Fix the Process: Real Cannabis Manufacturing Tools & Optimization Techniques ft.: Murphy Murri

Listen Now

Bryan Fields: What's up guys? Welcome back to another episode of The Dime. I'm Bryan Fields, and with me as always is Kellan Finney, and this week we've got a very special guest, Murphy Murri. Murphy, thanks for taking the time. How are you doing today? Murphy Murri: Doing pretty well, thanks for having me. Bryan Fields: Excited to have you here. Kellan, how are you doing? Kellan: He's doing really well. Really excited to talk to Murphy. Really excited that I got another Colorado person on the podcast, right? So yeah, we're going to dive into a bunch of topics today. How are you, Bryan? Bryan Fields: Yeah, I'm stoked. We're going to get this out of the way quickly because I've got nothing on Murphy here. So Murphy, we've got a little East Coast-West Coast battle. If you want to put yourself on the map, let's do it, and then let's get into the hot topics. Murphy Murri: I mean, I grew up in Utah, so I've been no-coast my whole life. I'm really anchored to this mountain range that spans the continent, but I do like the East Coast mentality. I like the pace. Bryan Fields: Maybe we should call it a draw, Kellan. So, before we get into some of the fun stuff — and I really, really want to dive deep into cannabis manufacturing, because I don't think people understand the nuances of how complicated it is, and I think the way you've been dispelling that information and making it clear and visible for everyone is really important — before we get into that, can you give us a quick background on how you found your way into the cannabis space? Murphy Murri: Yeah, I've been in the cannabis space my entire adult life, basically. I tripped and fell into it by accident. I was 21 years old and medical dispensaries were opening in Colorado under the caregiver model. There wasn't even a licensing thing — it was just a lease and a lot of paperwork. We were walking around with stapled packets that had our Social Security numbers on them. That went on for a couple of years. I got in as sort of a fluke, early on, when I was young, which meant I had no assets and no reason to be risk averse. I just jumped in, and from there we quickly got licensing, and that was straight to vertical integration — which is a crash course for anyone in business because it's a rare burden. I quickly learned every other aspect of the business outside of where I started. I was running all of the retail, I was running and overseeing all of the extraction and especially the infused products, and I was trying to stay as hands-off as possible in the grow, but constantly having to set up, relocate, rebuild, rearrange, tear down, and clean grows. So I've been in this industry pretty much since it developed in Colorado, and Colorado was one of the first. I've tired myself out through every stage. I've had to learn a lot about public policy and government that I never hoped to have to learn. I know more about building codes than I ever would have predicted. But most of it has always been in service of coming back to the plant, and I think that's why I love being in the lab so much now. Having worked in every angle of the business, the lab is where you can see progress the fastest — both as an industry and on a micro scale of one batch at a time. It feels very satisfying to set out on a mission, achieve it, and then look back at the results and analyze them all in the span of a day. There's no other part of the cannabis business where you can do that, except in the lab. Maybe taxes, if you're fast in math. Bryan Fields: I love that you highlighted that, because that's also the part of the business that's most volatile — the most open for opportunities, influx, challenges, leaks, and inefficiencies. Just grading cannabis manufacturing as a whole right now, one to a hundred, where are we today in terms of what's considered best practice, in your opinion? Murphy Murri: I always love to say there's no such thing as best — there's just a lot of worst. With cannabis, we've struggled to define best. If you ask that question from the perspective of a regulator, best practice is going to be compliance-heavy, and that's how you get a metric ton of data with no useful subtext, no useful context, no useful action items. So along those lines, I think the industry has actually done a really great job of proving we can comply — setting up really complex infrastructures and figuring out how to navigate them. Every time a state onboards a new licensing program, the fact that it ever comes to completion is impressive, because there's a tremendous amount of effort that isn't really in service of the cannabis plant. Maybe there's a reason for it, but the reason isn't cannabis. So on that end, I'd say we do a great job overall — the amount of diversion you don't hear about is significant, and those are outliers. But if we look at best practices from the perspective of occupational health and safety, my goodness — show me a facility in America that can tolerate an OSHA inspection today. We're so far away. We're far away in terms of actually implementing safety systems, but also in recognizing the role, because so many people got into cannabis from these disparate, random backgrounds, it gets really difficult to know who's actually an expert. Someone who's great at growing is rarely great at taxes. Someone who's great at taxes knows nothing about sanitizing a drip irrigation system. Someone who knows about sanitizing a drip irrigation system doesn't necessarily know how to keep people safe around flammable compressed gases. There's not a ton of overlap when it comes to employee health and safety. In most businesses we don't even have HR to establish a baseline — for working overtime, for paychecks showing up on time, for not being