Alexis
Hi everyone. Welcome to an episode of Data Dave Dives Deeper. I am Alexis, your host of this podcast, along with my dear friend data Dave. You know us both from our true podcast, Talk Tech with Data Dave. But today, we have an expert, so we’re doing a Dives Deeper.
Before we get started, I love to remind our listeners, please send us questions at talktech@d3clarity.com. any question you have for Data Dave, we would love to answer. And if you would be interested in joining us for one of these dive deeper episodes, please reach out to us at that same email address. We would love to hear from you.
Dave, good morning. How are you today? I’m happy we have Collin with us.
Data Dave
Good morning, Alexis, and thank you. And thank you for that introduction. Excellent. Yes, I’m very well, thank you. And we are joined today by the founder and managing partner at North Labs, which is a cloud provider and cloud consultancy, Collin Graves. So welcome, Collin. Very glad to have you here at Data Dave Dives Deeper.
Collin Graves
Appreciate it, Dave, thank you. And Alexis, great to be here as well.
Data Dave
Excellent. I start off with some of our normal sort of conversations. Often on this, we drill into how people got here because we find that very few people started in data and started in this world. We’ve come to love it and enjoy it and appreciate it and be able to share interesting stories. So please, let’s start out with your story as to how you got into data, and how you got into data management, data processing, and how you relate to that.
Collin Graves
Sure. I took a very unconventional track, that’s for sure.
Data Dave
I love those.
Alexis
Those are the best stories. Yeah.
Collin Graves
I was in the military when I first learned about the cloud, so I served for just shy of six years in the air force. I remember distinctly walking to an aircraft where I was. I just started as sort of a wrench-turner. I was a hydraulics mechanic.
Data Dave
Just, just one moment.
Collin Graves
Yeah.
Data Dave
Are we talking about the right kind of cloud here?
Collin Graves
We are. I’ve been in a bunch of them, but yes, the cloud as we know it.
But this was back in 2007. I was walking to an aircraft, not a very software-minded person at that time. And my father sent me an article about how Amazon was planning to open up its data centers for public consumption. So, this was sort of the advent of AWS, if you will. And I remember thinking to myself, I want to learn about this. And if you knew my father, you’d find that even more ironic. This is a guy who still likes to navigate with a Kings Atlas. Just got his first smartphone about a year ago. So we’re figuring out the texting game and things like that. He’s a brilliant engineer, but not super on the progressive side of technology, and he sent me this article very serendipitously.
I read it and went, I want to learn about this. So, I deployed to the Middle East about 60 days later and found myself just devouring the cloud. And, okay, if I’m going to learn about AWS, I need to learn about software engineering or computer programming and things like that. So, I spent the next four years sort of teaching myself how to write code. I used to fly around at my next base in a C130, and when we were flying, I had very little responsibility because I was sort of a mechanic on the ground. So, I would string up a hammock in the back of the tail cone of the C130, and I’d lay in a hammock and write code, little coding exercises, while there were 82 Navy Seals sitting, waiting to jump out of the plane.
It’s been a love affair ever since. So luckily, hitched my cart to a pretty fast horse in AWS, and 17 years later, here we are. The fire is still burning bright.
Data Dave
Nice. I’ve got many questions about C130. I spent too many hours in them myself.
Alexis
So, yeah, I’m over here going, C130 is a plane, right?
Collin Graves
You’ll recognize it as it’s got four propellers on it. It’s got the big cargo door that swings down, not as big as some of the other cargo planes in the military, but really known for landing in undeveloped or underdeveloped areas. Big, big, beefy tires on it that can handle a whole lot. And we could pack 82 special operations folks in the back of that and watch them all jump out either the sides of the plane or the back of the plane. And it was a pretty fun spectacle.
Data Dave
So, one of my most fun stories, just going back to my background, was bringing in a C130 in a clearing in the trees… by flashlight.
Collin Graves
Yep.
