The Digital Project Manager

Why Tracking Time Won’t Save Your Agency’s Margins

• Galen Low

Share episode

Feeling behind on your agency's data game? Marcel Petitpas, CEO of Parakeeto, joins Galen Low to reframe what it really takes to build profitable, data-driven operations. Hint: it’s not about buying software or pushing for time tracking compliance. Marcel breaks down why starting with metrics, models, and meaning—not tools—is the real shortcut to operational clarity.

With real-talk on utilization myths, data fluency for PMs, and how AI can actually drive measurable impact (without overwhelming your team), this episode is a field guide for agency leaders who want to do more than just keep up. It's about building a smarter foundation for the long haul—and doing it in a way that doesn’t burn you out.

Resources from this episode:

SPEAKER_01:

If somebody's listening and their organization doesn't have a culture of data, how long will it take to catch up?

SPEAKER_00:

The first step is getting clear on your framework. You can't make decisions about what tools to use, how to use them, what to track, what not to track, unless you know what metrics you're going to pay attention to. You know exactly how those metrics are going to be calculated, how you're going to define every object within every metric and what that metric tells you about your business when it goes up or when it goes down.

SPEAKER_01:

If you are operating an agency, how do you decide within this framework what tactics are actually going to help net the strongest impact to operational efficiency?

SPEAKER_00:

We think about this as the agency profit flywheel. And the flywheel has four steps. Number one is your assumptions, number two is your actuals. And those two things together is your quantitative feedback loop. That tells you where is there a gap, and that tells you what metric within the business do we need to move.

SPEAKER_01:

If you are an operator, you need to be very data fluent, very data-minded, data-driven, more technical, more of an engineer, or is that a fallacy as well?

SPEAKER_00:

There will be a point in the not so distant future that you will be able to build and manage a little data pipeline without having to be hyper technical or like a super duper data nerd. However, what will still be true is.

SPEAKER_01:

Welcome to the Digital Project Manager Podcast, the show that helps delivery leaders work smarter, deliver faster, and lead better in the age of AI. I'm Galen, and every week we dive into real-world strategies, new tools, proven frameworks, and the occasional war story from the project front lines. Whether you're steering massive transformation projects, wrangling AI workflows, or just trying to keep the chaos under control, you're in the right place. Let's get into it. Today we are talking about the relationship between profitability, operations, and data. And specifically, when is the right moment to ensure that your data is clean enough to understand profitability, as well as how can you catch up if you're already feeling behind on the whole AI-enhanced data-driven operations thing. With me today is Marcel Petipa, founder and CEO of Paraquito, and also my go-to person when it comes to profitable data-driven agency operations, and also explaining linear regression to me like I'm five. Marcel and his team act as the middle ground between having a fractional CFO and a fractional COO working together to get your agency growing predictably through sustainable and profitable operations. Marcel has also been everywhere lately. He was speaking at the Agency Growth Event Summit, the All-In Agency Summit, Mirin AI 2025, Consulting Success Live 2025, and various appearances with the Agency Management Institute. And he's also the host of the Agency Profit Podcast, which releases weekly discussions with agency leaders and industry experts. Marcel, thanks for coming back on the show with me today.

SPEAKER_00:

Galen, watching you do that intro was breathtaking. You were so locked in. It was like incredible. We're like chatting behind the scenes, and then you just like you went into the zone. Thank you for having me. It's a pleasure to be here. And thank you for the wonderful performance just now.

SPEAKER_01:

Thank you for the compliments. I'm actually AI this whole time. Marcel built me. Marcel, you've been on the show before. You've done events with us. I love having you on the show because I think you just have a really unique perspective and a pragmatic perspective on where agencies and just businesses in general are going in terms of data fluency, data-driven operations, and how that connects into value. You actually get down to business.

SPEAKER_00:

Well, I appreciate you saying that. It's something that I strangely spend an inordinate amount of my time thinking and talking about. And I'm very lucky that I get to do that. You know, I've gotten my hands dirty over seven years actually implementing these things in agencies. And I think that's where, especially in the context of project management, I have some perspectives that I think go against conventional wisdom in terms of like what best practices are, because I've seen where the golden handcuffs are. You know, the decisions that we make that seem like they make sense. And then in practice, it's like this just never works because it's too hard.

SPEAKER_01:

I love that. Yeah, we like keep using the playbook that's kind of broken just because it's there. Golden handcuffs, indeed. I was thinking that we could start off by just getting one big burning question out of the way, something controversial that everyone wants to know the answer to. And then I thought maybe we could like zoom out from that and talk about three things. Firstly, I wanted to talk about like data hygiene and what to do if it feels like your agency has already kind of like missed the boat. And then I'd like to talk about how exactly clean data and AI can make a measurable impact on profitability, as well as like how to score and prioritize operational efficiencies based on impact. And lastly, I thought maybe we could explore where the rabbit hole goes for agencies. In some ways, the agency game has changed completely. And I would love to talk about like the skills and structures that agencies need to look at to thrive in the age of data and AI. Big topics, but um, how does that sound to you? It sounds like a really fun time. Let's dig into it. Awesome. Let's get into it. Here's my big hairy question. I would say good clean data, this is what I learned from you. Good clean data is arguably the key to being able to achieve profitable data-driven agency operations. But if somebody's listening and their organization doesn't have a culture of data or isn't anywhere close to having data organized in any meaningful way, how long will it take to catch up?