harassed in the workplace. We're really far away from best practices for employee and occupational health. And then if we want to look at best practices in terms of the actual products we're producing, we've made up an entire category of consumer good without ever defining those best practices. So that can quickly become a West Coast-East Coast battle — a geographical battle. I think we're better with edibles than we've ever been, and I think we're probably worse with flower than we've ever been. Everything in between is very geographical — it depends who's doing it where — because we do learn something every time a new market opens up, but we never connect. It's still a la carte at every step. So best is so difficult for any business, because someone who's going to excel in one of those areas is lucky enough to have one or two people who are really well-educated and experienced in that category, with the power to execute it. And even if you have experienced people in each of those roles, if they don't have the authority to implement, if they don't have aligned priorities with the rest of the team, they can't be effective. In that way, we're all walking straight uphill. We are far away from each other on that. Bryan Fields: So when we talk about operational efficiencies and aligning KPIs — let's say for maximizing product throughput or efficiency — how do teams define that, and how do they understand on a regular basis whether they're hitting those metrics to be successful? Murphy Murri: The number one thing I see is that if I ask someone about quality-oriented data points, they'll return compliance-oriented data points back to me. People really are thinking about KPIs as they relate to compliance. They're not thinking about them from the perspective of their unique manufacturing capacity. They're rarely thinking of them from the perspective of their unique financial position in a market. They rarely have the tools to think about it from the unique scientific perspective of what's even possible within their given input material and potential output performance. So it's really tricky, and I think the one thing people measure most is just kind of blanketed everywhere else — it comes back to grams in, grams out. Straight-up compliance yields. The data people rely on most are these broad numbers that are outcomes — they're not actionable items, they're measured at the end, which is after I can do anything about it. It's an important data point, but it's supposed to be the climax of a much longer story, and I'm getting no prologue from any of these people's data sets. They're not giving themselves any context with which to apply it. So then you have people who will report back an average yield across their fresh-frozen trim and random remediation bud, and it's like — that's a number nobody needs. It's not useful. You can't do anything with it. Kellan: How much of this do you think is rooted in where the industry is from a maturity perspective? Because a lot of people got licenses and it was almost like, "Hey, I have permission to do this, and now I just need to do it — keep the lights on, pay my bills." And there wasn't any of the other assistance normal businesses get — banking, other kinds of capital from a federal perspective. So a lot of it was, "We need to make payroll next month," which creates a culture. So do you think what you're seeing is really the remnants of that culture, throughout a lot of these companies that are like, "Oh wow, we're still around ten years later — what do we do now?" Murphy Murri: I'd say yes, but I'd also bring up the fact that the traditional market was always good at this. Traditional market operators had to, at all times, balance the quantity they were producing against the value the quality of that quantity would return for them. There was an understood expectation of quality in every traditional-market deal. And because there wasn't a number to attach it to, they had to communicate that. When I think about the traditional market, there was so much transparency, because you couldn't collect on a transaction unless you knew what other people were doing, how much was moving, and whether it looked like it was supposed to or not. Everyone relied a bit more on the honor system when it came to reporting. Back in the day, you were motivated to do a good job trimming because you got paid by the pound for it. Your incentives aligned with the outcome that whoever was paying you decided you were worth paying for. To have that alignment in the traditional market evaporate under the pace at which people had to grow their small businesses is really tough, because what it illustrates is administrative management — used to banking resources — colliding with the slow growth that comes with that traditional-operator experience. They inherently always knew they had to run at a profit because they needed re-up money. They didn't have a bank account, but they had re-up money. And what we see in today's world is people without a bank account or re-up money trying to act like that's any way to get by. None of our collective experience supports that. But when we all get together, what we have to recognize is the compromises we make. When you have to earn re-up money just to stay afloat, you don't get to grow fast. It's really the speed of that expansion where people trip over big obstacles — they forget one category, they measure for square feet instead of cubic feet, and here we are, way, way off. Bryan Fields: You mentioned the outcome-based data point, but that you're not getting any of the earlier story. If I gave you a magic wand and said, "Murphy, tell me all the data points you'd want for a complete picture," what would those be? Murphy Murri: Moisture content, starting there, absolutely. Moisture content during cultivation, moisture content