Want some Doritos? For years, you might have only been able to get one or two snack chip flavors. But right now on Amazon, you can find numerous varieties, from barbecue to nacho cheese, spicy sweet chili, or Late Night Loaded Taco. And this is really just scratching the surface. There are now dozens of flavors of Blue Diamond almonds, including blueberry, smokehouse, toasted coconut, sriracha, habanero BBQ, and wasabi and soy. So how did this happen? It turns out that some of it is a tech story. Thanks to breakthroughs in automation at both the plant and warehouse level, companies are able to create and ship more varieties than ever before. On this episode, we speak with Ryan Harlan, the director of business development at the E Tech Group, about the rapid changes in the industry over the last decade and how that turned into so many more consumer offerings.
This transcript has been lightly edited for clarity.
Key insights from the pod:
What is the E Tech Group? — 07:01
What triggered the multitude of flavors? — 08:02
How a new flavor is developed — 11:27
How has tech changeover changed in the last decade? — 14:35
What opportunities for risks are opening for OEMs? —16:14
How technology's feedback affects product — 19:31
How the nut industry has expanded — 21:19
How easy is the flavor-switching process? — 24:14
How different countries get a head start on experimentation — 26:18
What’s the cycle when launching new products? — 28:23
How to profit off of DTC companies with an idea — 29:55
What’s next in snack food manufacturing? — 31:21
What’s the labor demand in snack manufacturing? — 34:31
Will AI be a force amplifier for robotics? — 37:51
How distribution manages an influx of varieties — 40:01
---
Joe Weisenthal (00:17):
Hello and welcome to another episode of the Odd Lots podcast. I'm Joe Weisenthal.
Tracy Alloway (00:22):
And I'm Tracy Alloway.
Joe (00:23):
Tracy, I think it's because of all of the illegal cannabis stores in New York City, but I...
Tracy (00:30):
This could go anywhere. All right, go on.
Joe (00:31):
I know, I know. But the thing about them — and I guess it's kind of funny, and I guess it kind of makes sense and maybe it's also kind of cool — is that the other thing is, so they sell various weed products, but they also seem to be the favored vendors of all kinds of exotic snack foods in New York City.
Tracy (00:49):
Right. So if you go into a corner store, probably…
Joe (00:54):
Or just look in the window...
Tracy (00:55):
Yeah, exactly. You don't have to go in, you can just look. You will see weed. Probably a lot of these stores are still unlicensed. You will see energy drinks, you will see maybe some Elf bars, things like that.
Joe: (01:09)
Nicotine, yeah.
Tracy (01:09):
And then you will also see lots and lots of snacks. It's like the ultimate vice shop, right?
Joe (01:15):
Yes, for a certain type of person.
Tracy (01:18)
You can satisfy all your cravings, or almost all of them, in one shop.
Joe (01:24):
Yes, absolutely. Not only do they have a lot of chips and snacks, [but] they have the type of snacks that used to be you would get only when one of your colleagues went abroad, right? So you'd have a colleague [who] took a trip to the Philippines or took a trip to Hong Kong or something. And then they would come back and they'd be like ‘Oh look, I brought Matcha flavored Kit Kats or Mapo Grilled Tofu flavored potato chips or durian flavored potato chips.’ And then everyone is like ‘Oh they're so cool.’ But now they're everywhere in the stores here.
Tracy (01:54):
They definitely are more available than they used to be. My experience with this personally is probably a little bit different from a lot of other people. As someone who, you know, spent a lot of time in Japan. I mean, I have vivid memories of going into like a Don Quixote, which is this big kind of department store, specifically to buy flavored Kit Kats. They were always in the basement and there were like dozens of flavors even in — this would've been like 2001, 2002.
And I feel like flavored Kit Kats are to me probably what shrimp is to that character from Forrest Gump. I could sit here for the next 20 minutes and just rattle off all the flavors that I remember. I love Kit Kats.
Joe (02:42):
So in doing research for this episode, I came across this great website, Taquitos.net, which basically reviews every single snack. And so right now they have links, and I don't know if all the links work because I get the impression that some of these flavors are cycled in and out. But there are links and reviews of 197 different flavors of Doritos that theoretically currently or at some point were sold on Amazon. Doritos Blue Grilled Steak, Doritos Golden Cheese flavor, Doritos Late Night Tacos at Midnight, Doritos 3D Crunch Chili Cheese Nacho.
I mean, and the list just goes on and on. It is nothing like when I was a kid in the US where there [were] like Doritos and Cool Ranch Doritos and that was it. We are living in a golden age.
Tracy (03:27):
So I disagree, a little bit, on this point. So yes, there are lots more flavors that exist, but as you just pointed out, they are not always available. And okay, in New York, you know, you can go into a weed shop and there's a novelty value to buying some exotic-flavored potato chips alongside whatever else.
But I feel like what this really kind of comes down to is this really, really intelligent marketing strategy where you create all these new types of flavors and you get people talking about them. And maybe you don't produce them in huge quantities, but people like to talk about them. People like to, you know, take pictures where they're eating Matcha-flavored or Sakura-flavored Kit Kats or whatever. So for the companies, it feels like this is kind of a win-win.