SPEAKER_00:

And is there a way to accelerate that process without being reckless? The answer is yes. The thing that I find so sad about so many of the clients that we end up working with after they've spent years trying to close this gap is that the vast majority of them try to wag the dog by the tail in how they approach this problem. And this is going to be especially true of, I think, a lot of the people that are listening to this show. And I don't think it's their fault. I think it's honestly like the software industrial complex has sort of conditioned them to believe that it starts with the software and it starts with the project management and the time tracking data and the accounting data. But those are all the last steps in the process and the least important. And that's gonna like shock people that here I am saying that. But it's like you don't need time tracking data to get 80% of the visibility into your performance. The first step is number one, getting clear on your framework. This is a little bit abstract, but like to put this in practical terms, you can't really make use of any of your data. You can't make decisions about what tools to use, how to use them, what to track, what not to track, unless you know what metrics you're gonna pay attention to. You know exactly how those metrics are going to be calculated, and you know exactly how you're going to define every object within every metric and what that metric tells you about your business when it goes up or when it goes down. And to give you a practical example of this, it's like it's great to sit around the boardroom table and say we're gonna measure utilization. It's another thing entirely to say, well, what exactly does that mean to us? What is a billable hour and what is capacity? How do we define those things? What's included, what's not included, what counts, what doesn't count. What is the logic statement by which we actually convert data into that depiction of utilization? And do we fully understand what makes it go up and what makes it go down and what that means about our business and how it influences our decision making and why we care in the first place? What will change about our decision making if we have this number in front of us and we see that it's going up or down? So that's actually the first step. Good news about that is that doesn't require any software, doesn't require any data, doesn't require any waiting to amass data. Those are just decisions that you can make. They're not easy to make, but we've tried to make it easier by publishing our framework for how we think you should measure those metrics and what metrics to measure and make them publicly available. So I'll say that that is step one. And there's a second step which also doesn't require net new information that I think is important to speak to. But I'm gonna take a moment here to uh let you get a word in edge-wise, Galen, because I could just keep going for like the next 60 minutes. You know that.

SPEAKER_01:

And I would be happy if for you to do that. But also uh one thing I love is that I maybe deliberately sort of did I lead the witness a bit in the sense that I think a lot of folks listening are like, gosh, we need to catch up, we need to buy software and we need to start time tracking right away because then we get a baseline of data, but we're still like months and months away from being able to like be anywhere close to having usable data that we can use to drive profitability. That's like the sort of like involuntary reaction, right, to feeling behind as an agency. So I love that it's not software as being the answer. I wonder if you could give me an example, especially on the utilization thing, because utilization, I think is probably a perfect example of those golden handcuffs, right? Because it makes so much sense. I've hired these people. My business model is selling services by billable hour. They are either spending most of their time making money or they're not. And if they're not, then my margins are collapsing. I might be losing money. I'm not leveraging the talent that I've brought on. And it seems so natural to be like, yeah, that's a good measure. But where you got me was what makes it go up and down? I was like, oh, it's such a complex equation. For folks who are like, Marcel, come on, what else am I going to track? Utilization definitely has to be it. What are some of the myths that you either debunk or run into that actually are fallacies? And does that still lead to still tracking utilization, but better, having asked the right questions, or is it abandoning it?

SPEAKER_00:

No, our framework includes utilization as a metric, but the way that we calculate it is probably not the same as what most other people would calculate it. And I could talk about that. I could talk about how we define utilization, but it's hard to do that in isolation because until you understand how we measure average build rate and how we measure profit margins on projects and how that's different than how we measure profit margins at the agency level and how we think about cost per hour, then you can't understand why we made the decisions that we made on utilization. The truth is we are thinking about this as an entire framework. So the relationship between the metrics in the business and the metrics in each other. And we've thought all the way through this and we've tested it with hundreds of agencies over seven years and worked through all the edge cases. So, like I could go through my definition of utilization, which is delivery hours. So any time spent on clients, regardless of if they were air quotes billable or not, we don't care. And those delivery hours are divided by capacity, which is all of the time that you purchase from anyone in the measurement that you're doing. So for the whole agency, it's everybody, including people that do no client work at all, and it's 2080 hours per year. So there's gonna be some people who are gonna hear that and be like, well, that wouldn't work for us. And maybe that's true because you made a whole bunch of other decisions about the other metrics, and that doesn't make sense. But I just spoke to somebody today, this morning, actually, that was talking about how every week their management team tells them that their team is fully utilized, but it's not actually true because half of that time that's being considered air quotes billable is internal time that's not earning them revenue. And so they're getting a false positive. Or this other team, they're most utilized in August, which is a false positive because they're taking a bunch of time off. And the way that they think about capacity, well, we should be stripping out time that's not available. Right. And so all of this to me is a signal that that conversation around how do we define utilization did not happen after a conversation about why and what decisions are we making and what questions are we trying to answer. Because if that conversation had happened first, they would immediately see that well, well, we can't pull capacity out if what we're trying to figure out is what is the financial impact of utilization and how does it affect our profitability? Because that would, you know, this is problematic. So anyway, I digress. I think the big trap is that you can discern the performance of your agency or of your business from a single metric. And I think the trap that people fall into is they try to make utilization that one metric, or they try to make project level profitability that one metric. And I think that that one is equally, if not more, problematic. You cannot infer the profitability of your entire agency from the profitability of one project. And trying to couple those things together is a fool's errand and a massive waste of time and resources, in my opinion, and does a poor job of answering any question.