at the time of harvest, moisture content at every point of repackaging — so I can see not just what the status is currently, but what has happened to it historically. I think moisture tells a really important story about the storage of our biomass over time, and it's one of the larger numbers in the mass of that biomass, which matters because I can't accurately measure terpene content, I can't accurately measure low levels of heavy metals, low levels of BTEX and other harmful aromatics. I can't really measure a lot of contaminants at the flower phase. But if I can measure water the entire time, then when I find contaminants at the end, I can go back and figure out where the concentration occurred. I think this is a really important angle we could take from both sides — as labs, to measure what we're doing and what's possible for us, but also to rebuild trust with our cultivators. Because I do believe no one wants to be doing a bad job. Most growers want to grow good weed. The days when growers bragged about being a grower have shifted, because a lot of people working in grows aren't doing things they're super excited about, and part of that is because they can't prove the value. If you work harder to grow more aromatic weed and we're going to over-dry it and store it in bins and trash bags and turn it into distillate, you just don't get any wins there. But if you could measure it and show it as weight, that would go pretty far in the meeting. So it's really hard to measure the terpene content you add, but it's not hard to measure the moisture, and moisture tells a really important story. So I would start there. As it moves on, water tells an even more important story throughout extraction, packaging, and formulation. Water is going to tell us a lot about the environment it's being processed in. Water is going to indicate whether our solvents are still clean. Water is going to indicate shelf stability in the final products. So moisture is lacking on everybody's pie chart, and it's one of the easiest things for us to test. I think that context would be huge and could be used at every single level of manufacturing, so we have something we can all talk about in a tangible way. A lot of times I'm talking to extractors about how to control the moisture content of their flower before they start extraction. But I promise you, if I talk to the growers who produced that flower, they'd tell you they'd love another dehumidifier in the room — that's where you'd actually solve the problem. Telling extractors to freeze everything in an open container for 24 hours gets me through this batch, but I don't get to fix the actual problem unless I go back to that hot, sweaty flower room and figure out how to offload some of that excess moisture. So simple data points like that can provide a lot of context. And the problem is, moisture isn't regulated — it's not mandated by anybody. But it ties directly to every other data point we collect. It's just outside the compliance bubble. From there, I think there are a lot of others we could pay close attention to, depending on what's important to us. Part of moisture control is aromatic retention, part of it is preventing microbial growth, and if I have one of those problems versus the other, there are different data points I'd want to track. An easy one to knock off the list is container changes — every time you re-bag flower, you've made a mistake, because a lot of time gets wasted in repackaging cannabis, and that's another one that would be really valuable for people to look at. If I saw how many times a bag had been opened, that would tell me a lot. Right now, that's why we used turkey bags — so we could reach in and grab a pinch quick. That wasn't done in service of the cannabis. So looking at things like the containers required for compliance, and pairing that with quality, and saying, "I want to make sure this never gets opened again, so we always package in these specific quantities to avoid that" — that goes so much further, especially when you're vertically integrated. I think about all the people who have growers put weed in a bag, and then extractors take weed out of that bag to put it in another bag. That's a lot of bagging they could have done once. That alone is probably twenty hours of labor a week in savings, and I don't even want to think about how many trichomes end up on the floor. All of them — so many. So little things about the life cycle of that product would go a really long way. We talk so much about seed to sale, but nobody's using seeds and nobody cares about sales in manufacturing — we're in the middle, where we're actually making it, and we're missing that entire historical background, which is honestly insane. Kellan: I think a lot of it, too, is that people go to record this data and then nobody does anything with it — they're just like, "Oh, I have five years of moisture data." So providing insight on how to actually use the data is super valuable. I think the other thing is that moisture is a unique topic because it isn't compliance-related, so it's become this lever where, potency increases are probably directly correlated to moisture manipulation. And moisture is a general category of water, right — there's free water, there's water activity, there's a whole slew of how water molecules bind or don't bind, which is a whole subject in itself — it's really about how the water is interacting with the plant. So when you explain to processors, "Hey, if you just monitor this moisture content..." — is it a lightbulb moment for them? What's that conversation like? Murphy Murri: I live for a whiteboard, because I usually have to draw a picture — I'm a big fan of analogies. I try to figure out what's going to land for whoever I'm talking to. If I'm talking to an executive or business-owner-level person, I want