On today’s episode of Thriving at the Crossroads, we have Allen Sulgrove. He is the director of the digital business unit for General Datatech, a company that basically helps people find out what they don’t know, and then how to use that data. He talks to us about what his company does with the innovation of things using the Internet of Things (IoT).
Listen to the episode here:
Welcome to Thriving at the Crossroads. I’m your host, Amber Christian. Today, I am pleased to welcome Allen Sulgrove, director of the digital business unit for General Datatech. Welcome to the show, Allen.
Thank you, Amber. It’s a pleasure to be here.
Allen, tell us a little bit about General Datatech. This is going to be fun. For our audience, we are live in a studio together. We’re not on my normal recorded line. Hopefully, there will be lots of laughter and a good time today. Tell us a little bit about the General Datatech and some of the types of problem that you solve.
General Datatech is a twenty year old hardware system integrator. Think network compute, data storage, we sell all that type of gear. In fact, we’re the number five largest Cisco reseller in North America. We like to think we’re kind of a big deal. In the past few years, we’ve created what we call the digital business unit, which is what I run. It’s all about analytics, big data, IoT. One of the problems that a lot of our customers came with is, “All right, we want to do big data. What do we do?”
We always help out build data centers, building out different server farms and things like that. But we decided that we really needed to start helping them with the data. They had all this data out there but they had no idea what to do with it. We’re trying to help them to figure out what to do with it. As opposed to trying to find random solutions using IoT and sensors, let’s work with things that we know about. We’ve got wireless access points that you can get connected to the internet through your phones, tablets and laptops. That actually gives you a lot of information as a company.
For instance, if you’re a retailer and you want to learn more about what your customer patterns are. Just by having that access point, its talking to all these different wireless devices on a regular basis even though you’re not connected to it. Now I can tell how many people are in my store or how long they’re staying in there, if there’s passersby. Now I’m seeing I’m getting more passersby than people coming into my store. How do I get them to come into my store? If you don’t know what the passerby’s are going by any tangible data, what would you able to do with it?
It sounds like you help people figure out what they don’t know. We don’t know what we don’t know. Gathering that first step, as you think about big data, we’re generating and gathering reams of data. It sounds like you help them figure out what they don’t know so that they can know it and then build up how to know it. Is that a fair statement?
That’d be a fair statement. The great thing about it is typically they already have all these devices in place. They don’t necessarily have to buy something that they don’t own. They could use the functionality of what they already own or they might need a newer access point to collect that data. I’m not going in there and saying, “Hey, you need to buy $10,000,000 worth of data or devices so you can get this data that you can use for this one little thing.” No, let’s start with what you already have, and then if you want to get more context from that, maybe we do add more sensors and BLE devices or RFIDs or things like that.
Can you walk me through a practical example of what it would look like for a customer that’s just starting? What they might start to gather, how might that evolve, what types of decisions they may make from that?
I used retailers as an example so I’ll try to figure out another. One of the things that we do a lot of work with is higher education, the university, the campus of the future. First, we need to figure out what type of data can they collect. We talked about the Wi-Fi access points. Most colleges have cameras all over the place for safety. We look at their networks and Wi-Fi. We can see whose coming in and out of their network. Are they students? Are they faculty? Are they people in the guest network? Now we have information about students, faculty and guests. What do we want to know about that?
One of the big things for universities is matriculation. I want to know how many of my students start of as a freshmen and actually graduate. They can get that information pretty easily through their different systems. But how can I impact that? Or how can I get ahead of a student that might not be going to class on a regular basis, might not be using the campus facilities? This is where those Wi-Fi access points could come in again. Now I can see if John Smith has got a 4.0. If he’s staying at home the entire time, I don’t really care. But I have John Doe over here rocking a 1.5 GPA and I can see that he doesn’t come to class or might hangout at the cafeteria all the time. Maybe the Dean of Students staff can see information like this to go out to John Doe and others like John Doe and get them into class and more engaged, or get them out.
Or figure out what’s going on. College is a time where a lot is changing in your life, there’s ups and downs and all around, is something going on? You could expose a problem. If someone’s a regular attender and now they stopped. What caused that?
Fantastic use cases. I’m sure there’s been many of times where a college student gets into that partying lifestyle and just drops off of the face of the earth. I’m sure not only the faculty but also the parents would like to know.