SPEAKER_01:

What I like about it is the Douglas Adams aspect of it, you know, when they're like, what is the meaning of life? And the answer is 21, but it's because you didn't really understand the question to begin with. That's like utilization to me, based on that explanation. All right, so part one, the framework, understanding what you're measuring and why.

SPEAKER_00:

Part two. Part two, I would say, is your model and your forecast. This is another thing that I think gets really overlooked, is and it comes back to this question. Okay, let's imagine we've defined how we're going to measure utilization. Then the next thing that tends to happen is, you know, the PMD gets all excited. They go, they roll out the project management tool, they hammer everyone for compliance. Three months goes by, they're finally like, okay, we've got some good data. Our utilization rate last month was 54%. And then everybody stands around and kind of looks at each other, waiting for someone to react. And then they're like, well, is that good or is it bad? And so again, you've gone to the end, the last step in the process, but you can't get value out of that measurement because you never established your expectation. So the good news is setting your expectation can be extremely insightful and it can tell you where a problem is without even knowing what the actuals are. And when you get to actuals, they are significantly more valuable because insight is a function of expectations contrasted to reality. So step two is let's build a model of your firm. And that model is very simple. We start with what we call a payroll grid, which is a list of everybody on your team, how much they get paid, how many hours they work, where you expect that time to go if you had unlimited work to do, how much paid time off they have. Then you look at your overhead expenses. What have you been spending on? Your lawyers, your accountants, your QuickBooks, your subscriptions, you know, all that stuff, your office. And then we look at pricing. What do you effectively charge? And what do you expect to earn for every hour of time that gets spent doing client work across all the things that you sell? And if you combine those three sets of information together, then what you should be able to arrive at is if all of these things went the way that we planned for them to go, how much money can this business earn with the current team? How much time will be spent doing client work? What will our utilization rate be? How much will we spend on overhead? What will our gross margin be? And what will our net profit be? It's like, what is the best possible outcome? That sets the foundation to say, well, first question, is this good? Is this what we expect? And this is an important question because sometimes we'll come across a firm that's like, no matter how hard we try, we just can't get profitable. And then we do this exercise and we're like, well, that's because you're planning to be unprofitable. Interesting. And they're like, oh, right. So, like, no matter how perfectly we execute this plan, we will still fail. And it had nothing to do with our execution at all. Actually, we just had the wrong plan. So that is step two is build a model of your firm, understand what it's setting you up for, how much room for error you have, what the targets look like. And if you layer in one last set of data, which is high-level understanding of projects, how much you're going to get paid, how when they start, when they end, how many hours you expect, then you can start to look out into the future and say, okay, well, are we planning to earn as much revenue as we could? Are we planning to keep the team reasonably busy? And you can solve about 70, 80% of profitability problems without ever tracking time or tracking project data, which is the most expensive and hardest part, just by number one, getting clear on your framework. And number two, getting clear on the assumptions that you make about your business today and just structuring those in a way that is accretive. And those are the first steps that we take with every client because you can't really make decisions about any of the downstream stuff until you set that foundation and determine what that structure looks like. So the good news is all of that is free. All it takes is time, and it only requires information that you already have in your mind. You just probably haven't taken the time to lay out. You don't need software, you don't need to wait six months for data to accrue. You don't need to change your accounting process. That's available to you. So that is the good news, is it doesn't take long to catch up, actually, because a lot of that is available to you right now.

