AI Won’t Save You If Your Process Is Broken

Episode Transcript

All right. I am coming in hot with a depressing stat alert. 95% of executives are saying they're struggling to see any ROI from their AI initiatives, and you wanna know where the solution to that problem is? It's in a book from 1984. This is the same tool that is gonna help us solve the we're-always-working-hard-but-nothing's-ever-getting-done problem. So yeah, it's pretty freaking powerful, and yet most people never talk about it, but we're gonna talk about it 'cause that's what we do on this podcast. We get beyond basic Agile tactics, and we explore the holistic operating system that's needed to embrace change like it's a fluffy bunny and not a stabby porcupine. That's right, my lovely listeners. Today, we are diving into the Theory of Constraints.

This is a powerful systems thinking concept that will help us pinpoint and then eliminate the real blockers in our workflow. So we're gonna dive deep into how to apply this theory in a modern marketing context. We're gonna figure out how to spot common constraints, then go into what to do once we find them, and really unlock the dangers of applying our effort in the wrong place in our workflow, and this is really cool stuff because once you remove the biggest bottleneck, that's when everything starts to flow.

So let's dig in. Welcome to The Agile Marketing Edge, the first podcast dedicated to turning Agile theory into real-world marketing breakthroughs. Every week, we unpack the how behind Agile. From building high velocity workflows and slashing waste to measuring what really matters and scaling success across teams, you'll hear quick hit strategies you can deploy today, plus candid stories from marketers who've traded chaos for clarity and never looked back. So lace up those virtual hiking boots, limit your WIP, and let's start ascending.

This is your weekly shot of practical, no-fluff Agile insight so you can deliver more value with less busy work and love your marketing again. And now here's your host, Andrea Fryrear. All right. So The Theory of Constraints was originally introduced by Dr. Eli Goldratt in his 1984 book, The Goal, and it originally comes from the world of manufacturing. But the core ideas from this book really apply to any complex system, and obviously, that includes modern marketing teams because modern marketing definitely qualifies as a complex system. Now I'm gonna hit the highlights here as we go through the five focusing steps and talk about how to apply them, but you really, really should read the book. I will drop a link to the Amazon page for the book in the show notes. It's really a quick read and really fun. It's written like a novel, so it's kind of fictionalized, and it's got characters. It's not a typical businessy book. You can get through it in a weekend.

Really fun read. Uh, highly recommend. So let's get into the nuts and bolts. The Theory of Constraints says that every system has at least one constraint at all times. You're never gonna get rid of all of them. There will always be at least one, and this is the thing that limits its ability to achieve more of its goal. Hence, the name of the book is The Goal, and if you don't identify and explicitly address that constraint, then the whole system is just gonna be kinda stuck.

It's never gonna get any better, and it's gonna be sort of spinning its wheels. And this part's important. If you apply your efforts to improving a part of the system that is not actually the constraint, you risk making the whole thing worse. Now this part is critical, and it's at the heart of why so many of those executives that we talked about at the top of this episode are having so much trouble realizing ROI on their AI investments. This is why many of us are struggling with our AI implementations 'cause we're applying it not at the constraint in the system. We're just slapping it in willy-nilly wherever we feel like it kind of fits maybe or wherever we found a tool or somebody got excited, and we put it there, and we're not thinking about where it might do the most good or where our constraint really is. And in marketing, constraints show up in all kinds of ways, and there's all kinds of places that AI might do us good. So we have to take time and think through the entire process. We can't just slap AI in any old place. We might have slow approval processes. That might be the primary constraint or the bottleneck. We might have a lack of a cohesive strategy. That might be the bottleneck. We might have overburdened creatives or unclear priorities. The list could go on and on. I could spend the entire rest of this episode listing where your constraint might be, but we have to find it first, and the key thing to take away right. here is? that working- harder...... somewhere where your constraint isn't is not going to help.

You have to find the one thing' that' is slowing you down the most and fix that' first. Here's one really good way to think of it: your marketing process is only as fast or as effective as its slowest and weakest point. If you make any other part of the system better, you haven't actually changed anything. The process is still only as fast or as effective as the slowest or weakest part.

I think the best metaphor for this is a really bad traffic jam. So think about a highway, like a four or five-lane highway, that' all of a sudden goes down to one lane. So we had five lanes of traffic that all of a sudden squishes down to one lane. So it gets really, really backed up on the five-lane side. Obviously, the one-lane part is the constraint, it's the bottleneck.

That's the part we have to fix. If we spend a ton of time and effort making the five-lane highway part better, if we make it six or seven or eight lanes, or if we make the cars on that side go faster, it doesn't help. It just added more pileup to that side.

Or if we make the part on the other side of the bottleneck better, that doesn't help either. So improving any part of this road that is not the single-lane portion didn't help.