There you go, there’s an add-on service. That could be quite a revenue stream. That’s a great use case. You start to gather this information off sensors, say at these wireless access points. What do they do? Pop it in a database? What do you actually do with it? I’m gathering it. Let’s take the next step. Where do I store it? What does that look like?
It could look like a lot of different things. It could look like putting it in a Hadoop cluster, it could be in a data leg, it could be in a sequel database. It could be whatever; it just depends on what you’re looking at. For the most part, a lot of the sensor data is really not that big. You don’t necessarily have to ping those devices every two seconds to get the data, maybe once an hour or once every half an hour. And then, we’re just looking at certain variables of the data that’s important to us. You don’t need a big data center to collect a lot of that. You just collect what’s important. It could be as easy as putting in to a sequel data base. Or if you’re looking of some real time type applications and like to run predictive analytics, I think you know where I’m going with that, you can employ something like a HANA in-memory platform.
We bring it into HANA or Hadoop, wherever, and then we can start to apply our predictive analytics, prescriptive, etc. down the chain to it. We’ve talked about a few different use cases. Are you seeing some of this adoption around IoT and devices? Is it everybody? Is it particular industries that seem to be investing the most? Who seems to have the most pressing problems that some of these IoT solutions solve?
Pressing problems, that’s always an interesting question. There’s a divide between nice to have and need to have. Need to have always falls when there is a dollar associated with it or a euro. If there’s money tied to it, the industry or the business will find a way. Retail is a good example. Retail in a Box is something that we do where you put in Bluetooth beacons and start using cameras and access points in concert to really get that granular customer information.
We’ve seen a lot of things in the hotel entertainment type industries, hotels and casinos, casinos especially. They want to know where their gamblers are at and how to keep them there and keep them there longer and attract those passerby’s that we talked about. Where the real money is, when we start looking at oil and gas, they want to be able to do things with sensors, which they’ve been really doing for years. IoT is not really anything new, but the way that we can get the data and use it and get to that real time standpoint, that’s the new part to it. They want to be able to put devices on pumps and sensors on pumps, oil lines coming through. They want to be able to put equipment on oil derricks out in the middle of the Gulf of Mexico. Some interesting issues too, guess what, it’s a big platform in the middle of the water surrounded by all kinds of metal things. Try to put Wi-Fi signals through that and get the data back.
The other thing that we’re seeing in the IoT is doing processing at the edge. As opposed to trying to get all that data from these disparate locations, I’ll do the processing of the data at the edge and then bring it back to wherever I need it to do the final mile of the processing. I’m taking gigabytes of data, shrinking it down to the megabytes that are important to me, and then be able to push that over a satellite line.
I’ve heard the term before but I haven’t heard much context around it, processing at the edge. Is it common across the board, only in cases where we have signal strength or bandwidth issues? How widespread is that concept of trying to do more of that processing on the edge to shrink it down? Talk about that a little bit more.
Clearly it’s applicable to those remote areas, those offshore areas, places where the bandwidth isn’t there. But there’s also an argument to be made that even when you’re not on sight, on premises, even if you have a good piece of bandwidth, it’s always better to do processing at the edge. That way, you can, again, get closer to real time by not having to move gigabytes of data over the air, internet or a direct link and do it there. The value is really in those remote locations with less bandwidth, but there’s a continuing argument to do that more in any remote site. That way, you’re getting closer to that real time.
The reality is there’s latency and other internet connectivity issues. I’ve seen it when system get shifted, infrastructures hoisted in the cloud. There’s other challenges now with networks or things where you can run into, if you’re pumping all the data through all the way and then presenting it and changing it at the end, you can run into a lot of challenges. You’ve got to do it upfront and then bring it through because there’s just a bandwidth capacity issue. We keep pumping more and more through that capacity. It’s not getting smaller. It’s going to grow exponentially.
It strikes me that there’s going to be a lot more of that just to keep up with the fact that we’re shoveling more data through, from devices, from everything. We’re to the point now where if I didn’t have my unlimited T-Mobile network, “What am I going to do? Where’s my signal? I need my data.” That’s what we’re all getting used to and having this ubiquitous access to be able to do it. In practicality, you’re right. In the middle of the Gulf of Mexico is probably not a place where you’re going to have a strong internet connection.
