any examples you can think of?
Saurabh Mishra 21:49
Yeah, there's a number of examples. So for coming, I'll give an example actually, that brings both of these aspects, the whole distributed analytics and the partner ecosystem to even go to market in the mix. So we started we're working with GE transportation, which has since kind of changed around given some of the changes that happened at GE. And it's actually owned by a company called walk tech, for they actually have a platform that they position to their end customers who are tier one railroad operators, so large railroad operators in the US and in Canada. So I mean, obviously, think about think for the locomotive going traveling in, you know, remote places, less connectivity, no connectivity, all of that stuff, but still generating a lot of data. Right. So that requires a scenario where we needed to position something on the train itself. So they have, they have a device that they actually called train on the train, it's a large computer that sits on the train all the subsystems on a train connect to it. So there's a fair bit of data coming into that thing. And they wanted the ability to be able to process that data locally, without the requirement of constant connectivity to the cloud. So they obviously run the back end in the cloud somewhere. So that required us to position the, you know, the streaming analytics capabilities that you may have talked about in the past. So we actually run the engine, on the hunters device itself. And in this case, our so they are not, they're not an end customer. For us, they are actually a partner, right. So they have their platform, and we augment their platform with the streaming analytics capability that is embedded inside their platform. And then they when they go to customers, we kind of go with them. And we joined forces to kind of implement customer scenarios. So that's an example where we are actually on the train itself, analyzing data. And this can cover a gamut of use cases, right? Because this is, again, this is more of a platform positioning. So we're not just solving one use case. But when we started doing something like this, I mean, some of the use cases that we got were like, you know, this is a sophisticated analytics platform, it can do like real AI on the edge. But the use cases that we were dealing with, like, you know, you locomotive has been stationary in the same place for over five minutes. Can you create an alert for that to the driver? Can we, so you have to just think about it, right? I mean, sometimes from from a tech perspective, we tend to kind of think about all of these sophisticated computer vision and AI at the edge, but really, was may require far simpler things that can really hurt if a locomotive has not moved for over five minutes in the same location.
Steve Statler 24:45
Well, I think anyone that's been watching Breaking Bad using this time and lock down to catch up, we'll know why it's very important to be able to track a locomotive stuck in one place because they might be right Important materials?
Saurabh Mishra 25:03
Exactly. Because if you think about it, the rationale is based on cost, right? So for for these large locomotive traders, fuel is the number one cost. So any small dent they can make on the fuel consumption is as big dollars saving for them. Right. So that's where their mind is, in terms of cost savings.
Steve Statler 25:25
Yeah, these these these devices that you're monitoring our platforms, aren't they I imagine you can extend the same thing to automotive and these cars. I mean, I'm just amazed at how much software is now in an automobile. And so I imagine the same rationale there would be would apply to that. What about manufacturing? I'm really interested to know, what does IoT looked like in a manufacturing context for you guys, you do some work in that area that?