Joe (04:17):
Well, I agree with that. And you do see this like on Twitter a lot. ‘Oh, Doritos just teamed up with, you know, Jack in the Box. Or Doritos just teamed up with so and so,’ and then there's like a short-run thing. So I do get the impression some of them are. By the way, I just came across [the] Doritos Gold Peking Duck flavor.
Tracy (04:33):
Oh, that sounds so good.
Joe (04:33):
Sounds so good. All right, we're going to have to try it. Anyway, it does raise the question of how did this happen? I mean, it's great. I love it. I mean, I don't eat that many chips, but in theory I love it. How did we enter this world where we just have so much variety? And although a lot of them maybe are short-run, a lot of them do exist. You have to work a little bit. But while we were waiting, I did order a 12-pack Doritos pack and a 16-pack different flavored Kit Kats. So they are getable, you know, at various times.
Tracy (05:01):
They are getable. I still have questions about the overall distribution strategy here, but it does seem like there is this trend toward a much wider and more exotic variety of flavors. And so I'm curious how that came to be, what is the corporate strategy that's actually driving it? And also the technology, right?
So how can it be that a company like Kit Kat — well, Kit Kat's not a company. In fact, I think Kit Kats are made by both Hershey's and Nestle depending on what part of the world you're in? But how is it that they're able to do, you know, dozens and dozens of different chocolate flavors and variations?
Joe (05:37):
Oh my God, check out this flavor. Doritos Roulette. Have you heard of this?
Tracy (05:41):
This episode is just going to be us like looking at Doritos...
Joe (05:45):
This is great. This is such a good idea. Doritos Roulette originally came in a nacho cheese version offering randomly placed hot chips among regular nacho cheese ones. That's such a good idea, like some excitement. Anyway, to me, this is a golden age. So I think we need to understand the sort of infrastructure and science and technology that has created this era of chip abundance. And I'm really excited…
Tracy (06:07):
Chip abundance!
Joe (06:08):
Snack food abundance.
Tracy (06:09):
This could be a different type of Odd Lots episode, but go on.
Joe (06:12):
Yeah, there's no shortage here. It's the lack of shortage that we're blown away by. We are going to be speaking with Ryan Harlan. He is the director of business development at E Tech Group, a control systems integrator. And he works with the various food and snack companies that are rolling these out across their assembly lines. And he also works on distribution. So Ryan, thank you so much for coming on the podcast.
Ryan Harlan (06:35):
Thank you for having me.
Joe (06:36):
So I think we're at a golden age, Tracy's a little skeptical. Where do you come down to this question?
Ryan (06:41):
I think we're almost in a golden age. I think we're getting there. I think within a lot of these snack food companies [they don't] quite think we're to the maximum capacity for variety.
Joe (06:51):
We still haven't reached it...
Tracy (06:52):
The best is yet to come!
Joe (06:53):
The best is yet to come.
Tracy (06:54):
That sounds good. Wait, so maybe before we begin, can you explain what it is that you do in the snack food industry, so to speak?
Ryan (07:01):
Sure. So we work on the industrial automation side. So we're an engineering services firm and panel shop. So we build control panels that contain the programmable logic computers, the PLCs, that have the programming that ties all of the various machine centers within a line that create the snacks.
So you'd have your fryers, dehydrators, fillers, cappers, baggers, palletizers, whatever you would have on the line for whatever type of snack you're producing. The control system has the program [that] tells what to start when, what to do, which process. And then eventually at the end, you get your product and then send it off to the warehouse where a different system picks it up.
Joe (07:42):
So I want to talk about, you know, the technology and how we got to this era. But first I want to actually establish the premise of the conversation. Is it true though, regardless of whether we've reached the true golden age or not, that we just have a tremendous more variety of snacks than, say, when I was a kid and I think there were like two flavors of Doritos?
Ryan (08:02):
Yeah, that's absolutely true. And a lot of the snack production is local. So if you're buying Frito-Lay chips, you're probably getting them from a factory within a couple hours from you, for the most part, if you're in a large population center. And these factories have been around for 50-plus years in a lot of cases, and it's a local supply chain. So the change from two flavors to multiple flavors with a similar population typically in an area is due to enhanced production and throughput.
Tracy (08:29):
On that note, when did this become a thing or start to become a thing? And what was the proximate trigger or development that kicked off — I mean, I imagine there was a company that was probably at the forefront of this and maybe it was the maker of Kit Kats or maybe it was a potato chip company or something like that — but what was it that kind of set this development in motion?
Ryan (08:53):
I actually think these soda producers were on the front lines of this for doing single runs. So you could find dozens of different Mountain Dew flavors going back 20 years, right? And limited time offerings. And soda is really easy to do this with because you have your recipe, you clean out your pipe, everything is the same. It's a really simple industrial process and it's a lot more difficult when you're dealing with snack food comparatively.
Joe (09:18):
Before we even start talking about variety though, how is a chip made? Let's just go think about like a single bag of Fritos. I like Fritos. You know when we think conceptually, what is the process that existed probably 40 years ago, 20 years ago, 10 years ago in which we got Fritos? What happens to make a Frito?