You’ve talked about a couple use cases. Any success stories that have really surprised you, where there have been either a huge leap or something that really took the customer by surprise? Sometimes they start with, “Yeah, we know exactly what we want.” Any real unusual ones that took you or the customer by surprise in a solution that you were able to put together around some of these devices and these data?
I’m going to talk about not quite a customer success story but more of a lab story. We worked with a local internet provider here in Dallas that said, “Hey, we really want to do some cool stuff for our community. We’re going to let you hang devices off our own poles, we’re going to let you have space on our data center and we want you to go crazy.” We got in touch with Cisco about it, they were very interested in it. What we’re able to do is create this lab as the showroom for basically anybody. Here at General Datatech, we’ve got lab data center. We’ve got a theater, we’ve got a network operation center, and we’ve got all these places where people can see technology in action.
But when you put it out in the community, it’s like a 10×20 square block area. We’ve got things hooking up to parking meters. We’ve got cameras. We’ve got wireless access points for the entire range. There are Bluetooth beacons, RFID tracking. You put all these different pieces together and create these beautiful visualizations. Now you’ve created a space where our customers and Cisco’s customers and the partners that we have involved in it can bring everybody and go, “I didn’t even know this was possible. You’ve talked to me about it but now I actually see it.”
We even have certain shop owners that will have a screen up in their window that show some of the analytics that are coming through. Just a nice to have. You can see the traffic patterns. The Dallas police department it has microphones strategically placed all over the city of Dallas. When thing happen, when gunshots go off they can immediately triangulate where those are happening. You put those access points and the cameras and all those together, the security in that neighborhood, nothing’s happening in there because they can see that this place is being monitored. I think that was one of the very cool things as we started to bring customers into that environment, whether they were cities or universities or companies with large campuses. They now see it in action. It’s not this concept, it’s not me walking in there going, “Trust me, it will work.”
“Where have we heard that before?”
“If you just wait until the next release.”
“That feature will be developed in six months to a year.”
“It’s in the disclaimer slide I just put in front of the presentation.”
“It’s subject to change at any time. Functionality may or may not be released with this release.” Sounds like the city of Chicago could probably use that right about now. It’s been crazy listening to what’s been going on there. It’s interesting though to hear what the city of Dallas has done and what you can with that. I was actually thinking, as a consumer, I would love it if a business put that monitor up so I could see when the people are going to be there, so I can go when they’re not going to be there. I have a whole different thought process. When is it not busy so I can get a really good service, etc.?
I’m not sure if you have this where you’re at, but there are many billboard signs around Dallas that have emergency room queue times.
Interesting. I don’t know that we have emergency room queue times. Usually you can get them on the websites for some of the local hospitals. They will say how long the urgent care wait is and some of them. You have access so you can know where to go.
“I cut my arm off and if I go to this emergency room, they might be able to sew it on quicker.”
Tie the tourniquet tighter. It’s an extra two minutes but there’s no wait. Nice. You’ve been building some of these devices as well. You have some SAP integrations with some of your IoT as well, right?
As someone said to me, every time I say “HANA” I get a dollar. I’m going to say “HANA.” That gives me two dollars. I mentioned Retail in a Box earlier. We’re putting these solutions together. I won’t necessarily say we’re building devices. VC is one of our partners where we can put in a full compute storage switches and all that good stuff. We put the platform for all these sensors to communicate back to. A lot of times we’ll use HANA or Hadoop, depending on the real time needs for HANA and depending on how much historical data we want to keep on Hadoop. It’s not necessarily we’re building devices but we’re building turnkey solutions.
Retail in a Box is one of them. Are there others you can talk about? Or is that all under wrap?
Campus of the future is a good example where you’re marrying all those technologies. We’re helping colleges out there that are building new campuses or extending on their campuses. We’re putting that infrastructure in place. That way, the access points are there, the lines are run, the cameras are there, they’re part of the construction of the campus as opposed to being an afterthought. Most students will have an ID card that they can swipe in to get into different places. You can put an RFID in that. As they’re swiping in there can be a ray so that you can look at the traffic patterns.