Data Dave
Pitch black, just with a flashlight. Bring it in the dark. Bringing in a C130 onto the ground with just lining what was our makeshift runway with guys with flashlights, and they weren’t even visible flashlights. They were all infrared.
Collin Graves
Yes.
Data Dave
So, you can’t actually see the runway. Then the plane just came in and landed, turned around, and we were on it and gone.
Collin Graves
Yeah.
If you imagine, like, little glow sticks that you could only see in night vision nowadays, technology’s improved since you’ve been in, Dave. That was too long ago. But they got those little emitter pucks, and we would make these runways in the middle of a rainforest, essentially you’re landing and hoping you don’t catch that one stump that didn’t get removed completely. And it was always adrenaline-fueled.
Data Dave
It was a tense moment when you’re staring down the runway. I know it was tense for the pilots and the guys on the plane, but it was tense for the guys on the ground, because if the plane goes off course or it bounces or anything, then you are the runway lining.
Collin Graves
Right? Exactly.
Data Dave
Back to data. So that is honestly, that’s how I got into data, was the collection and management of data from that kind of field. From that, you just studied AWS and picked it up. You came out way more recently than I did. So, AWS was only a glint in people’s eyes when I was serving, but. So, you just studied AWS and you decided that was the direction you wanted to go?
Collin Graves
I was self-taught for the most part. I actually got out of the military about nine months early, ended up going to school, finishing degrees in computer science and economics and things like that, while I was running this other company.
Sold that in 2014 as a mid twenties kid, not really knowing what a non-compete was. Signed my life away with a two year non-compete. Like you will not start a business other than maybe a landscaping company. And so ended up spending two years going back and forth between Minnesota, where I lived at the time, and the Bay Area, helping a lot of these scaling startups in the area leverage the cloud for the first time.
So, sort of acting as a cloud strategist in that capacity. And then midnight when my non-compete expired, I started North Labs. This was August of 2016, and here we are.
Data Dave
Nice. August 2016. So that’s about when we started D3Clarity as well. Yeah, we’ve been going for about eight years as well, so that’s an interesting coincidence. So, what have you been doing since? What does North Labs do?
Collin Graves
So, North Labs is a full-service cloud data and analytics specialist consulting firm. We offer everything from advisory and strategy to implementation to managed services and maintenance. I like to say we’re a data services company disguised as a management consultancy.
Data Dave
I resemble that remark, yeah.
Collin Graves
So, we really don’t do a lot of, “Hey, here are a few warm bodies to help you build a project.” We align with the executive suite within these organizations to help accelerate their data and analytics maturity. 80% of our customers, and this is not on purpose, are in manufacturing and industrial. So that’s been a very organic evolution for us.
Back in 2018/19, I sort of looked at our data and went, “Huh, isn’t that interesting?” Groups that all have complex supply chains that are building and taking raw materials, and their value chain consists of finished goods being delivered, anything from small consumer-type stuff to warships for the military.
Data Dave
Okay.
Collin Graves
Everything is sort of in that manufacturing, industrial. Industrial is a sort of loaded term because it’s biopharma, medical device, oil and gas, whatever.
Data Dave
Right, right.
Collin Graves
But that’s really where our strength has been. We work outside of that, of course, but that’s really where we spend a lot of time is now that these groups have it, systems, information technology, like your CRM and your ERP, as well as operational technology or OT systems, which are like your SCADA, and manufacturing execution systems, the stuff that makes the production floor work. There’s never been a greater need to combine those systems and derive insights and intelligence and automation from those systems to fuel efficiencies in the manufacturing process.
Data Dave
So, what do you think is the biggest demand there? What do you see as the biggest demand that’s forcing people to do things differently and move?
Collin Graves
We see sort of three core areas. The first is just more of that connected business type of experience.
Historically, there’s been a ton of silos where the executives don’t really know what’s going on on the production floor. They’re making decisions. They’re making very rough turns of the wheel with 90-day sort of lagging data. And so, having a more real time flow of data is really important. A lot of that ties into your SIOP process, which is sales, inventory, and operations.