Ryan (09:35):
So if you look at the ingredients for a normal Frito, I think it's corn, corn oil, and salt, if you don't have any other seasoning on it. So you're going to have your ground corn product, you're going to fry it. You could make a Frito at home, right? You could deep fry some corn flour and oil, pull that together in different shapes. You can squeeze it into the shapes and fry it.
Joe (09:51):
Tracy is going to try that now.
Tracy (09:55):
I’ve made Frito pie. I have not made homemade Fritos, but maybe.
Ryan (09:58):
I don't think they would turn out quite the same. But it's very simple, right? So you're frying the corn, sending it through and that's really it. So that hasn't changed for the actual production of a Frito. You've had enhancements on the consistency of the shapes, the volume you can do at once, the timings for various fries, and how you might move it through. But the Frito of 50 years ago looks a lot like how it's made today.
Tracy (10:25):
So you're talking about the actual manufacturing process. But I mean that is, in some respects, kind of a later step in these snack food product cycles. And the very beginning is, you know, I imagine there's a flavor lab somewhere at a large snack food company. And you have all these people who are brainstorming ideas, and then maybe there's research and development that goes into thinking about a specific one. There's consumer testing, product testing, seeing how people respond to it.
It feels to me like maybe one of the big changes that has happened is that people have also figured out that you can cut down a little bit on this huge R&D process just by producing a bunch of different flavors. So it's not incredibly important to your company if you unveil this new flavor and it falls flat, so to speak, if you're unveiling a new flavor every other month?
Ryan (11:27):
Right. And I think what they've gotten very good at is, so they would have small-scale versions of their factory, or digital versions even now, where you would produce something called the digital twin of your factory line. And you can simulate the R&D process through there for the actual production and then push everything through digitally to the factory.
So you can say ‘We know it's going to take this long to cook, we want to add this much flavor, we want to add this much salt, garlic powder, chili powder, whatever you're going to add to it, sugar.’ And we can test all that digitally. We know how it's going to flow through the line, and then send that recipe directly to the production factory to send it through a test run there.
And doing that process, I think has sped that up a lot and they're using ingredients they already have, right? It's just a different ratio. They're not affecting their own supply chain as much as they're affecting their potential total throughput when they make these runs. Because you do have changeover time, you do have recipe management, traceability concerns that you have to do for various potential recalls and things like that. So there's an opportunity cost, but it's pretty limited if you feel confident you can get this run in.
Tracy (12:31):
Oh, that's really interesting. So you can come up with a flavor, and then using digital technology you can basically simulate what the production would look like using your existing system or maybe figuring out if you need to put anything new in place — although it sounds like for the most part they don't.
Ryan (12:48):
Yeah, I think the real benefit of being able to put something new in place is, say you wanted to change the bag sizes. And you wanted to look at, maybe we have a line that goes through normally, but now it feeds into a 1-ounce, 2 1/2-ounce, and 6-ounce bag potentially, with quicker change over time so that you can have more variety running through a line. You could basically drag and drop the designs from an OEM into your line. And it runs sort of, I don’t know if you guys are familiar with the Factorio game? It's a big game for engineers.
Tracy (13:18):
No, that sounds amazing. Factorio.
Ryan (13:21):
It's really good. It's a computer game that lets you build a simulated factory with physics and everything and eventually your factory covers the whole planet. It's still a video game, but it's really interesting that way.
And we're doing that in real life now, where we can drag and drop different elements into the factory and potentially move pieces from other production lines to different lines for what we're trying to do. And that really speeds up the R&D process.
Joe (14:00):
So let's talk about the increased throughput or the speed with which changeovers can happen, etc. So let's imagine it's 2004 and someone at Dorito says ‘You know what I think that the consumer would like? A shrimp-flavored Dorito.’ And maybe they've done some testing and some surveys and they’re like ‘There seems to be some appetite for a shrimp-flavored Dorito,’ etc. How difficult is that changeover in 2004 to get the line to pause and go from regular Doritos to shrimp-flavored Doritos versus the difficulty and ease of that technology or that changeover in 2024?
Ryan (14:35):
In 2004, you're dealing with paper in that part of the factory. So you're getting, if you're an operator, a recipe on paper to go grab a 20-pound bag of salt, go grab a 40-pound bag of this, go do this. And there's a lot more checking, rechecking mistakes, and quality issues.
There isn't the digital feedback that you have now with the sensors down the line to check for quality. So now we have 3D vision systems that can dimensionally check a chip. Is it big enough? Is it unbroken? Check it for color. Does it look like it got enough seasoning? Did it overcook? Is it a little too brown? Did we send it through the fryer too long? Is it undercooked? Are we getting a bunch of chips that are stuck together? There's some issue on the front end that things aren't working right after we made the changes.
So all that can be detected in [the] process now and things can be stopped a lot faster and corrected. Where, 20 years ago, you were pulling something off the line, someone was standing there analyzing it all with the lines running and you could be producing 30 minutes of bad product before you catch it, right? The potential for mistakes was a lot higher and the risk of lost revenue was significantly higher when you were talking changeovers.