SPEAKER_01:

It's interesting because it's it's free, but it's not easy. In other words, some of these questions are based on knowledge that you may know, but answering the questions is actually the hard bit because it really, you know, you really have to stare your operation in the face. I actually thought you were going to say, think about how much money you could possibly make and then tell yourself that you are unrealistic because that is the utopia. Whereas you actually went, your plan might just not be profitable. Like how you operate the business might not be profitable. And the one thing that I think is really refreshing from my perspective as an agency PM guy, where we are accustomed to feeling like we've got the weight of the agency on our shoulders, right? It's like we are the economic engine, we are paying for all those people that work in the red, you know, like we need enough margin so that we are billable and usually like, you know, 80% utilization or higher, and also have enough margin to make sure that the administrative staff, the sales team, everyone also like we're paying their salaries too. Whereas I think the other thing you said in there, I hope, I hope I'm understanding this right. But it's like, it's not solely, it can't just solely be up to project profitability. There's this sort of like the COGS model, and you have to factor everyone in, and you have to factor in what you plan to have as a villable work or available work or money that you can make.

SPEAKER_00:

This is 100% true. And we've seen this before. We've seen firms that had every single one of their projects was wildly profitable, but the firm was not profitable because project profitability is only one of three factors that determine profitability. Just it can't come down to that. That's just how it is, right? It's like I could sell you this water bottle for a million dollars and do that once every 10 years. But if I then turn around and spend$800,000 a year on operating and on marketing to try and convince one person every 10 years to buy a million dollar water bottle, it doesn't matter that my margin on this was 99.999%. That's just the nuance of it. You have to zoom out and build a comprehensive model of the entire firm and pay attention to all of the levers that affect that number. Meanwhile, those Stanley Cups, that is their business model, right?

SPEAKER_01:

Tell some collectible water bottles at margin. Yeah, that are indestructible. This is actually probably a good launching point because I wanted to zoom out a little bit. Something that you had said, you had described what you do at Parakeo as being kind of like if a fractional CFO and a fractional COO had a baby together. And setting aside all the sort of conflict of interest implications there and all the cold play concert jokes, how can finance and operations leaders sort of get more plugged together to start using data to analyze and monitor profitability within an agency, or more to the point to like make these decisions together, to answer these tough questions together? And also, are those the key players or like who else needs to be involved in some of these conversations to figure out these frameworks and the model and to debunk any sort of potential myths that they've been operating around from a financial standpoint and from an operational standpoint?

SPEAKER_00:

I actually think it comes back to the first two things that I spoke about in my last answer, which is the conversation about the framework and the conversation about the model of the business, it is inherently related to the three most important parts of the business, which are the financial part of the business, the operational part of the business, and to some degree, new business. And so by sitting down and having a conversation that's centered around that, it sort of forces everyone to come together and come to an agreement. And in some ways, it also provides a backing to that conversation because it's just a math formula. It's not an opinion. It is nobody's opinion. It's just like this is the math. These are the levers. When they go up or down, this is what changes. And nobody can really refute that. That creates, I think, the foundation for everyone to then say, okay, well, if this is the expectation, then how are we going to measure against this? And everyone has to contribute to that. And it sort of forces you to now look across the departments. Well, if we want to be able to forecast, we need to know what's coming down the pipe from sales, right? So, how do we do that? What information do we need? If we are going to be forecasting our capacity, then we need to have an understanding of how much time is going to be expected. If we're going to forecast the financial side of this, we have to have an understanding of what we're going to charge and what our costs are going to look like and how much of that is our money and how much of it is not. And if we're going to plan for profitability, we need to know what the margin looks like. And similarly, if we're going to set budgets on the finance side, then we have to have an understanding of what our team looks like and what our hiring plan looks like. So all these things are sort of inherently connected. And you start to look across the organization to say, well, if we want to measure this, what data do we need? And inherently the answer is almost always some combination of operations data, because operations data precedes financial performance and then finance data, because finance data is essentially the summation that is very accurate of like what has happened in the past. And so to me, like the framework and the forecast and model are the foundation on which the feedback loops are built, and the feedback loops are where a lot of the cross collaboration ends up happening. But if you don't have a framework and you don't have an understanding of how to build a model, then what I think you often find is everybody just sort of stays in their lane and there's this little gap that just never seems to get filled and it drives the leadership team crazy. And you end up with just these like really disconnected things where it's like the way we think about selling work versus the way we set it up in QuickBooks and invoice for it versus the way it goes into clickup are all completely different languages. And it's like no wonder we can't measure anything. Like this data is just not matching up.

SPEAKER_01:

I like that. I like that the model and the framework is what brings people together. And I've experienced that small gap, right? Where everyone, it seems like everyone goes into the boardroom and agrees, you know, strategic planning. And we're like, okay, we're gonna do this. We've got the data, here's our plan, here's our goal. But the puzzle pieces don't fit together and the language is different. And actually, next QBR, everyone's like, yeah, but that's not my fault. That's finance's fault. That's HR's fault. I like the idea that because even coming into this, I'm like, okay, so a fractional CFO and a fractional COO get together and have a baby, and that's parakeetido. I'm like, okay, so then is it led through operations and finance? But really, it's the role you play is to bring that sensibility of the model, like building a model and building a framework, but then bringing everyone together from all the different departments so that you can look at the business holistically and using data that probably exists.