Saurabh Mishra 26:02
Yeah, we do. So manufacturing has been actually one of our strongest verticals, even going back in the history of SAS. And that continues to be a pretty strong verticals from an IoT perspective. So we're doing a number of interesting use cases, I can tell you some examples. So one example is a company that makes that makes wall boards, and they had this conveyor line where the where the wall boards are moving. And there will be times when the wall board should kind of collide with each other. And some of these collisions could result in a jam. And, you know, if that happens, that's bad, right? If the conveyor line gets jammed, it stops. That means the operation stop, and then you know, there's a, the these companies have a process to kind of get around it right. And things stop somebody comes in, and you may who knows how long the thing is done. So that's loss of productivity. And some of these environments are mean, you probably appreciate that they're not designed to have modern infrastructure from day one, right? So you're thinking off like, no, how can I fit something in this environment without ripping and replacing, because they don't want to stop everything and kind of just get into an infrastructure project at a time. So one interesting use case that we that we are doing is this, this happened to be a security camera on top of this conveyor line. So it was a security camera, right, I mean, just meant for security purposes, per se is picking up the field as the conveyor is moving around. So we were able to kind of grab that feed for the security camera, and retrain the model. So this is a classic computer vision use case where we are actually doing object detection. So we look at these wall boards, and we are saying how the wall boards are moving. But we go beyond the just the regular object detection, we're able to kind of ID, the wall boards. So we assign a unique identifier to each wall code will determine the edges of the wall board. And then we do some basic geometry to find the center of the wall board. And then we then start to track the distance between two successive wall boards based on the centroids. And then you're able to kind of forecast like if the distance is getting shorter than there is a tendency for this thing to collide. So that is we used to then preempt a collision, right? So we are able to kind of start to color that now there is a red thing coming if the distance is getting shorter. So that is, I think, a really good example of a sophisticated computer vision technology working in today's manufacturing environment where we didn't have to rip and replace the infrastructure using a security cameras feed, to act to detect the motion of the wall board. And to get to a situation where we are able to preempt a collision, and create an alert so that they can react to it and somebody can separate them out or whatever is the recommendation. So that's, that's one example that I'm like, this is really cool. This is commodity stuff, right? I mean,
Steve Statler 29:06
yeah, I love one of these. I love that my job is lesser now. But you get this excuse to get on site into these manufacturing facilities. And I remember being on size at a place that clothing one of the biggest clothing companies in the world. And it was almost like a science fiction movie. You go into the sleepy town and you're beautiful pizzas, traditional ancient town, and you go a few miles outside and you essentially go into this underground complex, which is seems like miles long, where it's all robotic. And you see just incredible technology that automates the flow of, of how things are produced. And I think, you know, we get into this technology business because we're basically nuts and then you get eatery that's used to making making exhaust pipe. So this is a, this is a ceramic filament that goes into your exhaust pipe and just see how these are made. It's a joy. But it's kind of almost like 19th century stuff furnaces and that sort of thing. But Oh, actually they using angle of arrival to track various things.
Saurabh Mishra 30:27
That is right. Absolutely correct, right. Yeah. Last year I was up in the New England area was being a customer, a shoe manufacturer, and they were trying to modernize their shoe production line. And this was, again, one of those scenarios where, you know, this is an existing production line. And they were trying to figure out what's the best way to kind of speed up certain processes that were holding up the production line, but they did not want to rip and replace. So we were thinking off, how can we augment this existing infrastructure, put cameras here, put some RFID trackers to be able to track like, first is to be able to measure where the slowing down is happening. And they didn't, and they were using, like a like a, like a shoe would be marked like this is how long it takes for a shoe to move from position a to position B because they will mark that shoe and somebody would they go, I got that shoe. It, I registered it at this time. And now I caught it at this time. So this is how long to so there is obviously some level of automation that you can bring in that process by being able to scan things or being able to, you know, just kind of have some tablet where people enter this. So a lot of these initiatives these days are starting with us and working with some of these customers, where other partners come in, that's where the partner ecosystem is such an important part. Before even we get to the analytics part of it. There's, there's a almost a phase where we are thinking about Okay, do you have the right infrastructure? And if not, who are the other partners in our ecosystem? Who can bring that say the right RFID tags are a gateway technologies, industrial PCs, you name it, cameras, things like that, that could first be defined, but first be installed to define your infrastructure before we can do any analytics with it.
Steve Statler 32:14
All right, let's double click and get even more specific about IoT. So you have this platform called the Event Stream Processing platform. Tell us a little bit about what that is the rationale between for how that's structured and the kind of projects that you're how it works, and the kind of projects that were that you're doing with that, you know, what is an Event Stream Processing?