Joe (15:44):
Yeah, I wanted to, can you talk then a little bit about the sort of business side of these decisions? So if you could catch mistakes faster, if you can make sure that the seasoning is applied consistently quicker on a new run, what opportunities does that open up for the OEM? Or the company in terms of like, okay, now the bar to make a decision is easier, we can try new things. Talk about the connection between that technology and then the confidence that a company has in trying new things.
Ryan (16:14):
Yeah, so when we're talking about how an operations person or a plant manager would look at their production facility, there's sort of a triangle between throughput, uptime, and quality. And they're weighing these three factors against each other for how they're producing to ultimately get the number...
Joe (16:30):
Is that a trilemma?
Ryan (16:32)
It’s a trilemma, right? Because you can never be better than your worst metric really. So no matter how much you produce, if only 50% is a good product, you've only 50% effective.
Tracy (16:42):
I’m saving that for my next performance review.
Ryan (16:46):
But it's like that in production. So what we're looking at really is the loss of uptime guaranteed when you are making changeovers. So you know, you're going to have planned downtime as you do these changeovers. So you have to outperform on your quality and throughput side as this goes.
So part of the digital transformation that you see in factories and the digitization of factories is that you can pull a lot more elements out of all of your various devices. In 2004, to use that example, if something failed in one machine center in the factory, if you were having an undercooking problem, you might not be able to correct it unless someone got into the program, made some changes, and then adjusted the timer, an operator caught it. And if you were 2 a.m. and it's the third shift, that operator might not be paying attention. It might be their second job. They'd just have probably not worked there that long usually. And the programmers also work during the day, so they might not be able to fix it until the daytime.
And now because of the digital feedback, we can create a sort of machine-learning environment potentially, or at least a closed-loop environment where we can use the quality data from 3D vision cameras later on in the system or any camera system we're using. And feed that back to the earlier machine centers to say ‘We're undercooking, go longer, make that change and I'll validate as this next batch of product comes through,’ right? Because it's usually a continuous process. So as I start seeing good chips, I'll tell you that we're fine.
And it operates a little like cruise control in your car, right? So you've got a loop system, you want to go 55, you're going 57, you slow down, you speed up until you get to a more ideal product, which in the past a really good operator does that on their own. But truly post-pandemic, we're at a loss of good operators and we're at a loss of people that want to be operators in the first place.
Tracy (18:33):
As you're going through all these new technologies, I'm sort of like mouthing to Joe going ‘wow’ for each of these. I'm coming ‘round to the golden age of snack food ideas, at least in terms of the actual manufacturing process.
But one thing I wanted to ask is, okay, so you described how new technology can kind of be used as a safety net almost for risk-taking and punting on new types of flavors and makes it easier in many ways to create these things. Is there a way that technology can also inform the flavors being made?
For instance, would you get data from a digital assistant on the factory floor which would say, ‘Okay, you're not using up all the, I don't know, like the cheese or the garlic when you're doing this particular product?’ Maybe you want to look into doing a more cheese and garlic-heavy version of this. Do you get that kind of feedback too?
Ryan (19:31):
Yeah, that kind of feedback can absolutely flow upward toward the business process as we're collecting all this data. So that tends to get visualized in dashboards for business operations to look at and see what they're doing. That wouldn't be an automatic decision, but absolutely, we could see that we might be able to produce different products because we have a glut of some supply with how something is working or how that goes.
I think what informs production more ends up being what leaves through the supply chain on the logistics side, what goes out of the warehouse, what starts being ordered more, and that now can be a driver, versus, I think in the past, you would have kind of a monthly production run planned out in advance. And now I think it's much more day-to-day — someone comes in and finds out what they're producing that day.
Joe (20:14):
So we've been talking largely in the chip idiom so far, so to speak. But another snack category that has really blown up, and I know that you've worked in this, is nuts. And particularly almonds and just the numerous flavors of packaged almonds you could get. I think it used to be you could get raw almonds maybe in a bag and maybe toasted almonds. And if you went to, you know, a hippie store, you could maybe get tamari or soy sauce almonds or something like that. But now there are just so many, they've just really gone wild with the almond flavors. What was the thing you were looking at Tracy?
Tracy (20:53):
Oh, there's a brand of Korean almonds. I think it's called Tom’s or something like that. And I am slightly obsessed with them. They come in all sorts of flavors like tiramisu, and honey, and there are savory spicy ones as well. Whenever I see them at any store, mostly Asian supermarkets, I buy at least five packs. They're so good.
Joe (21:15):
So talk to us [about] what you've seen in the nut space, the almond space.
Ryan (21:19):
Yeah, California [is] obviously well known for almonds as everyone heard about during the drought. So out here in the Central Valley, we produce a lot of almonds. And the production factories that handle nut production out here, not just almonds but pistachios as well when you talk about like The Wonderful Company. But when we're talking about almonds, it's typically you're going to hear Blue Diamond Growers is a co-op of growers that own several factories through[out] the area.
And typically what they have done is just produce almonds and send them off, large amounts to Asia, large amounts to everywhere else in the world because something like 99% of almonds are grown in California. So if you've seen almonds somewhere, they were probably grown by a few of these large farms and sent off.