SPEAKER_00:

And, you know, in larger organizations where there is a COO and a CFO and a CEO, what we do is we get them all aligned and then we allow them to go spend time in their highest leveraged place, which is not in spreadsheets figuring out what the hell's going on. It's going to do something about it with an understanding that everyone knows what they should be focused on. And in smaller organizations where they don't have the scale to have a CFO or COO, we're doing that same thing, but for whoever the senior most person is in each of those departments, which is usually the outsourced accounting firm on the finance side, and like the senior PM or like the PM that down just got an operations title because it's like, well, we're at a place where we feel like we need that and you're here and you're you seem good. So do you want to learn how to do operations? And the CEO, who generally did not get into this to deal with any of this stuff, and they're a subject matter expert. So it's about really kind of taking that messy middle part on so everybody else can just get clear on what they need to do and then go do it.

SPEAKER_01:

That's such an interesting one because I I've worked in several boutique agencies where that's definitely the case. The CEO is the founder. They just really liked digital technology. They're like building stuff, websites, you know, apps, things like that. They like design. And suddenly there's a CEO that they've never been a CEO before, they're not trained as a CEO. They join a club with other CEOs to kind of learn from them. But there are naturally with anyone, but definitely in that situation going to be gaps around, you know, what can be done, should be done. And that operations thing. Yeah, definitely. I've, you know, I've seen that a couple of times of, yeah, you could you could probably run operations. You're the COO now. We'll figure it out. And I like the idea that that's okay because everyone's on some kind of learning journey, but we need to have a plan to build that sensibility, fill those gaps, and drive those conversations so that we're learning from one another and making good decisions for the business together.

SPEAKER_00:

And I think there's something here that's worth noting, which is what I've seen far too many times is that senior most PM person that sort of gets voluntoled to become the operations leader, or is, or is you know, maybe that's what they want. They get into that role. And then all of a sudden the CEO is like, uh, yeah, figure out this profitability thing. That is an unrealistic expectation, and they are being set up for failure. That is not a side quest. It's a very hard problem to solve if you've never solved it before, if you have no framework. It's really hard. And it requires technical expertise in data, maybe not data engineering, but at least data analysis and data science. And it requires a very high degree of financial acumen. And it is just unrealistic to expect a person that has come up through project management to do that. And the other thing that I think is true, and we've talked about this in other episodes, is that project management sort of background becomes a bit of a disadvantage here because most project managers have been rewarded for being highly precise and detailed. And that is exactly the thing that will often hold them back from being able to do a good job of executive level reporting, which counterintuitively is successful generally when we relinquish a lot of the precision that we conflate with accuracy. A great example of this is the way to do a four-month capacity forecast for your design department is not to go assign 700 tasks to individual people inside your project management tool. That is not a good path. That's like using chopsticks to stack grains of sand together to build sand class. It's like it's not, it's not going to work. But if you come from a project management background, that's probably the first place your mind goes is well, let me go stack all these tiny little blocks together in our project management tool because that's all you've ever known.

SPEAKER_01:

Even though I know we've touched on that topic before in other episodes, now that like a bit of time has passed, my perspective on it is completely different, especially in the age of AI, I guess I would term it. And I'm glad you took it to where you took it because that's exactly where I wanted to go. And what I was teasing at before is like, you know, I know a lot of folks who may have fallen into an operations role and they'll be like, okay, well, I can do all these things that Marcel says, but I need to now hire a data scientist. And I've seen it happen where it's like, okay, we hire someone, you're in charge of data now because I don't have a data sensibility, you know, like I don't know where to get it. It's a gap for me. But then that data person isn't the person who understands a bigger picture on like what data to actually gather and how to package it. And then it's basically blind leading the blind. Is it just like a you said that sort of solving the profitability equation, someone could be set up for failure unless they have been trained to do it or have done it before? Naturally, when you think about it, it's like, yeah, if it was easy, everyone would run a profitable business, I'm pretty sure. But then tying it to the sort of data sensibility, the numbers, the accuracy, but not necessarily the precision. If somebody's in that role, is it table stakes now? Where like if you are an operator, you need to be very data fluent, very data-minded, data-driven, more technical, more of an engineer, or is that maybe a fallacy as well?

SPEAKER_00:

I think we're at a really difficult intersection right now. And I'll explain why. I think that data operations will become table stakes for small businesses, but there is a vacuum around that currently. And this is, by the way, the dirty little secret about Parakito is our entire thesis is just taking data operations, which has been around for a long time at the enterprise level and predicting that it's going to come down market and trying to package it up in a highly opinionated way for a very specific slice of people to make it available at a cost that's tenable. That's our whole business model, basically. It's like we're like a little data operations team. We've just packaged it up in a fancier way. And the problem is that like you can't go and sign up for a Databricks account as a million-dollar agency and afford the subscription for that and the technical people that are going to need to come in and build that out. But there is no Databricks for small business, really. However, those tools are coming and they're getting more capable. And so I think that there will be a point in the not so distant future that you will be able to build and manage a little data pipeline without having to be hyper technical or like a super duper data nerd. However, what will still be true is you will have to have an understanding of the framework and how to apply that framework to your business, right? So like there is no way around beginning with the end in mind. All of those tools, all of the data, all the spreadsheets, even if you were a, to your point, if you were a data scientist or a data engineer, if you don't know what you're measuring and why and exactly how to define it, then like it's all a path to nowhere. And it's all a big waste of time.

