The Handbook: The Operations Podcast

Growth Thresholds & Growing Pains with Marcel Petitpas

• Harv Nagra • Season 1 • Episode 29

Share episode

Growth is the goal for many agencies, but as your team hits certain size thresholds, your structure, systems, and workflows can start to break down.

Marcel Petitpas, CEO of Parakeeto, has spent years helping agencies around the world clean up their operations, tighten up forecasting, and build financial systems that actually support scale.

In this episode, Marcel shares what breaks as agencies grow, and how small optimizations can make the difference between momentum and mayhem. If your team’s growing fast, but your numbers aren’t adding up, this one’s for you.


Here’s what we dive into:

  • The critical headcount thresholds where agency ops tend to break - and what to do at each stage
  • The metrics that really matter when you’re trying to scale profitably
  • How to build a top-down forecasting model that’s simple, fast, and actually useful
  • How to align your leadership team around a shared, consistent view of performance
  • Why data hygiene matters, and how to balance accuracy vs precision to help you make smarter, faster decisions

Marcel also breaks down the operational traps that create noise, slow growth, and kill clarity. Plus what you can do right now to lead with better insight.


Additional Resources:

Follow Marcel on LinkedIn: https://www.linkedin.com/in/marcelpetitpas/

Follow Harv on LinkedIn: https://www.linkedin.com/in/harvnagra/

Parakeeto’s Website: https://parakeeto.com/

Marcel’s Agency Profitability Toolkit: https://parakeeto.com/toolkit/?utm_source=Earned+Media&utm_medium=Podcast+Appearance&utm_campaign=Harv+Nagra 

Parakeeto's Foundation Course: https://course.parakeeto.com/?utm_source=Earned+Media&utm_medium=Podcast+Appearance&utm_campaign=Harv+Nagra

The Agency Profit Podcast: https://agencyprofitpodcast.simplecast.com/ 

Stay up to date with regular ops insights. Subscribe to The Handbook: The Operations Newsletter: https://scoro.com/podcast#handbook


This podcast is brought to you by Scoro, where you can manage your projects, resources and finances in a single system.

Marcel Petitpas:

I can't tell you the number of times that we're like meeting with teams. We're starting in a discussion and we're talking about, let's call it utilization rate. We'll just pick a random metric, but this happens with all metrics and we start talking about utilization there's a moment where I have to be like, okay, pause. Project manager define utilization for me, and they're like, oh, you know, billable hours over capacity. I'm like, great. What is someone's capacity? Exactly? What does it

Harv Nagra:

include what does it not include? They give me an answer. Then I ask the CFO

Marcel Petitpas:

what's your definition of capacity? Oh, it's different. Fascinating. Then I ask the CEO, what's your definition of capacity? Oh, it's also different. Fascinating. Okay, so we're all talking about a metric that we think we're speaking the same language. We don't understand each other.

Harv Nagra:

​This podcast is brought to you by Scoro. Scoro is an agency platform that brings together your quoting, project tracking, invoicing, and agency reporting into one place. Imagine building your quote in the same place where you're tracking your budget, including time, expenses, and supplier purchases. Sending invoices that let you see how much you have left to bill, and knowing when you get paid, and being able to monitor your project margins. That's next level automation, and you don't get that in spreadsheets. Bonus, it's got a beautiful user interface. That's why I brought Scoro into my past agency. Sign up for a free trial at Scoro.Com or if you arrange a demo call, tell them Harv sent you. Now, back to the episode. Hey all. Welcome back to the podcast. Whether you're running a 20 person agency, you're creeping past 50 or surpassing a hundred, at some point you might find that your ops stop keeping up. Suddenly, you can't see the numbers clearly, forecasts start falling apart. People are busy, but profit is flatlining. Your management team is frustrated at juggling too much. And the systems that got you here, they're holding you back from what's next? It happens to the best of us. One of my old bosses, if he's listening, will remember a picture I showed him of a leaky pipe held together with duct tape and about to burst. I was trying to make a point. So in this episode of the podcast, we're breaking down the growing pains that hit agencies at different thresholds and how to overcome them before they become blockers. We get into what actually needs to change, whether it's your data practices, operating model, or how you make decisions and we bust a few myths along the way. So if you've ever felt like your agency's ops are built on duct tape and good intentions. This one's for you And joining me for this conversation is someone who probably knows agency numbers better than most of us know our own bank balance. Can you guess? It's Marcel Petitpas, CEO, and co-founder of Parakeeto. Marcel is the go-to expert when it comes to helping agencies improve their profitability, pricing, forecasting and financial systems. He's worked with hundreds of agencies all over the world, from boutique shops to global teams, and he's got a knack for making complex financial concepts feel surprisingly clear, he's also the host of the Agency Profit Podcast, a killer resource in its own right. Let's get into it. marcel, welcome to the podcast. Thank you so much for being here. You're the first fellow Canadian I've had on the podcast, so yay us.