Saurabh Mishra 32:44
Show? Yeah, so look, Event Stream Processing actually falls in this space called a complex event processing. So si p, right. I mean, so the underlying idea behind the SI P is that it's, it's designed to process complex events that are that are coming to it constantly. So I think the the origin of si P is from capital trading, because the spreads coming high frequency trading happening, and you're kind of kind of processing that and analyzing that to figure out what's going on. So that's that was the genesis, if I remember, right, for these platforms. So we actually invested in this time back right before we even formalized the IoT group that SAS because one of the one of the gentlemen who actually visited SAS for us the RND, behind this actually has done a number of si peas in his previous life. And he he knew that the combination of a CP where we are talking about just even processing with analytics is going to be a killer. Right? So he came to SAS with that vision, and he brought out start form, which because seven Stream Processing ground up. So his goal was to be able to bring world class CP, which is essentially being able to process millions of events per second. and combine that with analytics. So that's what resulted in SAS Event Stream Processing. Now, it's actually and I think of anonymously bias, but I think there's an incredible product, it is it is flexible, are you do actually have an engine, which is small enough, you can actually put it in a small Docker container that can give you an AI can run on a small gateway device. We have use cases that are happening on Raspberry Pi type things more more from a PLC perspective. But we run the same engine on that train example that I talked about, be on the same engine in a cloud scenario. So another customer example is a Volvo Trucks. They have a fleet of their trucks in North America closer to 100,000. Trucks sending telemetry data all the time. And we run this engine at the the back end environment there and analyzing this data all the time. So this is all the telemetry data, this is fault codes. So we're able to kind of understand and figure out like what's going on with these trucks. So when the truck pull in, we have already analyzed the data, and kind of helps reduce the diagnosis and repair time. But But going back to Event Stream Processing, so it's really this, this platform that allows you to bring analytics to data in motion, right. So again, the different paradigm earlier, we used to bring data to rest and then applied analytics on it. And we're turning that around, we are bringing analytics to data in motion. So that's the that's the core idea behind it. And, and it's a versatile platform, it allows you to do simple things like you know, if you wanted to do SQL type logic on a time motion. And so you could have a number of incoming data, data elements and you're doing SQL type logic, math thing, you can do that, you're looking for a particular pattern that in my stream, if event A happen followed by b followed by C, in this time interval raise an alert can do that. But where it gets really interesting is that we're able to combine all of this with, with machine learning now. So for example, we have algorithms that are trained to detect outliers, right. So even if you haven't trained this data, if you have just an A new source of data that started to throw this or these are items will understand the steady state of the asset could say from where the data is coming from. And then once it understands the steady state, it can begin to report against that steady state that I've seen a deviation and how bad the deviation is. Now, in this case, there's enough for you to go and do something with it. But what we truly want to do is that over time, if I have enough data about this asset, then I can actually create a predictive model, that when this asset fails, these are the parameters that are indicative of that failure, then I can actually bring that inside the Event Stream Processing. And that way, as soon as I begin to show signs of that kind of behavior, I can alert somebody. The other aspect where ESP comes in is that if a lot of times when we're going to customers, they have their data science teams working with open source, right, so they're doing Python stuff, they're using other open source frameworks. And which is fine that because that's that's just the nature of what what we live in right now. So we can also take new open source of Gotham, and operationalize that within Event Stream Processing. So we can create a pipeline where we are getting data in, you may be using some SAS ESP functionality to manipulate the data. But in a subsequent step two, you already have an analytic that you built that said, with Python, you can actually use that, and then do whatever else you have to do. So it can combine SAS plus open source and pretty easy way. So that is a platform that has a delicate environment, drag and drop, we also have programming interfaces, if you are python programming, you don't have to deal with the drag and drop UI environment. The engine is his waters, the the core processing, and that's lightweight, so it can go on the edge small devices. And we can also take into the cloud. So it's pretty versatile from that perspective.