And in 2020 during the pandemic, obviously Asian markets dried up as China shut down. A lot of the service industry that they might have sold to slowed down and they made a business transition into snacks, really out of necessity at first. And then it took off because people like yogurt-covered almonds and they like lightly-salted smoked almonds. They like, you know, spicy almonds. I think there's a whole wall at the airport now if you go in to see all of these flavors. Totally.
Joe (22:28):
Totally. It's wild, absolutely.
Tracy (22:30):
It was a massive mistake to record this episode right at lunchtime. Can I just say that?
Joe (22:35):
So on Amazon, you can buy a pack that has Smokehouse, this is just the Blue Diamond variety pack. Smokehouse, Bold, which is wasabi and soy sauce. Habanero Barbecue, Sriracha, Sweet Thai Chili, Salt and Vinegar. Blueberry, that sounds a little weird, I'm not going to lie. Whole Natural, Lightly Salted, and then Toasted Coconut — maybe. So how did this happen? How was a company that specialized just in like ‘We’re going to sell bulk almonds,’ then able to whip out like 11 different flavors like this?
Ryan (23:04):
They made the production choice to build new facilities and new lines solely for snack production. During this time the almonds don't stop growing, you have to harvest them. So when you're dealing with growers, they have the product. So they were essentially able to vertically integrate themselves into becoming a snack producer. And once you've decided you're going to make these little snack packs and you can make a 1-ounce or a 2-ounce or 6-ounce bag of almonds, you have your various lines and they're all producing different sizes.
And then all you're doing is roasting your almonds and flavoring them and then cleaning that out and changing it over and doing a different flavor. So you could run lightly salted and then you could run regularly salted without a real issue there, right? And then you could run maybe one more flavor after that before having to mess with things before you damage the flavor, right? So you can plan that out.
Joe (23:53):
So, you mentioned of course you have to clean out the flavor because you don't want to do the blueberry run and then the Sriracha run right after and there's still residue from the blueberry run. Has that process changed? The sort of switchover process of getting one flavor in and one flavor out since you started in the industry?
Ryan (24:14):
Absolutely, it has. Depending on what your product is and what your pipes look like, you're sending steam through, doing essentially what's called a Clean-In-Place process. So you're sending your cleaning fluid or steam through the pipe, and then you're changing out your nozzles and making sure that clean equipment is in there. Everything is stainless steel, it's food grade, it might get sprayed down, washed down.
But what we've been able to do now is detect cleanliness within the pipe that instead of running on timing, we can say ‘We know the pipe is clean when we see these certain parameters go through, or we can understand that this sensor as it comes through it's now clean.’ And that's allowed gains of, you know, if it was a 30-minute process, maybe you'd be able to gain four or five minutes.
But if you want to do that three or four times a day, you've gained a lot of production because you're doing maybe 200 bags a minute, right? So when you math that out, you're able to take these small, we sort of describe it as the power of gaining 1%. You gain 1% in 12 different places, you've gained the ability to produce a whole other product for an hour, you know, potentially.
Tracy (25:39):
So, Joe brought up the Korean almonds. And of course, we started this episode mentioning Kit Kats, among other things. But it does seem like Asia, in some respects, has had a headstart on experimenting with lots of different types of flavors. Can you maybe talk about what enabled them to do that? You know, places like Japan and Korea, how were they manufacturing all these different varieties? And is it a case that we're sort of catching up to that technology or is it a case that the market was just more amenable to experimenting with these types of varieties?
Ryan (26:18):
I'd say it's a mix of both. I think the factories when you're looking at production in Asia, tend to be newer. Where[as] the production in America is retrofitted [to] older factories from some of these legacy companies that a lot of these factories have changed hands. There's an ice cream plant down here that I think has changed hands like six different times and makes Häagen-Dazs bars. But it's been owned by a bunch of different companies and it's been around for 50 years.
And you didn't see that in the Asian markets. And I think that when they came in there with American flavors to produce for Asian markets, they didn't sell in the same way that they wanted them to. So they started experimenting locally by necessity to try and move product. I always think of it similarly to Proctor and Gamble making spicier toothpaste for India, right?
So they adjusted to the local flavor and I think some of that is market research and some of that's just sales data. They didn't perform, so they had other flavors. And I think it took a long time for them to understand that America doesn't just want to eat barbecue chips, right? That we'd love to get — I remember being very excited when I got my first Lay's All Dressed bag of chips that I found in the store because I had seen online that they were big in Canada. And I don't know why it was exciting that I had the very…
Tracy (27:30):
That's legitimately exciting. I'm excited for you.
Ryan (27:34):
And they were good. I mean, I never bought them again but they were good. And I think a lot of people have that experience where they want a snack variety, but they want it to taste good. And I think the fear is that variety will sell, but the flavor won't get another sale. And I think we've proven that sort of wrong, that we can always just create more flavors.
Tracy (27:52):
Right. This is exactly what I was going to ask next. So you mentioned you never bought that particular thing ever again. It does seem like there is a consumption pattern here where a lot of it is all about the novelty value. So you know, a new flavor comes out, you try it, you either like it or you don't. But even if you like it, how much repeat business are these types of experimental flavors actually getting? It feels like it's very much about the cycle of launching new products.