Marcel Petitpas:

What a time to be coming together like this.

Harv Nagra:

Exactly, exactly. Commiserating. Um, but we'll, we'll do that off the record. So Marcel, you talked to clients all over the world. are there any trends or shifts you're noticing compared to the past couple of years?

Marcel Petitpas:

we're recording this podcast on April 3rd,

Harv Nagra:

mm-hmm.

Marcel Petitpas:

obviously a lot of stuff changing in the world. And of course all of this is affecting agencies, but I think if we crop out from this moment in time, that feels very acute, there have been a lot of broad sweeping changes from my perspective that have been taking place over a very long time, at the most macro level, the industry has been maturing in the same way that every industry matures in that margins have just been getting worse over time. Competition has been increasing as all of the little barriers to becoming an agency have gone down. opportunities to do digital things have proliferated, and the cost of getting into those digital things and educating oneself on those digital things has decreased, what we found, especially in the last five years or so, is that there have been some major accelerants to that margin pressure. So

Harv Nagra:

Hmm.

Marcel Petitpas:

a lot of people are realizing that it's happening it's happened a lot faster in these last few years. And those two things have been the pandemic, which I think leapfrogged remote work and the

Harv Nagra:

Yeah.

Marcel Petitpas:

of competition by like a decade, almost overnight, where

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

people were like, wait, why are we hiring agencies locally? When our team isn't even local anymore, and now all of a sudden it's like, oh wait, we can get really high quality work from geographies that have a massive cost advantage. And so that really accelerated price pressure in the agency space while at the same time you might recall during that pandemic, the cost of onshore labor went up. to 40% across most roles inside of an agency, especially specialized in technical ones,

Harv Nagra:

Hmm.

Marcel Petitpas:

engineers, designers, project managers, account managers, like there were salary reports that got published that showed massive increases in staffing costs throughout that time. So you were getting squeezed from both directions through the pandemic. And then just as everyone's recovering from that tidal wave. AI shows up and enters the

Harv Nagra:

Yeah,

Marcel Petitpas:

and of course we

Harv Nagra:

I.

Marcel Petitpas:

still haven't really seen the full breadth of the disruption, but I think we have seen some impacts and the the worst is yet to come and the best is yet to come. You you could be on either side of that equation.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

forces, I think, have really recently put a lot of focus on much tougher it's becoming to run a profitable agency. How much less you can get away with not actually being sophisticated about your operations, having your data together. of course with the economy having been challenging, let's put it that way, the last few months, that has also, I think, increased the focus on this. And so the short summation of this is, I don't think anything has changed, but I think that recent events have pushed things to go a lot faster in a shorter period of time, and therefore people are more aware of it.

Harv Nagra:

Marcel, remind me that salary increases that you're referring to. Was that due to that"great resignation" thing that was happening where people were moving around a lot and getting quite, pushy with their expectations?

Marcel Petitpas:

Yeah, so I, I'm not gonna pretend to have a full kind of understanding of the macroeconomic factors behind this, but my high level understanding of why that happened was twofold. Number one, there was not a very strong incentive to work during Covid because there was so much relief fund available, right?

Harv Nagra:

Right.

Marcel Petitpas:

of that meant that there wasn't a lot of pressure for people to work. There was this moments in time where the the power dynamic in the labor market, especially the skilled labor market, shifted in a big way towards The employees. We did see

Harv Nagra:

Absolutely.