SPEAKER_01:

For our operators listening right now who are like, oh my gosh, Marcel just described me and I don't know how to work backwards from the end. And I have been fixating on waiting for my data breaks to become within reach, where should they start? Where can they fill that gap? Or is it just like get out of your career now?

SPEAKER_00:

Uh no, no, no. This is an interesting thing, actually. Like there's the question of what happens to PMs when AI is a thing. I actually think project and account management is like one of the most insulated disciplines inside of these businesses. And is actually when you strip away a lot of the execution, is some of the highest value stuff to be doing inside of these firms. I could talk more about that on a side tangent. But to answer your question directly, like where do they start? Shameless plug, start with the agency profit toolkit. The framework is in there and it's free. It's simple. It's not easy to your point. And that's the reason that people hire us to like take them through the process of installing this in their business. But if you want to know what the metrics are and how they're calculated and what the definitions for the objects are, I've given all of that away for free. So that's where I would start.

SPEAKER_01:

Love it. I'm gonna shift gears a little bit here because you had also just recently done an interview with my team on the digital project manager. I'll link it in the show notes. It was great. We covered a lot of ground there. And one of the things that we got into that I think is related here is that I think a lot of people are looking at the profitability equation in the context of AI right now and going, well, my business is built on dollars for hours, billable hours. We measure utilization. AI is here now, and it's promising operational efficiencies to make our profit. It's gonna 10x our profit, it's gonna 100x our profit, but there's just so many things you can do, right? Like it's either death by a thousand paper cuts or pick one and hope for the best. I'll use as my example. You mentioned like AI for context gathering, right? Just like pulling together all the pieces from different tools so you're not spending time there. But I thought I'd just like kind of skip to the main question, which is like if you're like operating an agency, right? And everyone's kind of like overwhelmed with how many options and use cases there are for AI, how do you decide within this framework, within your model, how do you decide what tactics are actually going to help net the strongest impact to Operational efficiency. And is that even a path to profitability in your mind?

SPEAKER_00:

It absolutely is. I've seen AI impact all four levers that can move profitability in a firm. So I think actually you're really going to like this answer. And the PMs listening to this are going to like this answer because it's very tactical, actually. We think about this as the agency profit flywheel. And the flywheel has four steps. Number one is your assumptions, number two is your actuals. And those two things together is your quantitative feedback loop. So what did you think was going to happen and what was the plan and what actually happened? That tells you where is there a gap between what we expected and what actually happened. And that tells you like what metric within the business do we need to move? We can come back to like, there's really only four choices. Once you know what metric you need to move, then the question is well, you have reports and meetings with the team about, hey, this is the metric that we need to move. And then as a team, you ideate, well, what could we do to move that metric? And we could talk about the tactics to move the metrics. Once you've discussed those ideas, it should become clear what you think, you know, the sort of quick wins are or the things that have a good cost benefit. That's these are not new things. I don't need to explain any of that to PMs. And then once you decide on what those things are, you go and you execute them and then you repeat the flywheel and you keep doing that until the metric changes and then you shift. And that should just be a muscle that's going on in perpetuity in the business in a perfect world. So at a high level, that's how we think about it. But we can talk about like the specific metrics and how to move them if you'd like.

SPEAKER_01:

Yeah, I was wondering if maybe we could dive into if we used, for example, like context gathering as our example from end to end, where I don't know, let's just say, yes, there is a gap between what we expected our profitability to be versus what actually happened. And I guess maybe picking one of the four. But I'm thinking like, okay, well, my mind immediately goes, okay, well, team is like, yeah, we're spending so much time just like finding things across tools, and that adds up. And that's why we're not getting things done within estimates. Again, I'm wearing my PM hat right now. So the thing we need to do is spend less time gathering assets from everywhere, and then we'll become more profitable. But then I'm like, before we even start looking at profit again in that flywheel, my instinct is like, okay, well, how can we measure how much time we saved by not having to gather all these assets from different tools because AI is taking over? And then I remember what you said earlier, where you're like, PMs are about precise data and that doesn't always matter and it doesn't always make it accurate. And I'm like, oh, did I just jump to that thing where I'm like, okay, we're gonna save time. Let's measure how much time we save using AI, and that will tell us a leading indicator of whether that's going to actually help us become more profitable.