to translate this data collection, which they see as a burden of extra work, into dollars. Sometimes that's a shortcut to time savings. Sometimes it's maximizing yield and not throwing away extras. Sometimes it's simply avoiding compliance failures — reducing overall risk. But it's usually converting the number into whatever outcome they actually want. When I'm talking to operational staff, they see my request as more work too, and what they want to know is whether that work will be seen by their managers — will it be valued and recognized, and how will it actually impact their outcomes? Because not only do they not want to do more work, they don't want to do more work for no reason. Executives don't tend to care about the reason why something works — they care about the outcome. So being able to say that controlling humidity in a grow room reduces dry time, so you don't have to build a bigger dry room — you can just dry a little faster because you're not evacuating ninety-eight percent water content in the dry room, you're only evacuating eighty percent because you've got much lower humidity in the grow room to begin with — that lands. Often what a grower will say is that the plants are unhealthy and unhappy in high heat and humidity, and nobody has sympathy for that — "the plants are uncomfortable" isn't a line item in a spreadsheet. Discomfort of plants isn't a category in QuickBooks. But if I can translate that into faster dry time, reduced microbial potential, staying under 0.5 water activity the entire time so you never fail a test — that's what it takes. In most of these cases, the grower already knows that controlling humidity will help. I don't have to convince them of that. In most cases, the extractor already knows that freezer-burned fresh-frozen runs poorly and causes problems. I don't have to convince them either. What I have to convince their upper management of is how that turns into worse numbers, because management often averages it out — people make enough good decisions to compensate for their bad ones, and they still end up right around where a functioning business should probably be. So it's always two pitches: how do I impact your outcome and skip the backstory for the people who don't need it, while making sure the backstory is covered for the people in the middle, so the people asking for data are asking the right questions and the people providing it understand how they're impacting it. A lot of times the experts on the floor aren't the experts in the room making the decisions, and while that's unfortunate, it's actually kind of normal. A good CEO doesn't necessarily have to know how every piece of equipment works — but he does have to know how to find out whether it's working at its optimal conditions. As long as the questions we're asking are aligned, and the numbers we're asking each other relate to each other, it works out. Not everybody needs to know everything, but everybody needs to know why what they're reporting is important. That delegation of responsibility is pretty rare — a lot of times you have a decision-maker who lacks context, or a context-owner who lacks any ability to make decisions, and it isn't effective. It doesn't do anything. Bryan Fields: Those conversation points about data are so important, right — exactly like we talked about: does the data get used, is it valued by the manager, is it actionable enough to make a difference going forward? And what seems to happen is someone like yourself comes in and says, "Listen, there are problems here, here are the fixes," and short-term, it's, "Cool, we'll start recording everything." The manager says, "We're doing everything way better." Then a month down the line, someone stepped out, people stopped looking at the data, and everything gets kind of lost. So do organizations need to become more data-centric — is it a mindset shift, or is it about understanding that early on, you need to collect enough of the right kind of data to make decisions, so you can build an organization, potentially with a tool, that's forward-looking and successful? Because what we've all seen is margin compression happening, numbers getting even tighter from a COGS standpoint. If you're not doing enough to see the full picture, you only get outcomes after the fact and you can't go back and fix it — new information just keeps coming. How do organizations get in front of that and continue to be successful moving forward? Murphy Murri: I think it's always easy to say we don't have enough data, and the way people try to solve that problem is where a lot of the opportunity — and a lot of the mistakes — lie. One way to get more data is to collect it from more places, which is what a lot of people try to do, and that's how you end up with the worst-quality paper towels I've ever seen and the most excessive solvent waste I've ever seen, in the same business. It's like — why are we saving five dollars on a box of paper towels when we're burning five hundred dollars a day in nitrogen for no reason? A lot of times, looking at more data doesn't make it easier to find the point of impact. In my opinion, one of the best ways to use more data is to dial in tighter on one piece of data that we know has impact, and we can do this one at a time. It may not be that you have to measure the moisture content of every bud in a flower room forever, but if you measured the moisture content of a bud on every plant in a flower room for a month, you'd create a data set you could tie to your CO2 usage, your environmental conditions, and the yield at the end of that harvest. If we dialed in enough, if we got enough data points close together, we could actually start to impact them. One of the biggest issues with our data is timing, and that's why compliance data is so ineffective — it's not about what's happening now, it's