Marcel Petitpas:

resignation, people

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

The push to remote work, I think really opened up a lot of opportunities for people to

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

in different geographies. And so, Yeah, all of those things really contributed to this pretty steep hike in the cost of talent, especially as the demand for digital talent went through the roof in that moment. cause you

Harv Nagra:

Right.

Marcel Petitpas:

there was. Massive industries that were way behind on digitization that all of a sudden needed to catch up. And so the demand for agency services went through the roof. I know a lot of e-commerce agencies that did very, very well. Those two years had multi

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

percent year over year growth rates. So they're trying to hire talent. big organizations are trying to hire that talent in-house. So those things all coalesced in a perfect storm and that shot the kind of leverage for people that had digital skills very, very high in that time period.

Harv Nagra:

you also work with people in all different markets agencies in the uk, Canada, and the us, Australia and New Zealand, can you share any observations about differences that you might see in these markets?

Marcel Petitpas:

be candid, the, the differences are very minor. the UK, I would say is suffering a little bit worse than everybody else, mostly because of how devalued the pound has become over

Harv Nagra:

Hmm.

Marcel Petitpas:

few years. I. has kind of been an additional blow on top of the broad sweeping economic factors that I think are affecting everyone. Mm-hmm. major one was the bottom falling out of the tech industry in late 2023 or late 2024, where you might recall like Facebook's valuation, like all these big tech companies, the MEG seven, their valuations got cut. of a sudden, investors were using the word profit, which they'd never really done before. IPO market dried up. M&A started to dry up under the, the previous administration where, they were basically shutting down any major acquisition. And so the flow of venture capital into the tech industry really stopped,

Harv Nagra:

Hmm.

Marcel Petitpas:

is what has rebounded that. But we, we still haven't really gotten back to m and a. We still haven't really gotten back to IPO, so there's still a bit of a, a, tenuous environment there. That was one big blow. And then the second thing that happened was a lot of the economic uncertainty that came with this most recent election and that we're still kind of dealing with now, where there's essentially a global trade war going on. Nobody really knows what's gonna happen next, and therefore we're not really in decision making mode. So the thing we're hearing across every place in the world is no one's making decisions because no one knows

Harv Nagra:

Hmm.

Marcel Petitpas:

next. Therefore, we have stalled projects, we have delayed projects, we have contracts out that haven't been signed,

Harv Nagra:

Yeah.

Marcel Petitpas:

sales cycle has basically slowed to a crawl.

Harv Nagra:

Mm-hmm. Yeah, good point. Because I've been reading a couple of LinkedIn posts that, you know, don't worry, service-based businesses are exempt from tariffs and all that kind of stuff. But,

Marcel Petitpas:

Yeah.

Harv Nagra:

our customers may not be. So before we get into today's discussion, bringing it back to kind of agencies and profitability, can you think of any mistakes agencies make, of any size when thinking about their kind of finops?

Marcel Petitpas:

I can think of a lot

Harv Nagra:

Okay.

Marcel Petitpas:

that they make when they think about their finops, I'll summarize the biggest one

Harv Nagra:

Mm.

Marcel Petitpas:

thinking of finops in, I'm gonna call it the old school. Way, which is this linear way of thinking,

Harv Nagra:

Okay.

Marcel Petitpas:

we have to kind of go back to time and materials era in services, which was not that long ago, couple of decades ago, when the defacto business model was time and materials billing. You get a client, you give them an hourly rate. You work that number of hours, then finance reconciles all of that.

Harv Nagra:

I.

Marcel Petitpas:

Like that billing model is inherently retroactive. We don't know how much we're charging the client until we've collected the time sheets. Stacked it up, billed the client, right? So made sense at that time for finance to be seen as sort of the the primary owner of measuring the performance and profitability of the business because the finance workflow, which is inherently retroactive and is about reconciling data from the past. Was well suited to that function, and so you really kind of had this like sales went out and sold things. Delivery did things, and. things flow from sales to delivery and then to finance

Harv Nagra:

Yeah.