We talked about the access points, we talked about the cameras. The cameras are really getting interesting because we’ve gotten to the point where there is facial recognition and it works well. From a security standpoint, making campuses as safer places, that’s a big part of it as well. Another thing, on the SAP side, one of the things that we are prototyping here is putting RFID in our warehouse. We’ve got probably $30,000,000-$40,000,000 of equipment in our warehouses at any given time. What we can do is put our RFID tags as they’re coming in, they can be pre-processed in SAP.
We are an SAP customer, we use HANA and Concur and SuccessFactors. But with those RFID tagging and arrays around the warehouse, I can pre-process my inventory as pieces come into the warehouse. With the arrays of our RFIDs, I can know where any box is or any piece of equipment at any given time. It removes the need for inventories and cycle counts.
The time and productivity and everything that’s involved in having to go do all that manual activity. Especially also I would think you’ve got a lot of high dollar value items there as well so it’s good to have all those RFID tagged and tracked for other potential issues.
Other things that we’re working on, the really cool cutting edge things, I’m not sure if you’ve seen the Intel drones that do the light shows? What we’re working with now is, for instance, water tower manufacturer. They need to be able to do inspections of their equipment. They can now use drones to go do inspections. Let’s take it up a notch, let’s automate those drones. As opposed to having to have somebody that actually controls it or monitors it, we can have a pre-designated pattern working around and then report back. Now, someone can look at it, and then we can use some of the analytics on the video to see discrepancies on what’s up there. Maybe a bird’s nest is up there or there are some sensors there that need to be cleaned off, or even worse it might have some lightning damage that wasn’t showing up. Those are the cool stuff that we’re doing these days.
It seems like IoT has a lot of good uses around maintenance and prediction of failure or things of that nature. It can help you observe things that you wouldn’t always see because you’re not going to send a guy up the water tower every day. Its ability to help you predict and then get in front of that maintenance, and deal with problems before they happen. That to me is an interesting concept.
I live in Minneapolis and we had a rather large catastrophe a number of years ago when the 35W Bridge collapsed. It strikes me as very interesting when we get into the things that affect our infrastructure for where could devices and things actually help predict or catch things that would be much harder to do with the human eye. That was a series of factors that caused it, but for me, that’s intriguing to prevent some of those types of things that unfortunately happen every now and then. Then we only figure out in hindsight what was the series of things that happened. But if you had more capacity to predict in real time, gather more information to be able to see the effects of what was happening, it seems like some pretty amazing use cases you could have for it.
SAP has a project called IT Operation Analytics. There’s a study from Ponemon Institute. They say that it costs, on average, $9,000 a minute for a data center to be down. Think about that. Extrapolate that out to an hour, that’s $500,000. All that being said, it’s a lot more cost effective to prevent an outage or issue before it happens than after the fact.
Absolutely. With some of these interesting platform solutions, Retail in a Box, and things that you’re building. A lot of these are new and just getting established in the market place, particularly in SAP integrations. We’re hearing a lot more about. Now that we’re a few years in we’re hearing more and more use cases for it. I like to ask about the practicality of, as you build some of these new solutions, just where you’re at customer-wise. For our listeners, we have a rating scale in terms of how many people actually use some of these solutions or how far you’ve gotten into the market.
I interview startups that don’t even have customers yet and are so early in their life cycle, to established companies that have lots of customers. The rating scale is ABC, for alpha-beta customer, i.e. “I’m getting my first live customers. We haven’t really done any integrations with this. We’re just working on the first ones.” Then we have our rating scale D, which is one to five customers on the solutions, which means, “Yeah, I’ve got a few live, I’ve learned some lessons, we’re fining it.” Then I have category E, which is existing customer base. “I’ve got more than five which means I kind of know what I’m doing at this point or should know what I’m doing at this point in integrations with customers.” Where would you say you are at in terms of your customer base?
It depends. I would say we have existing customers that have, if we take the drones out the equation, we have well over five customers. We’re a worldwide company. They have solutions in place that can get this type of data today. The question that might be better suited is: How many customers are actually using some of the advanced functionality we talked about? I would say we’re probably in D, maybe a couple E, but that’s growing every day. Especially when we’re talking about new construction, expansion of data centers, moving of companies where we’re starting from the ground up. That’s coming along a lot faster than anything that we have to rework or replace. From the drone’s standpoint, we’re still alpha-beta.
Drones, part of the challenge has really been around the regulations and it’s such a grey area before you could even register. Is there anything that I haven’t asked you that I should have?