about what you're allowed to do and what you've already done. That's theoretical to the people in the lab in the moment; those batches are already over, or yet to come. Real-time data has real-time action. The thing about taking action is you can't do it in the past, and you can't really do it in the future either, because we have these huge gaps in our understanding of who we'll be and what conditions we'll have in the future. I always have more time tomorrow than today, in my mind, in spite of the same twenty-four hours existing endlessly, and that same bias gets applied within the business — we always assume our overall cost will go down because our overall sales will go up. We have these optimistic views of the future, and that's fine, but we can't react to that today. The action has to be as fast as the data comes back to us, and the speed of that data is very painful. When I look at analytical testing — testing for microbial activity in a gummy after I've made oil extract from flower and processed and infused it — that is just the worst time to find out I have microbes or heavy metals. There are a thousand places it could have happened. The cheaper option is always to test at every stage and never have to fail at the end, but the only way we seem to learn that is by failing at the end, finding out the real cost, and working backwards. A lot of times our action is, again, compliance-oriented — we don't take action until there's a compliance risk. I understand the importance of that, but it should be our last line of defense, enforced at the final gates — not the way we wage our battles in between. We've got to be approaching this data faster, more often, on a more micro scale. Kellan: I think it's also important to mention that with compliance data, it's a mindset — people go to do their compliance data entry and check out. It's, "25, into that cell — here we go, I got it done, now I can go home." There's not the thoughtfulness that needs to go into the data-management side when you're actually recording it. If you're recording numbers with the goal of optimizing a specific process in mind, you're recording those numbers with a completely different mindset, and that's where a lot of the insight comes from. So it's really hard to decouple — it's almost a completely different data-management system at that point. Murphy Murri: Right, it can be. I think compliance data is a different job than operational data, and to that end, they can be different people. I think compliance data should be automated, and probably AI, and it could be extrapolated from operational data. Kellan: Don't help operators too much — not in California. Bryan Fields: We've got handwritten logs, guys. We write things down on paper — it's fast! Murphy Murri: It's such a burden. Think of it in the context of a sprint, a race. If you're the runner, you do need to know how fast you finished, but more importantly, you need to know how fast everyone else finished, because knowing you won is great, but knowing you only won by a tenth of a second matters compared to winning by half a second, or a whole second. And if you didn't win, that gap is even more important. So the operator needs to know not just whether they met expectations, but how far away they are — because if they don't believe the expectations are reasonable, they're not going to pursue them, and if they don't trust that what's being asked can actually be done, they're not going to try as hard to achieve it. So you end up with self-defeating goals, where we say, "We want to average six percent returns on all our fresh-frozen." Great — not helpful. That pizza party is never coming. We have to give people a goal they can actually hit — increase the yield you had last quarter by 0.1 percent, or close the gap between last quarter's potential yield and your actual yield. This quarter, we want to give you a smaller, incremental goal so it can be achieved. The impact of that goal won't be small — the impact, in the second half of my pitch up the executive ladder, is that 0.1 percent is thousands of dollars, and it's thousands of dollars across ten different categories in your accounting spreadsheets. You're saving money everywhere it counts. So if we can get a directive from the top about which areas are the biggest burden, and a directive from the floor about which areas are the biggest burden, they often overlap — they just aren't called the same thing. It's almost a conversational problem. Part of it is that the minutiae of constant data collection doesn't need to be in the spreadsheet that goes up the executive chain — it needs to be in the spreadsheet the people interacting with that equipment every day actually see. I don't need my CFO reviewing the temperature log for my distillation system. I've spoken to many CFOs who suddenly get excited that they want to learn about distillation — it's not helpful to them, and it's not helpful to the staff who then have to teach this guy how molecular distillation works. What matters is for him to understand that we're hitting our goals in yield, in energy, in whatever's most important. What matters for the staff is knowing that when they track those data points, they can follow it through to those outcomes. A lot of times what you have is staff tracking data points and then someone else coming in to analyze them — that's insane. Of course they don't trust that guy, of course he doesn't have the right perspective and is going to miss the impact. What you need is the people who can impact the data to be interacting with the data. A lot of people only collect it — it's like the cleaning checklist in a public bathroom with twenty lines on it. Nobody's ever filled out twenty lines of a bathroom cleaning checklist. Nobody's actually wiping the mirror and checking it off, wiping the toilets