Ryan (28:23):
Yeah, it's definitely about the cycle of launching new products. And we've seen over the years that certain types of flavors have tended to hold up. Like chili lime flavors on a lot of different chips has shown up and stayed. Flaming hot varieties of every chip have stayed. So it seems with spicier flavors, they tend to stay.
A lot of the novelty, dessert flavors I don't think do very well, they kind of bounce in and out. You know, cinnamon churro chips sound great and then, they're fine. So I think if you trend it, if you walked into your store and took a picture of your chip aisle every week for a year, you'd see a lot of different flavors and you'd see some new flavors never leave. And those kind of gets added into the production rotation.
And that's all built into the business analytics from these companies that they can see they're selling, they move off the warehouse shelves, you know, they have to move these chips within a certain amount of time for freshness reasons. If they're not leaving the shelves, they don't get made again.
Joe (29:17):
This might be a really stupid question. But obviously, in the era of Instagram and DTC companies, you know, we know about Frito-Lay and these big companies. But there’s obviously a million smaller brands that maybe try to pretend to be a little bit healthier or whatever it is. Is it easy now? Like, if Tracy and I had an idea for a chip shape and a chip flavor, is there sort of like off-the-rack infrastructure and capacity for us to just sort of buy some time and space on a line and sort of, without having to do any infrastructure investment ourselves, have a chip brand?
Ryan (29:55):
Yeah. So, we call these companies co-packers, the same way you would use for beverage producers. And there was a fun example I ran into the other day of a company like this deciding they wanted to make a pea protein mix of a Cheeto. The company had this idea, this manufacturer built it and lo and behold, now you can buy PeaTos at Costco. So they sold this to Costco, had this idea, found this company to make it, and the production facility was there for them. So it's way easier than you think.
Tracy (30:27):
I'm sorry, I'm losing it a little bit at PeaTos. Sorry.
Ryan (30:31):
I know, I wanted to bring it up because I love the name. I thought it was so clever.
Tracy (30:37):
That's a serious question. Okay, so you walked us through all the technological enhancements that have enabled us to develop these new types of flavors and manufacture them in a more efficient and faster way. You know, things like sensors detecting cleanliness, the ability to observe the chips that you are producing for quality in a sort of automated way. What's next? What's really exciting in terms of that technology?
Ryan (31:21):
I think the thing that's moving the technology along now is the movement of robotics into the production process in as many, many places as we can get it. So where I'm starting to see it is, so you would now see robotic arms in most manufacturing facilities at the end doing the palletizing process. So taking the finished boxes, it's a big robot arm, picks them up, puts them on a pallet, arranges the pallet, and then sends it off to the forklifts or automated guided vehicle back to the warehouse, right?
And then we're starting to move where now we're even using robots to start picking packs of things or stacking different parts of the process to get them, you know, stacked 10 rice cakes and get them in a line so they can get wrapped and moved to the process. So when you start to think of the technological innovation in a production line, you want to work backwards. So you can start at the warehouse because you'll never have a problem where you've created a bottleneck because you're already at the end.
And then you can move to the palletizing process and the wrapping process because that's also the end. Now we're starting to move further and further through the production line. And we're having robots that move alarmingly fast as you stand next to them, that are doing a lot of these processes that maybe weren't done by hand, but were done with a series of blocks and conveyors where a line would make the product turn three times and then it would get it into the right position to get into a bag and get everything moving.
And the speed that we're creating on the backend has driven a lot of the speed demand on the front end now because we've got more production capabilities and we're ready to move. And when I think about the future, it's just going to be more of this variety.
I think that the food manufacturers have looked at things like Ozempic as a challenge and the way that they're going to tackle it is through novelty, right? Nestle just came out and talked about how we are going to innovate, and I don't know how Nestle can innovate except by making more products.
Joe (33:18):
So Ozempic is going to kill our demand for snack food, but the way they're going to respond is by just such tantalizing, incredible creative flavors that it will overwhelm the medicinal suppression of our appetites. I'm excited for this cat-and-mouse battle of whether they could just sheer innovate their way back to consumer demand.
Ryan (33:37):
It's worked for me so far.
Tracy (33:41):
I'm just thinking this is such a thematic Odd Lots episode because we talked about pallets, chips, although a different kind of chip. And then we're also talking about Ozempic and automation processes.
But Ryan, you alluded to this earlier, what is the labor picture like right now for snack food manufacturing? And one of the reasons I ask is, I don't think my husband will mind me mentioning this, but one of his afterschool jobs in the UK was packaging chips. And he said it was the most boring and tedious job that he's ever done. And I think he did that for quite a while and just spent hours and hours putting bags of chips together. But what is the labor picture like right now? How easy is it to find workers for this type of job?
Ryan (34:31):
It's very difficult to find workers that want to staff these facilities. It's a boring job and it's a stressful job sometimes. You know, things are going down, you're getting hollered at, you're holding up production depending on what you're doing. And you may not know enough about the process, especially now, where the machines have gotten so difficult. It's not a push button anymore, it's an iPad or it's a touch screen that you have on the line that is controlling things.