SPEAKER_00:

No, you are actually thinking about it in the right way. So, and this kind of comes back to the model, right? So the model is effectively just a really big math equation. So I'll take a quick step back and then we'll dive back into this question. When you think about the way an agency gets more profitable, there's two macro things that you can do. You can either cut your overhead or you can increase what we call delivery margin or what other people might think of as gross margin. Almost no agency has an overhead problem. I could tell you from looking at over 300 of them, there's like a handful that actually had an overhead problem. It's almost never the issue. You might not believe that because your PL isn't structured properly and everything is being called overhead, but like when you look at your PL in what I believe is the correct way, there's a very clear distinction between we spend too much on overhead, which almost never happens, and the real problem, which is we are not earning our revenue efficiently enough. And therefore our delivery margin isn't strong enough, or our gross margin isn't strong enough. There are three ways that you improve that. Number one, you can improve your average billable rate. Average billable rate is a function of how much money did we earn and how many hours did it take us to earn it. You divide those things together and it tells you for every hour of effort, we earned roughly this much money. So, how do you increase that? Well, you either charge more money without changing the amount of time it took you, or you leave the price the same and you figure out how to get it done in less time. So that would be a way that you could measure the impact of this. The second piece is utilization. So, what percentage of the time that we buy from the team actually gets used to earn that average rate? And then the third piece is the average cost per hour. How expensive on average is an hour of time to get these things done? And so a perfect example here is if you were to apply AI to all this context gathering, you might find that it impacts all three of those metrics, actually. So, okay, 20% less time to get this thing done, therefore our average billable rate goes up by 20%. Also, our utilization is increased slightly because for whatever reason, this makes us like able to get more things done and keep the team busier because there's not as much context switching going on. And then the third thing that it might do is well, actually, now because there's not as much judgment required and we're getting all this context in one place, and then we could train an AI on how to analyze that data and output a result. We don't need a senior person to do all of these things anymore. We can get a junior person to get it done because the heavy lifting is already there. So then the formula for profitability is average buildable rate multiplied by utilization equals how much money you earned. And then average cost per hour multiplied by delivery hours worked is the cost. And so you could measure the absolute increase in revenue, the absolute decrease in cost, and the relative improvement in profitability by just like doing that simple math. So you're thinking about it actually in the perfect way. And this is why it's so important to have a model because you can't do that math unless you know what your model is and how you think about that.

SPEAKER_01:

That makes sense now. I'm wondering the model in terms of placing your bets. Can an agency use the same model to place their bets when they know that they've got, you know, they've top graded 10 options where like, okay, we can invest in AI-driven operational efficiencies, but we can't do all 10. Is that kind of the same approach where you're like, okay, well, what do we expect to happen? Well, we expect 20% greater efficiency because we don't have to go and gather assets from all these different tools. That's going to equate to this versus this other option where we'd expect it to only net a 15 to 18% efficiency gain. But then the other thing you said, which is a lot of the time we're pretty bad at estimating, there's also the reality piece as well. Like, can this model be used to like decide where to invest?

SPEAKER_00:

Yes, I think it can. But what I will say is that it's not as simple as looking at 10 options and looking at the impact that they have on various metrics and just like picking the one that has the absolute highest ROI. And the reason I say that is within those three levers, right? I talked about average cost per hour, average buildable rate, and utilization. There is sequencing in terms of like if you have opportunity in all three, there is absolutely a correct answer in terms of which one you should tackle first, second, and third to actually get results. And the answer is you should always start with utilization first, then go to average buildable rate, and then average cost per hour. And so the reason for that is, and I'll just use an analogy, right? So, like if I have a pipe that has water flowing into a reservoir, and my problem is we need more water flowing through the pipe into the reservoir. Well, if currently the issue is that there's only half of the capacity inside the pipe being used, it doesn't matter if I make the pipe bigger, therefore more efficient. I don't actually benefit from that at all. And so I see this happen all the time with firms is they fall into this trap of like, we got to increase our price and we got to get more efficient and cut down the time. But it's like your utilization is way down. So you're not actually gonna make any sure your effective rate and your profitability on projects is gonna change, but the number on the PL won't change at all, right? This kind of comes back to what we were talking about earlier. It doesn't matter how profitable your projects are in this context, you're not realizing the benefit because your utilization hasn't increased. So while that does create a more profitable model, theoretically, from a sequencing perspective, if the goal is to return ROI to the bottom line, that is not the correct sequencing. So in that instance, this is why we do what we do the way that we do it. The first step is what is the metric that we need to prioritize? Then we have a conversation around that specific metric of like, what can we do to get this to the right place? And once utilization is where we want it to be, then everything we do to improve average billable rate goes straight to the bottom line. We realize that efficiency. And also, from a practical perspective, all the things that we usually have to do to improve our average billable rate, we have a far better context to do them in. Because if we got to have a pricing conversation with a client, way easier to do that when we don't care if we lose them because we're maxed out and we have another client coming in and we're gonna have to hire people anyway. So it's like, okay, you could leave, we'll just replace you. That's fine. If we got to cut the scope down on their engagement, we are in a much better position, practically speaking, to actually even execute on the things that will likely come up in an average billable rate conversation. So that context, I think, matters outside of just the objective conversation of like what is the potential ROI here.