and checking it off — that's never happened, and it doesn't make the bathroom cleaner to put that burden on somebody. Reporting it in a more functional way is way more important. You get much cleaner bathrooms when there's a button you can push that says, "This bathroom needs attention." Instead of reporting how often you clean it, you report how often it gets dirty. It's a big shift. Bryan Fields: Yeah, it's a mindset shift, and there's a few things I want to highlight there. You talked about real-time data and how valuable that is, and expanding on that — the integration of digital sensors inside these processing facilities allows you to skip those manual steps. Now the operators don't need to track twenty things by hand; everything gets transmitted automatically to a system, and all that data gets piped in. The most interesting part is that cannabis isn't the first industry to do chemical manufacturing, so we can borrow tools from other industries for real-time process control to implement best practices. We still see companies a little hesitant about that change, but I think it's exactly what you described — it's the mindset shift. My anticipation is that a few companies will start picking this up — we've already seen it start — and it'll become an avalanche of momentum, with everyone asking, "How did you accelerate so fast? How did you get your COGS down so low? How did you get so much visibility into your process? How did you know what to do with all your data?" And it'll turn out they borrowed tools from other industries to implement best practices here in cannabis. Murphy Murri: Right, that's why my motto is "more hash, faster." How do you get better at anything? You make more of it, faster, and you'll either learn more or get better at it faster — you're either going to change your technique or improve it. Creating robustness through repetition rather than wide expansion is exactly the position we should be in, especially because the only unique thing about cannabis is cannabis. So let's dial in specifically on cannabis and let everything else be very standard — we can borrow that from every other industry that's already done the work. Why wouldn't we? Everybody loves to think their data is proprietary, but not understanding something isn't the same as having a proprietary method. Sometimes people get so deep in the weeds about the complexity of their own issues that they can't step back and look at it, and that's a great opportunity for AI — you can type in all your data points and say, "Imagine these were the conditions of my direct competitor — what would I do to compete with them, or take advantage of the flaws in their business model?" Sometimes getting outside your own perspective is critical, and if you have the data, you can get that perspective easily. A lot of times, as a consultant coming in, my first goal is to teach people how to collect data, and only after they've collected some do I come back and teach them how to analyze it. If they're already proactively collecting it, every time I come in we can have a much more effective conversation about analysis, because every piece of advice I give is a step on a ladder that leads to the next one. Most of the time my advice is, "You have three potential causes of the issue, three to five ways you can approach it — pick one, tell me what you did and what happened," and the decision tree goes on from there. It gets very ineffective when we try to do that with ten different things at once. But if we dial in really tight on one step in the process at a time and get it perfect, the decision tree becomes reasonable — now it's yes-or-no instead of if-then. From a manufacturing perspective, we want a lot of yes-or-nos. We want that binary: is it ready to go, or not ready to go? Is it approved, is it not approved, is it rejected, is it in progress? We're pretty far away from that. We have extensive if-thens in cannabis, and most of them are by design — but accidental design, usually as a consequence of a different decision made in another department. Sometimes it's, "This is the building we could rent," so we do a lot of crazy things to deal with a short ceiling, or to deal with not having enough power. Somewhere along the way we made a compromise, and sometimes you can overcome those compromises and start choosing differently. Sometimes we have to start from a shaky foundation and support it in other ways, but it's all very doable. I think what a lot of businesses miss is that they overestimate their ability to identify their own weaknesses. We're all better at identifying other people's problems, and at determining what other people should do to solve them. We're rarely good at doing that for ourselves, and the same is true in these businesses. Just because you have managers in every room doesn't mean they have the insight, and even if they have the insight, do they have the authority to do anything with it? These are tough questions in this industry, because nobody does it the same way. The term "standard operating procedure" gets used to mean almost any piece of printed paper that people don't read. So we're starting from a very tricky, falling-down staircase as far as... Kellan: You're not wrong. Bryan Fields: But it's super dialed in, Murphy. We've got it really dialed in. Murphy Murri: The more binders you have, the more I know that nobody's ever read them. You know what I mean? Do they have color-coded tabs? Nobody's ever read that — it's hard to find. Kellan: Not even the person who spent four days making the color-coded tabs read it. But they did a job, and it looks good. Murphy Murri: Their job was binders, and they did it excellently. Bryan Fields: So then teams go through their process — specifically in extraction, right — they've got biomass variability, they run it through their system, they've