Marcel Petitpas:

But today, with the proliferation of alternate billing models, very

Harv Nagra:

few

Marcel Petitpas:

firms are exclusively billing on time and materials. Most of'em have multiple different billing models. There's also been a proliferation of different staffing models, so most firms have full-time employees, part-time employees, contractors, white

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

people that are somewhere in between all of those things, right?

Harv Nagra:

Yeah.

Marcel Petitpas:

of those complexities have made it such that. You don't have such a linear flow of information and timeliness is becoming more important and the complexity of operations has gotten much higher. And so finance really should no longer be seen as the place that we push all of this information and then have them try to figure out what's going on. And I think

Harv Nagra:

Yeah.

Marcel Petitpas:

people are recognizing that. This is not really a fair thing to be asking them to do because it requires so much context and it requires so much complexity to really interconnect what's happening in sales, what's happening in operations to what's happening in finance, and get a real picture of how the business is performing and how changes to the business affect the rest of those things.

Harv Nagra:

Mm-hmm. Absolutely. And I think it's just so important that that kind of education trickles down to literally everyone, reminding themselves that we're here to make money and here are the ways we need to do that, and the rules we need to play by in order to make sure that we're, profitable, right?

Marcel Petitpas:

Yeah. Yeah, and I think that the biggest change there that has really increased the level of ownership that I think business development and operations need to start taking over This is that it's no longer as simple as an hour multiplied by a rate equals revenue

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

an hour multiplied by cost equals cost,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

things have become abstracted from each other, and so there's a lot more nuance to figuring out like how are things actually going and how

Harv Nagra:

Right.

Marcel Petitpas:

is our business relative to how healthy it could be. Theoretically.

Harv Nagra:

Yeah, I wonder if you can generalize a bit that when you start working with a new agency, can you make any kind of broad observations about where financial maturity tends to be?

Marcel Petitpas:

yeah, so to your point, the level of financial maturity is quite broad. I'm gonna abstract that even and say profitability management is quite broad.'cause you could focus on finance. I've run into lots of firms that have, I. Incredible finance departments and they still have no idea really what's going on from like a profitability management perspective because finance is only one small piece of

Harv Nagra:

that puzzle

Marcel Petitpas:

And I would argue the least important piece of that puzzle, far less important than getting a grip on operations data. that varies a lot. I spent a lot of time in the software industry. My co-founder, Ben, he spent a lot of time in the home services industry. And you might find this kind of alarming, but I would argue that the average level of sophistication for like a plumbing or landscaping company when it comes to this stuff.

Harv Nagra:

is actually

Marcel Petitpas:

quite a bit higher on average for the same size of company than what we've seen in the agency industry. The margins are a lot worse in home services than they have been in agency. But now

Harv Nagra:

Hmm.

Marcel Petitpas:

to get to that same place where it's like, oh, we really need to take this seriously, otherwise gonna start failing as a business. We're gonna start losing money, and it's becoming increasingly challenging. So I think agencies are behind a, because they didn't really need to pay attention to this, and B, because a lot of creatives that start firms, this is not really what they're interested in. And so they're gonna put

Harv Nagra:

Not at all.

Marcel Petitpas:

long as they can. Right. As long as the money's

Harv Nagra:

Yeah.

Marcel Petitpas:

they don't have to deal with this stuff, then that's generally gonna be their tendency.

Harv Nagra:

that's probably the trigger that gets'em to kind of start working with yourselves and stuff like that is when things get a bit scary maybe.

Marcel Petitpas:

That, or they wanna sell, and then all of a

Harv Nagra:

yeah.

Marcel Petitpas:

start doing the math on like, oh, every additional dollar of profit I make, I can like 10 x that when I sell. Well, okay,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

that's a little more interesting.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

to actually rewarded for that discipline.

Harv Nagra:

Right. So. I talk about business maturity here a lot, and that's not necessarily tied to headcount, but I'm wondering if we were to look at headcounts, Marcel, what thresholds do you notice? Agencies start to outgrow their systems and processes and need to level up.