SPEAKER_01:

I love that there is a logical sequence in that pipe analogy. And chef's kiss. I just wanted to kind of like land out looking ahead because we've talked about a lot of things. We've talked about a lot of things today, feeling behind, catching up, where to start, software versus frameworks versus actually getting a deeper understanding of your business. We talked about AI and operational efficiency. You work a lot with agencies. And I wanted to just ask a broad question because from my standpoint, I feel like right now, almost every agency is in a state of transformation or are under pressure to be transforming. Some folks are doubling down with AI as part of their service offering. Others are just starting to bake data-led decision making into the process. But if we fast forward like five years from now, what does a profitable agency look like in five years' time?

SPEAKER_00:

You know, it's interesting because it's hard to say what it looks like at a tactical level. What are the services that they're selling? You know, it's like it'd be like 20 years ago. Could you have predicted that there'd be an entire industry of running Facebook? Facebook didn't even exist. We never would have seen this coming, but here's what I know for sure the business model of arbitraging labor will not go anywhere anytime soon. It's like the oldest industry in human civilization. It's been around forever. And it is the most efficient medium in the economy, in my opinion, for filling in the cracks that emerge as things change in the economy, right? Expertise is that like very fluid thing that is so efficient at solving problems in the market. So I don't think expertise is going anywhere. What I do think is happening and has been happening, will continue to happen, is that margin pressure will continue to increase in a lot of the verticals inside of professional services. And this has been happening, but I think everyone's acutely aware of it now because we've experienced like 20 years of maturity in this market within the last like six. It started with the pandemic where we had onshore talent rates go up by 30 or 40% across all the skilled categories, while we were also seeing a complete globalization of competition overnight. So you went from like you had all your onshore costs go up. So that creates margin pressure from your cost basis, while the price pressure increased because now there was a firm in Eastern Europe doing work that looks just like yours that's charging a third as much and are still profitable. And you're like, well, shit, what do we do about this? And all of a sudden, the firm that used to care that you were in New York doesn't really care anymore because everyone's working from home anyway. So that was like a quantum leap in maturing of the industry in terms of margins going down. And now AI has entered the chat and we all kind of can see where that's going. So, what I think is going to be true, and I have this graph, we have a point of view article on our website. There's a graph in there that shows, you know, time. Over time, the margins in the industry go down. And that's just a function of every industry that matures over time. And therefore, the level of sophistication required to run this kind of business goes up. And so I think a couple of things are gonna happen. You're just gonna have to have a grip on this as a smaller firm and have more financial acumen if you want to survive because it's not gonna be as forgiving. There's not gonna be as much free money flowing around unless you're in one of these very nascent, cutting edge, very like low supply, high value industries. The other thing that I think is gonna be true is we're gonna see a fragmentation of the supply chain. We see that in most industries that have really low margins. Like I look at grocery and I look at airlines, for example. Like you sure as shit can't be the CEO of an airline unless you really know your numbers because you have a 3% margin to play with, right? And then when you think about the experience of going on an airplane, okay, I bought the ticket from Air Canada, but they didn't do the food, they didn't clean the plane, they don't run the airport, they don't do any of the maintenance, all of that is outsourced. And every player in that supply chain is super specialized in terms of their operations. That's the only way for them to eke out a margin. So I think that in agencies are going to increasingly move in that direction as well. So they're gonna have to be super disciplined about their finances and super disciplined about operations. And I think there's a lot of people that are listening to this that just like want to make nice things that are cringing listening to this. But I just think it's the reality of where the industry is going, unfortunately.

SPEAKER_01:

And I like that perspective of it happens in other industries. This is not unique and it didn't happen just because of AI it started six years ago in the agency world because it's it's been going on for a hundred years, but it's just going way faster now.

SPEAKER_00:

Yeah. That margin.

SPEAKER_01:

Marcel, this has been so great. Before I let you go, where can people learn a little bit more about you?

SPEAKER_00:

Head to paraketo.com. There's tons of free resources there, including parakido.com forward slash toolkit, where if you want to dive in, learn about the metrics, how to install them, it's all there at no cost. And if you like podcasts, tune into the agency profit podcasts. And if you want to talk to me directly, LinkedIn is the best place to do that. Connect, send me a DM. I'd love to hear from you.

SPEAKER_01:

Amazing. I will put all of those links in the show notes. And thanks again. I really appreciate it. Thank you, Galen. It was a pleasure. That's it for today's episode of the Digital Project Manager Podcast. If you enjoyed this conversation, make sure to subscribe wherever you're listening. And if you want even more tactical insights, case studies, and playbooks, head on over to thedigitalprojectmanager.com. Until next time, thanks for listening.