got some sort of challenge there, they want to soak for a certain amount of time but can't, then they've got quality issues, so they do a second pass. Where in that process would you say to yourself, "This isn't going the way I want — maybe I need to reconsider our flow, because if I make this adjustment here, maybe I don't need a second pass, and I'll lose a little bit there, but I'll improve quality and cost"? The speed of the industry, moving as fast as possible, sometimes blends into the challenge of, "Okay, we've got to get another run up, we don't have time to go back and change it, Bob's not here, he's our best guy, we'll wait till he's back" — and now it's three days down the road and it's, "Alright, well, next time we get it greenlit, we'll run it." If you were in charge of a business, where would you be the one to say, "I don't think this is going right, the alarm bells are going off in my head — now is the time to really reconsider how to change this"? Murphy Murri: There are quite a few opportunities throughout the course of a business where people could address something right away, and I think one of the easiest ones people miss is that they don't put a lot of thought into their purchase agreements for large-scale equipment. So when the performance parameters, details, and data points that the equipment manufacturer promised them aren't met in real life, they have no legal recourse, and they're usually bullied into paying for more training, more consulting, more ongoing support from that company — or they just decide they don't like that company and try to save up to buy a system from someone else. They don't hold these manufacturers accountable for the performance of the system. That's something the executive team is built for, my goodness — get the binder guy, let's go, this is the time. Hold manufacturers accountable for what they promised you, and if they can't deliver, build that into your contract so you at least have some financial compensation to cover the gap. Because while I appreciate that so many of these manufacturers fail to meet expectations — it gives me a job — it would be even better if those manufacturers were paying me instead of the customers, who now have to pay two people twice, because by the end of my training they have a new list of things they need to buy and they keep running further out of money. So start by holding people accountable for the promises they've made. Equipment manufacturers should be held to a higher standard when it comes to promoting what their equipment can actually do. A lot of people get in over their heads on cost and ROI on day one, because the equipment was sold based on a set of operating conditions that will never be met. That's a huge place to shift from right away — don't even start making financial projections until the equipment is installed and validated in place under real operating parameters, and then once those exist, you have to revisit them periodically and set a depreciation schedule in effect for that operational efficiency for certain types of equipment. A lot of people compromise their maintenance in favor of extracting more today, because they figure they'll have time next month, next week, tomorrow to take on those extra tasks. Not only will you not have time later, you'll actually be less efficient every day as you fail to meet those goals — it's like not getting your oil changed, you literally use more gasoline. Everything gets less efficient. The compromise doesn't happen all at once, later or today — it happens every day, over time. I see this a lot with expansion: people bought equipment, they had a disappointment, they put someone in charge of the lab who couldn't do what they were hired to do, now they're in a deficit and they want me to come in and help. I always want to start by asking what was working before — what was working so well that led to those other decisions? Sometimes what worked well is they were great at cultivating, and then they said, "We didn't know that much about extraction, we hired the wrong person, we trusted the wrong company." I want to figure out where their background is — that usually gives me good insight into where the issues are. If they started in cultivation, that's excellent, because I know it's probably not a starting-material issue — it's more likely a handling and operational-support issue once they got into the lab. If they don't have a cultivation background, maybe it is a starting-material or input quality-control issue. Everybody's got different strengths and weaknesses, but even with a lot of diversity on staff, one or two people usually have the greatest impact on decisions, so that's who I look for — whoever sets the tone. Then I need to figure out how to convince that person to help everybody else see their vision, because we do tend to all want the same thing — we all want to make the best product at the end of the day. It's just a question of how we get there. Here's a literal example: someone could go from making really high-quality distillate in a short path — they dial in that short path, it's all glassware and different components, they dial it in tight, they know exactly what their incoming crude needs to look like, and if it's not winterized the way they want, they won't put it in the short path. They're very specific, because they know that when it's time to clean that short path, it's not going to be a good time if they lowered their standards on input material — and they can see it, because it's glass. They make so much of that distillate that they hit a limit, and the question becomes: do we go continuous, or keep doing sixteen-hour passes on the world's biggest short path? So they pivot technology, and instead of glass where they can see, they get a wiper where they can't see what's going on inside. Instead