Marcel Petitpas:

the simple answer is every time a layer of management is getting installed in the agency, that tends to correlate to one of these sort of"ceilings", and

Harv Nagra:

there's a lot

Marcel Petitpas:

of vectors to those ceilings in terms of what makes them challenging. Some of them are tactical, some of them are strategic, and some of them are psychological and emotional, and they have to do directly with the founder's ability to grow to that next level of leadership in the organization. So

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

one. It It depends on how many founders there are, but let's assume that it's like a, a one founder company. The first one's around 10-ish give or take a handful of employees. So you're getting to that like maximum number of direct reports and usually at that stage, the founder's having to install that first layer between them and the client work. So you need some level of financial acumen at that point in time to know what's going on because you're, you no longer have your finger on the pulse directly with every client. the next major one, this is the one that we deal with a lot at Parakeeto, is that first real level of middle management, which tends to

Harv Nagra:

Hmm.

Marcel Petitpas:

around 30 ish, give or take, let's say five to seven employees where you're, you're finally having like. or an executive suite, and then a layer of people that are kind of managing each department in the organization, you probably have like a person overseeing delivery, a person overseeing sales, a person overseeing operations, and maybe the CEO's playing one of those roles. But you really kind of have a sophisticated layer at the middle, and that is usually where. You really start to feel the pain of not having any way to measure how everyone's doing and hold them accountable. And I think more importantly than holding them accountable, empower them assess their own performance and to be able to make decisions and make judgment calls based on objective information that everybody agrees on. And that last part that everybody agrees on is nuanced, but should not be. cause it tends to be one of the biggest challenges. then similarly around 60 ish employees, give or take 10, that's when you're installing a C-suite. So you, at this scale of business, typically you have three levels of management. You have your executive level management, you can have your middle level management, and then you might have some kind of leadership at the delivery level of the organization. And

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

there needs to be. Much more structure, much more clarity to how not only we think about measuring the business, but now we have to get into the systems measuring at the different levels of the business. And there's so many traps at that point that have to do with actually letting go of trying to connect every little thing that happens at the bottom of the organization to the

Harv Nagra:

Hmm.

Marcel Petitpas:

at things at the top of the organization.'cause those lenses are often quite different and require very different structures of data.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

where we start to get into a lot of sort of architectural systems level challenges that also require a change of thinking, in terms of philosophy around managing data. So that was a long answer, but 30, 50 is, is roughly where we see those major roadblocks.

Harv Nagra:

That resonates with me as well, Sometimes I think it tends to be quite reactive. These kind of challenges are being addressed when It's already started to cause some issues. Right. So, hopefully we're giving some pointers to anybody coming up to those thresholds today. I wonder If we can go through each of those in a little bit more detail. Marcel. I think you were saying 10 to 30 size initially. What are the things that agencies of this size might be experiencing and facing, and what are some of the problems or pitfalls that you tend to see?

Marcel Petitpas:

so agencies need to understand that it's alarming how often they don't understand this is their business model is capable of. And so I ask this question all the time to agency owners, which is, if everything went perfectly. How profitable would your agency be? How much revenue would you make? What would you spend on delivery on overhead? How much net profit would the business generate? And

Harv Nagra:

they often

Marcel Petitpas:

don't have an answer to that what they respond with is their goal. And I'm like, okay, well I can appreciate that that's your goal, but mathematically, what is the limit of what your current business model is designed to do? And it kind of catches them by surprise.

Harv Nagra:

Hmm. This

Marcel Petitpas:

is actually just a math formula. You have a certain number of people on your team. They get paid a certain amount of money, they work a certain number of hours per year. There's a

Harv Nagra:

certain number

Marcel Petitpas:

of those that could possibly be used to earn revenue doing client work. You have a way of pricing that sets a target for what you expect to earn from that, that formula. It's pretty straightforward. you how much money you can make, and then you should theoretically have a model for what do you expect to spend on things overhead, your office, you know, sales and marketing, et cetera. So

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

be an understanding of there is a best possible outcome, the way that this business is currently structured, what is that? and that container is so important because without it, do you interpret any information, any actual measurement of data? Oh, our utilization rate is 52%. Is that good or is it bad?

Harv Nagra:

Yeah.