And you might need to click through several screens and make adjustments or start or stop something to cycle things to get everything back running. To go back to our 2004 example, what you might have seen in the past is a staff at every single machine center. So you would have someone at the fryer, you would have someone at the bagger, at the palletizing station, someone might be putting chips in a box, like you mentioned Tracy.
And now what we're seeing is we can probably have an operator controlling two to three lines potentially on an iPad, walking around, making sure that nothing's breaking, looking at and acknowledging alarms on their system.
And when we talk about things like robotics in the palletizing station, typically you're going to use two to three people per shift to do palletizing if you don't have a robot there. Someone stacking boxes or you've got a machine and what you're doing. So one robotic arm eliminates somewhere between six and nine jobs for a three-shift factory per line. So these products and these processes are removing operators, but out of necessity really. These jobs are very difficult to hire for and are always at a staff shortage. Every factory I walk into, has a hiring sign ‘We can interview today.’ You know?
Joe (36:09):
So, speaking of robotics actually, you know, actually there was an interview with Jensen Huang, the CEO of Nvidia, about this and I've [heard that] the power of robots will really be amplified with AI. And I imagine the idea is, right, that if you have some sort of robotic arm picking up a bag, it helps if the bag is maybe at a 45-degree angle and it has to shift a little differently or something like that.
Can you talk a little bit about that more? How AI or maybe machine learning sort of makes it so that robots are becoming more efficient? Is that something you see so that they can respond to deviations? Or a weird chip shape comes through and it's still able to pick it up? Can you talk a little bit about this sort of intersection of AI and robotics and whether you see that as a force amplifier?
Ryan (37:04):
Machine learning certainly, when we're talking about it, is already being used. And you'd be surprised how often a very slight tilt in something will disrupt a robot entirely. So that does make sense, it's looking for a very specific shape. Sometimes you have to really build the process out in front of it for a long lead to make sure you get things...
Joe (37:21):
This is what I heard from someone recently about why we still don't have, for example, like robotic pizzerias. Which is just that they’re still at the point where if one of the pies is a little bit smaller or a little bit bigger, etc., like the robots still aren't nimble enough or agile enough that they would be able to respond, and then that would break the whole process. And so the argument is, well, if you have some sort of AI sensor vision and stuff like that, then in theory it can respond to these deviations faster.
Ryan (37:51):
Yeah. Right now, where we are at with robots is that you'll see when they start to make deviations, sometimes they can respond for the first few issues. But they have trouble coming back to a home position to reset when the product is good again. So they can sometimes function for up to an hour in a sort of failing state and then they fail entirely. And then the process has to be they have to be put back onto where they are in the grid and relearn where they are and then you can fire it back up.
So with robotics, we still have a lot of limitations there and I think those limitations don't have to exist. But for the mass-produced robots that we're using right now in production facilities, we still have those limitations. So I don't think Jensen is wrong that we could start using AI in these processes. And I think robotics is probably the first place to look.
When we talk about machine learning, we're using that all the time where we're a closed loop process that the machine can determine. Dairies are especially good applications for it — cheese making, where you're heating the product and moving it through. You're able to adjust times depending on variations of the product and you can visualize what an ideal product would look like. And know that and make adjustments with all the different sensing units you have. You're like ‘I need this moisture content, I need this temperature, I need it to get to at least this, it needs to be this color, if it's not this continue.’ All that's happening, but we are not to a point where we're letting anything sort of make decisions on its own.
Joe (39:19):
By the way, there are a bunch more Blue Diamond almonds than I said before. So anyway, there’s various chocolate-covered ones too. Alright, I have one last question. One thing we haven't talked about is the distribution side because I have to imagine that it's great if the manufacturer can quickly change over their lines and not have to worry about downtime as much when they're shifting from one stable process to another.
But then you have this explosion of SKUs and I have to imagine that makes a warehouse, you know, before you might have had a warehouse that had two kinds of Fritos and now you have a warehouse of probably, you know, a hundred kinds. What is required on the sort of distribution side to handle the explosion of variety of snack food?
Ryan (40:01):
So we're seeing the rise of fully automated warehousing. And essentially what a fully automated warehouse looks like when you're dealing with pallet-sized products is you're going to have a series of conveyors and you're going to build it up in a multi-story elevator situation. And you'll have carts that are not operating dissimilarly to what you would have in an Amazon warehouse, where they're picking and placing something, they're taking small totes, right? And we're dealing with pallets.
So we're almost running it like trains and elevators moving to certain locations that are then visible on a 3D grid for the warehouse operator. So typically, we can have one or two people operating this warehouse now and calling down for the packages they need in an automated way. So saying ‘We're going to fill this truck with these pallets in this order so that when we get to the various stops it's going to make, it unloads in the proper order.’
All these different pallets are located in bin 114 on floor five of our system and pulling through. Where in the past, right now at a sort of a traditional warehouse, you'll have a sticker on the outside of a pallet and someone drives up with a forklift and scans it and then pulls it in based on what order they're pulling something from. And it's a lot more hectic to deal with multiple SKUs. And you see a lot of forklift drivers moving pallets to get to a pallet behind it, because of something that happened, where you don't see that in an automated situation.