So how do you plan against those things?
The second, for sure, is more on the line itself. So think like scrap reduction or quality control, or predictive maintenance, or things like that. How can we use data to get ahead of the things that really cost us a lot of revenue opportunity every year, like broken machines and running lines at half speed versus full speed. And then the third is IoT, and the automation that can come out of the Internet of Things, or connected systems movement, where we can actually take prescriptive action on those lines based off of the data being collected. So instead of Dave having to sit at a terminal and go, the temperature of this pump just turned yellow. I better slow things down. Now we can use data to prompt those actions instead.
Data Dave
So, the first one is connected business, right? So, we see that tremendously connected business. We see it in our everyday life, which is we, as consumers in every facet of the world, are no longer tolerant of a business that doesn’t know what the hell it’s doing.
From CRM, connected to manufacturing, connected to ERP, why did you send me this invoice? Well, I sent you this invoice because you bought 17 things and we know exactly what those 17 things are. There’s that connected business. We see that all the time, which is the pace of business that is fueling the economy that we live in.
The second one was really getting into operations, the level of increased opportunity for removal of inefficiency and waste through a production system. And the way we look at that often is how do we use data as an analog to the business process and the business structure and then get back to process engineering, six Sigma waste elimination structures like that, starting to drive that and turn that around, which is huge.
And then the third one, the way I look at that, is starting to make automated decisions based off of sensor feeds directly. So making the system fully automated by, okay, the pressure is going up in this pipe, therefore we’re going to turn this valve and drilling that through. Talk to me a little bit about what you think about the impact of the current wave of machine learning across the top of those.
Collin Graves
It’s interesting. I still view GenAI from a very wide aperture. What I hope doesn’t happen is similar to what we saw in the 2008 through 2012 timeframe, which was this tsunami of cloud interest where everyone just said, “I don’t care what type of workload it is, I don’t care where it lives, I don’t care if it’s 25 years old, get it to the cloud.” And we very quickly realized, okay, terrible path, very expensive, lots of risk and.
Data Dave
I think… so let me stop you there. Yeah, very expensive. I thought the cloud was supposed to be cheaper.
Collin Graves
Yeah, there’s gives and takes, right?
Certainly, if you are saying, “Hey, take this 15 year old as 400 IBM system and just find a way to shim it into the cloud and will find efficiencies,” you’ve got another thing coming, right? There are trade-offs. There are the capex intensity of on-premise data centers, but perhaps lower OPEX costs or operational costs versus, okay, lower CAPEX, larger OPEX, but maybe more agility or possibilities. Organizations need to figure out on a workload-by-workload basis which is going to work best for them. The answer is not just put everything in the cloud and it’ll magically fix all of your problems. It’s not a panacea of pain-killing; it creates other pains.
Data Dave
So it just moves pain around.
Collin Graves
Right. And especially if the systems are old and kludgy and full of skeletons, that just gets exacerbated when you’re no longer on a capex basis where you can just amortize down the cost of some databases. And, yeah, it’s slow, and yeah, it kind of.
Now you’re paying by the second for this old kludgey system, so you’re being constantly reminded of how kludgy the system is. So, as far as AI and machine learning, the biggest things that we’re seeing are still very much in the machine learning realm, as opposed to more of this GenAI conversation, but I can touch on that. But machine learning for us is a huge trend analysis for operational analytics. Like, I talked about, the biggest thing for us is how can we use forward-looking projections to add context to a business situation. And here’s what I mean.
Yes, in predictive maintenance, the way things typically work today are Alexis is looking at a screen. It’s got 50 green squares on it, meaning all 50 machines are humming along. Great. All of a sudden, one of them flips yellow, and Alexis picks up the phone and calls the mechanics to report to that system and see what’s going on. But instead, with machine learning and context, we can say, “Okay, Alexis, this machine just turned yellow. But actually, this other machine is projected to turn yellow in 17 minutes from now. And that’s a much more critical system. And here’s where it’s going to turn yellow. The vibration in this certain case of this pump is starting to increase.” Now, Alexis has context when she sets those priorities to the maintenance folks to go, look, you’re going to notice machine 14 is yellow. I want you to go hop on over to machine eight. It’s still green, but not for long. And that’s the one that’s going to cause the downstream bottlenecks. 14 can live as yellow for the next couple of hours.