Marcel Petitpas:

answer the question unless you know what it theoretically could be. Our average billable rate on this project was$178. Is that good or is it bad? Well, you can't answer that question unless you know what you expected it to be in the first place. And so the first fundamental thing is, do they understand their model? is that model capable of accomplishing their goals as an organization? And is it based on assumptions that are realistic? that's kind of the first major thing. And then it ties into pricing, which is another area that I am. I'm alarmed at how few firms actually understand how profitable they expect to be when they've sold something. They know what their rate is. They often don't know why their rate is what it is. They

Harv Nagra:

know how

Marcel Petitpas:

many hours they think it's gonna take. They multiply it by the rate. But when I'm like, okay, but what's the margin that you expect to make on that? And how do you know that that's the right amount of margin to support the rest of the business? Very few firms have an answer. So those would be the first two major components is make sure you understand your model. and

Harv Nagra:

make sure when

Marcel Petitpas:

you sell things, you have an understanding of how much money you expect to make. we can kind of start to get into more advanced things like forecasting and some basic feedback loops to make sure that we're on track.

Harv Nagra:

so at that size, Marcel, what do you think are the most important things to address that can have the biggest impact?

Marcel Petitpas:

Yeah, so assuming that you have a basic understanding of your model and you're pricing things intentionally, right, so you

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

you're selling things in such a way that you're setting the business up for success. The the next three challenges that we tend to see people face in that journey from let's say 10 to 30 plus, is number one. are having to step back from the day to day, so they're no longer gonna be in, be in touch with what's going on in the business. They need to now empower other people to be accountable to and measure results. And the timing of hiring and letting people go and, and really managing staff relative to demand,

Harv Nagra:

Yeah.

Marcel Petitpas:

significantly more important. So the ability to forecast and in particular forecast at the executive level, which I think is, is very different than forecasting at the delivery level. And we'll, we'll talk about that. So. On that point, the first major piece at this point that we like to install or that we find is very helpful is something I call top down forecasting.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

top down forecasting is really the idea that when you're at the executive level in an agency, the conversation at that level around staffing tends to look a lot like, what if this, that, or the other thing, right? What if we close these four clients that are on in the pipeline? What if we don't, what if they

Harv Nagra:

Yeah.

Marcel Petitpas:

month instead of this month? What if this person who seems a little disgruntled quits, what if we were to shift this person's role and put them over here instead of over here? what, if any, combination of those things were to happen? And so it's, it tends to be a very fluid conversation where we're in evaluating a whole bunch of different potential outcomes and then making decisions based on our analysis of all of those potential outcomes, because we very rarely have perfect information. the system for forecasting in order to facilitate that kind of conversation and decision making needs to be really quick, needs to be really fluid, and often needs to be abstracted away from detail, the big thing and the problem that I see most agencies face is the way that they're forecasting is by trying to take what their project managers are doing

Harv Nagra:

mm-hmm.

Marcel Petitpas:

to decide, you know, who's working on what task and using that to have this very fluid conversation. And it does not work. It doesn't

Harv Nagra:

Hmm

Marcel Petitpas:

because you can't go and update a thousand task assignments every time you wanna run a different scenario because it takes. In some cases, hours to do that.

Harv Nagra:

mm-hmm.

Marcel Petitpas:

majority of the time when you get to the meeting, something's going to be out of date because there's too much change happening at that bottom level. And so it's not to say that that bottom up resource planning that project managers do isn't useful. It's extremely useful for the job that they're doing,

Harv Nagra:

Right.

Marcel Petitpas:

very useful for the job the executive team is trying to do, which is over broad time horizons. Highly variable sets of uncertain data,

Harv Nagra:

Hmm.

Marcel Petitpas:

bode well for a system like bottom up resource planning, which is kind of inherently based on the assumption that we have a short time horizon. We have a high degree of certainty with the

Harv Nagra:

Yeah.

Marcel Petitpas:

that's the first thing is, is installing a top-down forecasting system that is radically simplified, and

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

kind of mental shifts that we have to work through with clients is The fact that this is really simple and lacks detail is the feature. Not the bug and the

Harv Nagra:

Hmm.

Marcel Petitpas:

to a degree from what your project manager doing again, is the feature, not the bug. So it doesn't rely on an unrealistic expectation, which is their forecast will be super duper up to date quick to update at any given moment in time.'cause that's just not really what that system is designed to do.