of thermocouples on the vapor in the distillation head and in the oil where it's being heated, they go to no thermocouples anywhere, just set points on heaters. They start trying to run the same distillation product on equipment with none of the same features. They can't run it visually anymore, they can't run it by temperature anymore, and suddenly their quality drops, because it runs continuously and needs constant feeding. They don't do as good a job winterizing anymore, they don't do as good a job making sure it's fully decarboxylated anymore, because since they can't see the eruption that would have happened in a short path when it hits the wiper, they're blind to it. All of that leads to continuing problems — not only did they make several bad decisions in a row about equipment and process, they're also still accumulating the same grime, so the lessons they'd already learned aren't getting fixed anymore. All of that is because they didn't collect the data points until the end. If you wait to find out the distillate is going to be low quality at the very end of the process, it's already made. But if you had real-time checks on the glass, the thermometers, your real-time vacuum data, you would have known and adjusted. The people running short paths before were interacting with it constantly, keeping it bubbling at the same speed so there weren't gaps where it was just getting hot and steamy. As soon as we switch to wiped film, you have operators who are literally just moving kegs around and connecting hoses — they got away from those tangible details, and when we took away their tangible, real-time data and feedback, we took away their connection to the output and to the impact they have on the process. You can do the same dialing-in with a wiped-film system that you do in a short path, but you have to look at the same data from different sources, and sometimes the equipment doesn't support that goal. So for me, it's always trying to figure out when and where the initial shift or deviation from the good plan occurred. Sometimes it's very early — bought the wrong stuff day one, and it's not getting better; hired the wrong person day one, and it's not getting better. But sometimes it happens as a result of success, which is where I think a lot of people are — they succeeded somewhere, and it allowed them to get in over their heads in another area. Bryan Fields: Last question for you — process optimization. How fast is the return on investment for most companies? Murphy Murri: It depends on where they're spending their money. If we assume the equipment costs are sort of depreciated and sunk, the return is pretty fast. But because of the banking issue, a lot of people are having to pay cash for equipment, and that's where we're very far from other industries — nobody else pays full price up front for their equipment. That's usually absurd. So because a lot of people pay cash, that's where the biggest gap comes from. If you overspend on equipment day one, if you have to buy multiple versions of that equipment to get it right, that can eat you up. But if we're talking about consumable costs, same day. Payroll costs, maybe a month. And if we're talking about downstream rework and avoiding failed tests, that will be as fast as your access to data — if it takes a week to get a turnaround on test results, you could do it in two weeks; if it takes a month to get enough testing data together to look back at your batches, it's going to take a month to turn around. It's really going to be about as fast as you can take the time to collect and review the data. That's where it gets granular, in terms of how you've set up your business — do you have people in charge of data who have context for it, are you prioritizing the right data? Those two jobs might mean adding another human, but also removing a lot of rework. You really have to balance the whole picture to see the true impact. Bryan Fields: So Murphy, give our listeners a quick pitch on the type of services you provide, and let them know where they can reach out and find you if they need your help in the future. Murphy Murri: Sure — I've got a website, MurphyMurri.com, easiest place to find me. I'm also on a bunch of social media, and I do a lot of podcasts to answer questions in longer form, so I'd recommend those to people with specific questions. The services I offer are general consulting services. A lot of times I'm hired in moments of transition for businesses — they've just purchased equipment, or are about to purchase equipment. To those people I'd always say: if you can ever choose, hire me before you buy something. Let's optimize what you have before you buy something new, as much as possible. I'm happy to come help after you've bought something too — it's just worse news to deliver, personally. If I can catch you while you still have a lot of potential, that's awesome. Same goes for hiring and training people — if you're unsure whether your staff are performing where they need to be, if you're losing trust in management, that's where I can provide a lot of support. That might be because your operational management doesn't trust the executives, or vice versa, and in both cases I can come in as a third party to more objectively explain both sides, because it's almost always both, rather than one. Occasionally there's a bad apple, but usually you already know that before I show up. So my services range from training on the equipment you have, to developing and designing a lab flow and workflow for new equipment, so you can optimize an existing process, add more products to the workflow, or control costs — especially when they're getting out of hand. Bryan Fields: I love it. We'll link it all in the show notes. Thanks for taking the time — this was a lot of fun. Murphy Murri: Definitely, appreciate it.