We have to fix the single-lane area, or we didn't do any good. So that is the most helpful metaphor. Keep that in your brain. That's what we have to figure out. We have to find the single-lane part of our process and tackle it first. Anything else, putting our effort anywhere else isn't going to help.

So before we move into the five focusing steps of the Theory of Constraints, I want to ground us quickly in where we are in the Agile marketing operating system. This won't take long because it's really simple and obvious this week. We are squarely in how. We are focused on the how. We do work, just one pillar, just one area, and it deserves special attention, because as you can imagine, in the traffic jam example, this can slow everything else down. We can have the best teams in the world, we can have the best strategy in the world, but if we've got terrible execution, we risk having none of that matter. So, really important that we get the how right, so we're gonna spend today's episode really just focusing on that. All about how today. All right, so let's get into the meat of this episode and talk about the five focusing steps. So, the first one is super straightforward but must not be skipped under any circumstances. You must first identify the goal of your system. So, this is really actually step zero, so it doesn't even get into the five focusing steps. It's the pre-work before you even get into the focusing steps. Because if you don't know what you're trying to accomplish, all of your other work may be in vain.

So for marketing, this is really important. What are you trying to do here with your marketing work? Are we thinking about brand awareness? Are we thinking about lead generation? Are we thinking about customer retention, conversion rate optimization? All of these are very different sets of activities that are going to be optimized in very different ways. So, goal first and foremost. Obviously, the name of the book is The Goal, so this is critical. Once you know that, and everyone has agreed on that, then we move into optimizing for it, and then we have to first identify the constraint or the bottleneck. So remember our traffic jam example. This is the place where all the cars are stuck. So ask yourself, "Where is the flow of work slowing down most often?" Where do we have excessive waiting? Where are there a lot of delays? This might also be where you find yourself having a lot of rework. Where do we have to go back and redo things?

Or maybe there's a lot of back and forth edits, stakeholders are always coming in and making changes, maybe that's the area. Or we've got a spot where a lot of other tasks are piled up behind this one spot. Another thing you might want to ask yourself to help identify where the constraint or the bottleneck is, is if I could speed up one step, what step would I speed up to make everything else faster or easier or clearer? Right? So, take your time here. Don't race through identifying the constraint, 'cause as we've said, putting your effort into the wrong spot can actually make everything worse, right? You might just make more traffic.... in the traffic jam if you're not really focused on the constraint. So take your time here, really identify the constraint, and then we move on to making it better.

And now, the way that the next steps are structured is important. They go in order of their level of disruption. So we wanna go through them in this order, because the first ones we're gonna talk about are the least disruptive. They're going to cause the least impact to your system. So if you can improve the bottleneck with the least disruption, that's what we want to do. So try these first ones first. If they're not giving you the impact you're looking for, then go down the list. But if you can do a little bit and see a big impact, then great. Then you can move on to the next bottleneck. So the first one is called exploit the constraint. It's a little bit of a problematic name, to be honest, because when people hear this they automatically think it means, like, chain the person to their desk and don't let them leave, but that is not what it means. This is not just make the person who is the bottleneck work more.

That's not what it means. What it means is, can we make the bottleneck more efficient? So could we take away work that is distracting to them? Could we clarify instructions or criteria? Uh, things you've probably heard on this podcast already are things like definition of done or definition of ready, so providing more clarity around what it means for work to be ready to get started or ready to be finished so that the person or people who are doing the work at the point of, of the bottleneck don't have to spend a lot of time guessing or a lot of time chasing people down to figure out if the work should be started or if the work is done. They already know. So we wanna eliminate any and all distractions so the only thing that the bottleneck workers are doing is bottleneck work. Another easy thing to do in the constraint, or excuse me, in the exploit the constraint step is to try to figure out a way to streamline any reviews that are happening in this step. Can we use AI to do a first pass so that reviews are kind of just a final check box moment? Or can we do them async so that we don't all have to, like, get on the calendar and wait for those things to happen?

All of these things can streamline the bottleneck workers' time. So again, we want them to only be doing bottleneck work. All right, so those are some of the easiest things, we wanna try those first. The next step, the next focusing step is to subordinate to the bottleneck. So we are making all other work in the system subordinate to bottleneck work. Anything that we can do to help the people working on bottleneck work, we want to do that. We don't wanna just pile more work in upstream, or to come back to our traffic jam example, we don't just want to add more cars onto the road.

That is not helping. We are not making anything better by putting more work into the system. So if we are backed up on and we have 10 campaigns that are stuck in the bottleneck, launching 10 more campaigns that need to go through design isn't gonna help, or those 10 campaigns need to go through setup in the smart tech stack, not gonna help.