Steve Statler 38:18
So understand how you could use this for predictive maintenance, you start to see patterns that require attention. And how am i is this something that brands or retailers could use as well what sort of use cases
Saurabh Mishra 38:37
with? Yeah, so we have a few use cases going on with retailers right now. So then the use cases around, let's say, next best offer customer intelligence kind of use cases. So again, combination of location, prior purchase history, could be data elements that you are combining to create, let's say in what's the next best interaction you should have with this customer. So this is where we kind of tend to link together some of the SAS offerings that we have in the retail space. Combine that with the seven stream processing. So we have customer intelligence capabilities that you know, be positioned into retail and other hospitality. But the use case that I'm talking about is focused on there. So you have your retailers app on your phone, and you enter a store and you are spending some time near a particular product, particular aisle. So all of this data can be analyzed in real time. And if you will, the Amazons of the world are doing it online, you could replicate the same behavior in the store. So this is a kind of a next best offer kind of use case that that may have some real time requirements that can support. The other use case that we're seeing is around data. Inventory part of it, because shrinkage continues to be a huge drain was from some of these retailers. So to be able to understand where my inventory is, in real time is of big importance. So a simple scenario that we were looking at is that sometimes that this was inventory missing on the shelf, but you might have inventory in the in the back room. So to be able to track, what's the status of the inventory on the shelf, be able to generate some alerts in real time could be of value. So so we're seeing some interest in that. And that really is something that can be extended in multiple fronts, because ultimately, you're talking about some level of computer vision, that is keeping a watch on the inventory. I'm seeing use cases and this is where can retail especially it gets a little interesting because of privacy concerns. But now I'm coming across use cases where people are saying that we could actually monitor the expiration date of perishable products. So as because, you know, you've probably been at times when you've gone into a store and like notice stuff is expired or close to expiry and bad days for the customer. So be able to create alerts for that. We actually have a simple scenario where we talk about missing inventory, misaligned, product misplaced product, be able to catch all of those scenarios in real time and create alerts. So there's there's a few use cases in retail that we're dealing with that.
Steve Statler 41:26
Yeah, this is almost segwaying into another area, which I'm interested in, which is logistics and supply chain and so forth. Are you seeing IoT analytics in that area? I mean, so much if we're thinking now about now we're trapped in our homes, that whole supply, we're very, very acutely tuned to how well a supply chain is working, how long we have to wait for something to be delivered. And things missing on shelves in, in, in supermarkets. Do you seeing any IoT applications in for analytics in the supply chain? No.
Saurabh Mishra 42:07
New Faces in quality monitoring, that is a that is an area that there is I mean, I tend to look for patterns where there is a need for you to monitor something constantly, or there's a need for you to monitor a constant stream of data, right? Or, for us to kind of really have an IoT story. So cold chain monitoring is one of those use cases. So fair bit of requirement there for you know, healthcare. pharma companies have requirements around that. We came across an interesting use case from a from agriculture perspective, where when they're transferring livestock, they have requirements to maintain a certain level of temperature, humidity, and that kind of a combination. So poor quality monitoring is kind of the closest one that comes in.
Steve Statler 42:58
Very good. Well, fascinating to hear what you are working on and how SAS has evolved into serving this area, which we find so interesting. You just before we sign off, you did mention Bluetooth beacons and and the retail what retailers were interested in with that, can we go back to that is anything else that you can add to? What is it that retailers wants to know from analytics perspective about Bluetooth beacons?