Joe (41:25):
Tracy, are you convinced now that we are indeed in the golden age of snack food?
Tracy (41:30):
In terms of technology? Yes. Yes! And I am very intrigued by the idea that maybe in an era where people are eating less if Ozempic and the GLP-1 drugs become even more popular, that food companies are going to have to offset that loss of consumption with creating new and tempting varieties of things to try.
However, I will say not all of the flavors are evenly distributed. And you and I were experimenting with this earlier, trying to order particularly exotic varieties of stuff and you can't find them everywhere. You can get most of the things off Amazon, but they're not in every grocery store. And also prices are still pretty high. Like, I don't know if you've bought Lay's or something recently, but it's kind of insane how much it costs nowadays.
Joe (42:20):
By the way, there are Elote Mexican Street corn-flavored almonds that you can buy from Blue Diamond. I'm going to buy them. Ryan Harlan, thank you so much for coming on Odd Lots. That was an absolutely fascinating conversation and a great explainer of how technological innovation in the backend truly gives us this wonderful world of variety. So thank you so much. That was a lot of fun.
Ryan (42:40):
Thank you for having me.
Tracy (42:41):
Thanks, Ryan. I've never been more hungry.
Joe (42:45):
Same.
Tracy (42:57):
Joe, you know, you mentioned the Mexican Street Corn flavored almonds. So this reminds me, I was in the supermarket the other day and I saw there were Mexican Street Corn Elote flavored Cheetos and I was tempted. But I didn't buy them because I think it was like eight or nine bucks for a package and I just couldn't justify the novelty value.
Joe (43:21):
Just like maybe a one-off though. Tracy, by the way, I did order some Kit Kats and Doritos before this and I'll bring them into the office.
Tracy (43:30):
You'll subsidize my chip consumption?
Joe (43:33):
No. I'm saying now you get those and...
Tracy (43:35):
Just like a good industrial policy should!
Joe (43:37):
And return the favor of my buying...
Tracy (43:40):
Actually, I will bring you some of those Korean almonds. Because now that we've been talking about them, I really want to go get some.
Joe (43:46):
But just this idea. So there were so many fascinating things in the conversation. For one thing, I just loved how clear Ryan was about explaining how all these processes work. So that was really cool. But then you think, okay, you have this factory and you try to get some sort of maximum like 99% uptime, or maximum throughput. You can understand once you've achieved a certain level of throughput why you'd be really hesitant under the old regime, 2004, to try something new.
Tracy (44:16):
Yeah. This is what I find really fascinating. So the idea that, in this one respect, in the world of snack foods, feels like technology has enabled risk-taking. And so often we kind of see it go the other way, like people perfect stuff and then they kind of double down on efficiencies and then they get sort of hardwired into doing one thing perfectly. It's really interesting to me that in this world, we're using technology to enable variety in a more efficient and maybe interesting way.
Joe (44:49):
Totally. And I love the simple example that you don't think about, like okay, let's say you have to clean the pipes, which of course you do between different flavor runs. Before you’re like ‘Well, the solution and making sure you have clean pipes is to just let the washing process go along.’ So, maybe 30 minutes. But what if you could do it in 20 minutes and those extra 10 minutes are a waste? And then it only seems like 10 minutes, but what was it? The rule of 1%? If you could find 12 1% efficiency gains, then you can get an extra hour of throughput through [the] factory. And if you're doing hundreds of pallets or whatever per hour…
Tracy (45:20):
Right, it's exponential.
Joe (45:21):
Yeah. So, it's pretty cool.
Tracy (45:22):
But also, you know Ryan mentioned that game, Factorio. And the idea that, okay, it's a video game, it kind of — I've never played it — but it kind of looks like SimCity but for factories, and you create your own simulation of a manufacturing process. The idea that that's actually happening in real life, so you can come up with an idea and then you can use software to basically figure out how you would produce this with your current resources in the cheapest and fastest way, presumably.
That seems really interesting too. And a big difference to how things were previously manufactured, where you would come up with an idea, you'd have to do all this research and development, lots of customer testing; but you'd also have to talk to the factory floor and figure out how you would make these new things.
Joe (46:12):
Totally. I want to see one, first of all, I want to play Factorio or at least check out the trailer...
Tracy (46:16):
It looks good.
Joe (46:16):
It looks really fun. And then also this idea of, digital twin, I think was the term he used? So you could just create an entire digital version of the factory and if the simulation is good enough, then you just run the whole test in the digital twin. That's a really cool idea.
Also, this idea that, the point at the end that robots still aren't very good [and] that eventually, it becomes a problem if too many of the bags are tilted and then they have to be manually reset. Seems like there's potentially low-hanging fruit there if, I don’t know, AI can make it so that the robot can just automatically do whatever. It's really exciting. We should have Ryan back in 2034 to talk about the state of snack food manufacturing.
Tracy (46:59):
I'd be into that. And then maybe at that point, we really will be in the golden age of snack foods — of cheap and abundant and interesting things to eat.
Joe (47:09):
Can't wait.
Tracy (47:10):
Alright. Shall we leave it there?
Joe (47:11):
Let's leave it there.
You can follow Ryan at
@ryanharlan
.