So that means subordinating the upstream work to the bottleneck. Don't launch new things if the bottleneck is backed up. And same thing, what else could we do downstream to maybe help take some pressure off of the bottleneck? So remember back to some of the core principles here, the whole system can only go as fast as the bottleneck goes.

So everybody has to be on board with helping the bottleneck go faster because everyone benefits. So we, everybody should be keen to change their own work to help the bottleneck.

So this isn't just the bottleneck's problem; this is the whole system's problem. So we wanna subordinate everything in the system so that the bottleneck can go faster.

So that's the third step. Fourth is to elevate the constraint and now we can start to maybe add more AI stuff in here because this is gonna be more disruptive. This isn't necessarily just changing the way that the existing system works. Now we're starting to change the system itself. This might be adding more resources in, so this could be hiring more people with the skills needed to help the bottleneck resource. This is kind of like adding another lane to...... the traffic jam spot. That's a big deal, right? Construction of a brand new lane is not a small thing. So we don't wanna do that until we've tried the other stuff first. But maybe that's what we need to do. Maybe it's a hire, maybe it's a freelance thing, maybe it's a tool as well. So maybe this is an AI tool opportunity for some automation, for some additional support of some kind to help expand that bottleneck so it's not quite so bad.

Can A- AI help with drafting or triaging the work so we're really only working on the highest priority? Could it summarize or do an initial proof on some of the work? Could it remove some steps entirely, and could we have some agents doing some of the more manual, repetitive steps that are part of the bottleneck's work?

So these are more complicated, more complex ways, but they may be necessary if none of the more simple solutions are working for us. But again, we don't wanna jump straight here. We wanna try some of the simpler things. If they're not working, then we can try these more complex, um, more intensive ways to address the bottleneck.

So when we get through this fourth step, now we get to step five, which is repeat. Because once we have smoothed out the flow in this one bottleneck, we will find another one, and that is great, because this is an iterative process. Fixing one bottleneck doesn't mean that everything is great and that we now have a perfectly moving five-lane highway without any delays or any issues. I'm sure if you've ever driven on a five-lane highway, you know that's not how highways work. Something else is gonna come up. Fixing one bottleneck doesn't make the whole process perfect. What we're really doing is just making it a little bit less broken one piece at a time. Okay, so let's talk about AI and continuing to apply the Theory of Constraints, because AI is really not a flying car.

It is widening the bridge so that we can ease the backup, and there's gonna be a backup somewhere else, so we're going to k- need to keep doing this over time. So a couple of places where we might wanna keep doing this as time goes on, because more bottlenecks will show up in new and different places. So maybe we have slow content reviews. There's still gonna be definitely needs for human interventions, human oversight into content. I think we all know better than to just hit publish on AI-created content at this point. I hope you know better than to do that. Um, but AI summaries so that we can have digestible action items from meetings, all of those things are really, really helpful. We don't have to dig through a whole bunch of clunky, big piles of text, right? So these are great use cases, um, not having to go through pages and pages and pages of reporting data, being able to have auto-generated dashboards or only flagging critical data that is too far off of a baseline. Again, we don't have to wade through tables and tables and tables. This is a great use case. So if reporting is our emerging bottleneck, then this is a great use case for AI. If we are constantly being, um, barraged by incoming requests and we're struggling with prioritization and we spend a lot of our time going back and forth trying to get data from stakeholders or get requirements or briefs, we can use AI to triage, categorize, and even prioritize based on the information that people provide and help eliminate that bottleneck. If we find ourselves doing a lot of rework because we didn't have the necessary data people provided, requests, then AI can make sure that we do, right? And it can be the one to go back and forth.

It doesn't ever get frustrated. It doesn't ever get tired of going back and forth and back and forth, again and again, until it gets everything that it needs, and not submitting the form request until it's all filled out.

So wherever the bottlenecks start to emerge, we can apply AI as a precision tool. We identify the bottleneck. We apply AI intelligently and with precision, not as a sledgehammer. That means that we can have a system that thanks us and gets better all the time and we are not a cautionary tale of mismanaged AI implementations with no  ROI. If even after, you know, we've talked about all of these focusing steps and Theory of Constraints. and you go and read the entire Goal book and you're still not sure how to find the bottleneck in your system, we do have our Agile marketing assessment, which will give you a score across all nine dimensions of our operating system, you know, across the what and the why and the how. It's gonna give you a score for teams, process, data, practices, mindset, customer centricity, people, leadership, and planning.

So you will definitely know where the bottleneck is, then you can focus your AI application efforts where it will matter most. Link, as always, will be in the show notes, or you can find it at AgileSherpas.com. In the meantime, thanks for being here. The struggle, as you know, is real, but so is Agile marketing. See you next time.