I think we’ve really gone around the horn. It’s a fun space to be in right now. It’s called the “something gap” and basically it looks like your standard linear progression except there’s a big gap in that progression. I think we’re finally getting to the point where were jumping the gap. A year or two ago, we were just before that gap and we all had these cool things that we could do. But now we’re finally jumping the gap. I think we’re changing IoT into actual things now.
Maybe yesterday or last year, we would’ve called the things that you can do with cameras IoT because you’re getting data, “I’m just monitoring camera.” Now, I’m getting traffic patterns and anomaly detection. That was IoT. Now that’s just part of the camera functionality. We’re getting to the point where we’re getting less of IoT and sensors out in the ether to more productization. I think that’s probably the one thing that we haven’t talked about. It’s getting less of this nebulous kind of thing and now it’s becoming more productized. As soon as it gets more productized, the more velocity we’re going to see and the more things that we’re not even thinking about are going to show up.
We’ll see the transformational leaps in the technology, in the use cases, in the creative things that people will do with it. But you have to have those first ones that bridge it for a lot of people to see the possibilities. “That was a really creative use of that data.” It starts to expand your brain a little bit once you can see it, because so many of us are visual and we need to see things to understand them. We always talk about that in design and systems and how often you can spec and design something out to your heart’s content and then you actually show it to someone and they’re like, “That’s not what I meant. That doesn’t really do me any good. Let’s try this again.”
That sounds like project 101.
It is and it happens a lot. It seems like until you get some of those first ones, once you get those with that real transformational sense to them, you were talking about the campus of the future, the first buildings. All of a sudden now other people can see, “Oh, that’s what you do with that.” Once you have all that data and everything in that one place, that should really start to unlock a lot more use cases, I would hope.
As would I. The other thing to think about as people are launching into these IoT projects is do it an phased approach. Don’t try to boil the ocean. As with most projects, don’t try to do this big bang. Let’s start with the basics. Get used to that, get working with that, get the data we want out of it and then start looking at those other variables that we can do. That’s the better way. Because one of two things will happen. One, it will lead to successful implementation of the use of this data. Or two, you’ll realize that that’s not what we thought it was going to be and we saved some bucks on the way.
Abandon that before we get too far and before we sink too much money into it. I like that thought process of an incremental approach. When I look at analytics, and I do a fair amount of writing about some of it, in the working capital space, about getting your programs, getting those first baby steps established. In the analytics row, I talk about you don’t get straight to prescriptive analytics. You start with just describing what you have. Get your arms around it. Even understanding the data has got to be a challenge. You start to get into that IoT world and now we’re in a different series of data than the traditional invoice payment, etc. Now we’re dealing with sensors and different levels of data. We’re having to reconstruct things. It’s just different. There’s a skill set gap or skill set learning.
In predictive analytics, I can see something is happening or I can see a trend now but what’s the context? That’s the one thing that is great as predictive analytics. Grade A machine learning and artificial intelligence could get to that context level. But we still have to figure out what it means.
Right, we still need that human adding that additional value to that information to put context around it.
We still have a job.
Did you hear that? Anybody who’s listening is like, “How are we going to remain employed here?”
Figure out context quickly.
That’s your key takeaway, context. I have one final question for you. I like to ask everyone on the show this question. At some point, I need to create a map of all the very interesting answers I’ve had from people about their favorite destination they’ve ever travelled to in the world and why it is your favorite?
Unfortunately, I haven’t been in a lot of places in the world. I haven’t made it to Europe yet, across the pond. What I would say, and I say this a little bit begrudgingly, the happiest place on earth, Disneyworld. I say that because my daughter just absolutely loves it and my wife absolutely loves it. I can deal with some of the frustrations of the land of the mouse. It’s actually a really cool place to go. I was there a few times in my youth so it’s cool to go back and go see Figment. He’s a pink-greenish dragon.
Is he a part of my imagination?
You hit it. He’s Figment of your imagination.
Wonderful. Thank you so much for joining us today. It’s been a pleasure having you on this show, Allen.
Thank you very much, Amber. Have fun.
Did you know that at Consult Ace we do monthly webinars on all things SAP? If you’re curious to learn more, check out our website at ConsultAce.biz/resources/webinars