So, all of a sudden, think of that efficiency that’s gained, as opposed to Alexis sending a team out to 14, they have to go get the parts. Then eight turns yellow, and you go, stop what you’re doing. I don’t care if it’s your lunch break, get over there. It just adds a whole lot more streamlining to something so complex as multi-line, multi-facility maintenance. And that’s just one example.
Data Dave
Right?
Alexis
So, I just had that conversation with Preston, Dave’s partner, my boss. He was talking to me about like a call center. He gave me that same breakdown that you did. The idea that what if the system knows that when Dave calls, it’s likely going to be a, b, or c? And here, let me give you a, b, and c ahead of time, so that when you get on the phone with Dave, you’ve got the answer as quickly as possible. And if we can teach the system to start to expect those sorts of things, so that idea can be applied in so many different areas. I love hearing the different application ideas here. Dave, I didn’t mean to stop you. I was just.
Data Dave
No, you’re welcome. No, that’s perfect.
I was just going to say. I was going to give another analogy because I was talking to a gentleman the other week who was flying helicopters in Vietnam and what they were doing even back then. He’s so excited by the machine learning aspects. Very similar to you, because what they were doing then was when they got a helicopter in for service, because it had a bullet hole in some part of it, they would predictively look at the parts of the aircraft, because while it’s on the ground, if a part is going to fail in the next 15 hours, change it now versus wait and bring it back in again. He was actually. We’re having a similar conversation. He was talking about how they were doing it back then and how they could do it in the seventies, and comparing and contrasting that with using machine learning, etcetera, to do the analytics, to bring it full circle.
Collin Graves
Yeah, they’re the pioneers, not us. We’re just, we’re just carrying the baton now.
Data Dave
But that was interesting.
Alexis
I’m super enjoying listening to the two of you because I keep drawing connections between you. I’m like, oh, look, you’re both founders of your own companies. Oh, look, you were both in the military. Oh, look, you both have huge backgrounds that didn’t really start in data. Oh, look, oh, look. I’m really just kind of enjoying listening, so I just wanted to throw that out there for our listeners. If you didn’t catch all those, there are quite a few here.
Data Dave
So, you alluded to another thing, which is the trinity of ML, but you said something, and I can’t remember exactly what it was now, but around you. Hope it doesn’t go the way that we did with the cloud back then. Elaborate on that a little bit. Just because we do see these fads in technology where everybody rushes down a path before they understand it and the pendulum swings and then it swings back and eventually we get into that sort of stable state. Is that the kind of thing you’re alluding to? Where were you going with that comment?
Collin Graves It is now. Just for clarity’s sake, I do believe that Genai is a transformative technology and I don’t want that to be lost on anyone. I think it’s an amazing feat. I think it’s got staying power, and it’s going to really change a lot of businesses in a lot of ways. But what I’ll say is, just like the hype cycle of the cloud that I alluded to in the early 2010s, we’re seeing that here where people are sort of saying, if you are not on this gen AI bandwagon right now, you’re going to be left behind.
I don’t think that’s true in the next five to seven years. I truly don’t. And the reason for that is we still don’t understand as an industry where this is going. So it’s not like wealth management where you go, okay, we’re going to add technology in and we’re going to have robotic portfolio automation. And if you don’t have that and you’re still manually doing your clients portfolio allocations, you’re going to be left behind for sure. But we understand the future state. We’re still trying to figure out what this future state looks like.