Saurabh Mishra 43:30
So I mean, I, I tend to think about that. They're interested in the location data, right? I mean, location is a pretty cool element of what what they're interested in, and being able to kind of understand where the customer is, right? I mean, that customer journey is really important. So there's a there's a couple of aspects where I think, I think Bluetooth or location plays a role. One as people are kind of navigating a physical store, right, being able to understand where customers are, what path are they taking in a store, how much the time that they spent in this area versus the other area, just being able to understand that. And that has ramifications both from what offer you might serve up to that customer? Or in the long run, how should you design stores or together by I mean, is there a different layout that's going to facilitate facilitate that that motion is important. The other aspect where that is important is the what we call omni channel analytics, right? Because predominantly, most of the retailers also have an online presence. So if you are, Steve, if you are in the store, and you've also been online so to be able to kind of combine that word pair what you did online kind of has to have that understanding before you serve up an offer or an attraction to customer in the store that you just signed defied through a location Or due to timing is important. So there's a fair bit of analytics that goes in there in terms of like, what's the what's the right offer the right promotion or right interaction in general, that you can offer to the customer? So that I think is a is a pretty broad area. I mean, the the one area that I think about from an IoT perspective that, no, I try and understand every time I think about use cases is that they have to do at a very simplistic level one of two things, there has to be some either cost saving, right, you have to save cost at some level, or you have to make the experience better at some level, right. So I think instant order priority in time we talk about cost savings, that will that will trump the experience for most of these organizations. Great. So that's where we see slightly more acceleration in terms of the industrial IoT side, we talking about assets, you're talking about reducing downtime or slowing or reducing in scenarios where production has to be down, things like that. That because there is a direct impact of something like that on cost, right. So going back to the fuel example, even if I can prevent the fact that a local motor has to be idling for x minutes, has a direct impact to the bottom line of that of that operator. The challenge with some of these retail use cases that I see, especially when you're talking about next offer or next gen traction perspective, they're making the experience better, right? So in the long run, yes, you can tie that to revenue when we increased conversions or things like that. But in the short run, that's a little bit of a challenge that you have to have a champion in the organization that understands the benefit of that, and can take take this message along inside the organizations, otherwise, you might be better off focusing on things that do have a direct impact on on the revenue. So loss prevention or shrinkage or things like that. So that's a piece, you know, kind of a simplistic lens that I put on i times just to kind of bifurcate these two words.
Steve Statler 47:10
Yeah, I mean, it's great to having simple mental tools that you can use when you're when you're looking at the business case, find these things. I like it, whether you can save money or make money. Very good. Well, Sarah, thanks so much for your time, I've really enjoyed the conversation, great to hear what SAS is doing, and IoT, and more, more broadly. Appreciate it.
Saurabh Mishra 47:35
Yeah, thanks for having me. And I enjoyed the conversation, and I look forward to seeing what we can do together. winboard.
Steve Statler 47:49
If you would have three songs that you were gonna listen to on a long journey, which ones would you listen to?
Saurabh Mishra 47:58
Hmm, okay. So let's say if I had three songs, I would do. I mean, these are all probably going to be bollywood songs. So it depends on the audience if they relate to it. And I know there is a there is a song that came out a couple of years back. It's called they'll be a gala. sisal is a ballad, soft song. So so I like that a lot. And my daughters have been trying to play that on the piano. So I've been listening to that. And I do. I used to listen to a lot of Bryan Adams when I was growing up. So so 18 till I die is like, you know, one of my all time favorites. Okay, I would I would keep that I actually used to listen to a lot of Nirvana when I was growing up as well. So So Kurt Cobain is, you know, one of my absolute favorites. So any song of his would be another one that I will okay.
Steve Statler 48:53
Name, name, name, one that springs to mind.
Saurabh Mishra 48:59
It I was just listening to something that a bit was I didn't realize it was his death anniversary. And I forget the song. He was playing the song on the stage. I just I just don't remember it. I mean, I can I can have the music playing in my mind right now. But I I just don't remember the direct. So if you aren't, if
Steve Statler 49:20
you see united spirits, it's not one of these. I can't remember.
Saurabh Mishra 49:23
I don't know. It was not that. Yeah, I'd have to look it up. But. But
Steve Statler 49:29
fair enough. Well, that's good. I think Bollywood is amazing. And I feel like our modern video industry, there's a lot to Bollywood that the sort of classic dance sequences where the whole cast in a movie is dancing. And then you kind of watch MTV type stuff. And it's like the same thing.
Saurabh Mishra 49:51
Yeah, it's like this larger than life, kind of a setup where you have 200 background dancers and maybe changing costumes made way. That is a pretty bold word for it.