And the biggest thing is just seeing some of my peers in the space saying, hey, we have this offering for you, misses customer, and it’s going to take four years and it’s going to cost you $16 million, and you’re going to have a GenAI center of excellence. We’re going to GenAI all of the things in your organization, and won’t it be great? I view that as a losing proposition. And the reason why is imagine a year ago an organization saying, we’re going to go all in on Chat GPT-2 and they spend all of that time in this waterfall approach, implementing GPT-2 and tuning the LLM and really getting it crispy. Well, nine months later, a year later, now GPT four is out, and the people adopting GPT-2 and having tuned that system are like this. Things way better than this two ever could have been. We’re going to be in an accelerating maturity curve for the next five to seven years with this stuff. There’s going to be more and more LLMs coming out literally every day. Every research institution on the planet is working on their own LLM. MIT, Stanford, Harvard. So it’s about starting small, just like people should have with the cloud in the early days. Small and pointed. If you don’t have a sound ROI thesis, a sound business case, I say learn about it, but keep it on the back burner.
Data Dave
So, I agree. I agree with that. So, you’re saying start small, focus and have a defined problem set that you’re solving for and you’re drawing that parallel back to the cloud, back to other things, which is really what I’m hearing there is do it for a reason.
Collin Graves
Yes, exactly.
Data Dave
Know what it is, understand that reason, and go for it. Magic doesn’t usually happen.
Collin Graves
Right? We’ve never seen a technology panacea. Right? Think about this for a second. In 2016, the average industrial organization had something like 68 or 75 data-producing systems in their organization. Today, that number is 148. Years later, that number is 140. Every single one of those platforms that a company adopted claim to be a panacea. A magic pill.
Data Dave
Yes. Right, I’m with you. I use the term silver bulletin. So many companies have been sold silver bullets, and none of them have slayed the werewolf.
Collin Graves
There’s not been one werewolf killed.
Data Dave
No, there’s not been one werewolf killed.
Collin Graves
Right, right, exactly.
Data Dave
It’s missed the target or it’s been whatever happened to it. Because one of the things we always say is we don’t sell you a silver bullet, we’ll take all those silver bullets that you had because everybody owns two or three of them in the same space.
Collin Graves
Right, right.
Data Dave
You probably own all the technology you need. We just have to augment it in a couple of little places. But really, it’s teaching you how to use it and teaching it how to work for you.
Collin Graves
I can’t tell you how many times we have customers come to us and go, look, the data initiatives sound fantastic, Colin, but first we need to go through this ERP migration effort. We’re going to migrate to a new ERP. And I go, well, your current ERP isn’t 30 years old, running on a DoS green screen. Tell me why.
Well, it’s got this one additional module that we really think could impact the company. And I’m going, look, your ERP is the heartbeat of your organization. Changing ERPs is quite literally a heart transplant. No one has ever finished an ERP migration. Gone. Geez, that went well. We stayed on time and on budget, right?
Data Dave
No, never.
Collin Graves
Right?
Data Dave
Never.
Collin Graves
So how can you use data systems like you and I are talking about, like our organizations build to serve as almost the connective tissue.
Data Dave
Absolutely.
Collin Graves
Between these systems to go, don’t change, augment, be additive, and give yourself that sort of composability, that modularity to adapt over time. But you still have the same heart, right?
Data Dave
Yeah, exactly. And focus in on the applications. What have you got that application doing? What is the actual job of that ERP system? If it is failing and not being optimal in that job, then hire another something to do that job, but leave it doing what it’s good at and then wire it together into the ecosystem where everything has a defined job. If you define the job that those systems are doing, two things happen.
One, your ecosystem becomes stronger, but it also becomes more loosely coupled because your interfaces become defined and it ultimately becomes easier to replace any one component because you know what job it’s doing.
Alexis
Thank you so much for joining us today, Collin. I absolutely have enjoyed this conversation. I’ve loved hearing from both of you and just kind of drawing those parallels that I was talking about earlier in my brain. So very, very good. Thank you both so much for being with us today.
Data Dave
Yes, thank you, Collin. It’s been a pleasure. Let’s stay in touch and keep moving forward.
Collin Graves
Absolutely. Thanks, Dave. Thanks. Alexis has been awesome.