Harv Nagra:

correct me if I'm wrong and I am probably wrong, but the way your kind of approach to this works is based on trends with past projects of a similar size and scope. Is that kind of the idea?

Marcel Petitpas:

So There is a point, there is a level of sophistication where that can start to happen. We, that typically is something that we'll work on later. So, and when we talk about this next stage of growth, the kind of 50 plus,

Harv Nagra:

Okay.

Marcel Petitpas:

scale you have enough maturity and enough data to be able to start using Yeah. Time tracking data that's

Harv Nagra:

Okay.

Marcel Petitpas:

well structured

Harv Nagra:

Hmm.

Marcel Petitpas:

your own algorithms to say, oh, we're selling a website.

Harv Nagra:

Hmm.

Marcel Petitpas:

is roughly$400,000. Well, if you have enough data, you can literally just build linear algebra that says, okay, well that's roughly this many hours of design, this many hours of engineering, this many hours of project management, because that's what

Harv Nagra:

Yeah.

Marcel Petitpas:

in the past. then,, dial that up or down a little bit. so you can start to, with well structured data, do that algorithmically in the early

Harv Nagra:

Okay.

Marcel Petitpas:

It's really as simple as saying, let's take the scope of work that sales put together and just simplify it down to. Three to six buckets, how many

Harv Nagra:

Hmm.

Marcel Petitpas:

hours, engineering hours, project management hours and strategy hours is this going to take? Then let's also resource plan like bucket all of our team members into one of those three to six buckets and

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

the same. And so now we have this really simple system where everything that sales produces. It's going to eventually get broken down into a gazillion tasks. Don't worry, PMs like you still get to do that

Harv Nagra:

Yeah.

Marcel Petitpas:

right? Great.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

it all has to ladder up to this much simpler data schema, which is this abstraction layer of what we call these role categories. What role category. Is this time estimated to. And then similarly, all of our resourcing, we can break it down and assign individuals to tasks, but everybody has to ladder up to, again, one of these role categories. So we have this much simpler space where if we wanna update a project and say, oh, what if this project closed next month and the scope increased by 20%? Well we can update a date field and five estimates of hours.

Harv Nagra:

Hmm.

Marcel Petitpas:

of updating a date field and an estimate of hours for 300 tasks,

Harv Nagra:

Yeah. Yeah,

Marcel Petitpas:

the

Harv Nagra:

yeah.

Marcel Petitpas:

that, that, and you multiply that across 50 projects you have a, an exponential decrease in the cost and complexity of maintaining a, not precise, but an accurate forecast for the purpose, which is trying to navigate and simulate outcomes with a lot of uncertainty that is

Harv Nagra:

Right.

Marcel Petitpas:

those outcomes.

Harv Nagra:

I was talking to, somebody a couple of weeks ago that was trying this approach. I think people have that tendency to want to go bottom up when they just don't know any better and, they were struggling with this, saying it's not super precise. and the point was that it doesn't need to be that precise at this stage and the approximation is better than not having it at all.

Marcel Petitpas:

A hundred percent. This is kind of the fundamental idea behind not just this, but a lot of the other ways in which we approach things, which is that

Harv Nagra:

hmm.

Marcel Petitpas:

and accuracy are not only not the same thing, they're often inversely correlated with one another. Right?

Harv Nagra:

Hmm.

Marcel Petitpas:

so a good example of that is you could try to get to this high level insight of how does our design demand change over the next six months if we do or don't get these, you know, projects. if. There are 5,000 task assignments that make up that insight, and each of those task assignments has a 1% chance of being out of date. Well, there's a 5000% chance that that system is fucked at any given moment,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

my French, right. that is a perfect example of the precision of trying to build that castle out of granules of sand

Harv Nagra:

Hmm.

Marcel Petitpas:

exactly the reason that you'll never have an accurate output. the precision is the problem. And so the goal there is to say, okay, well let's round the edges, let's broaden these time buckets. They won't be as precise, but for the type of insight that we're trying to look at, which is inherently uncertain, we have a much more accurate lens on that because the context is different than, you know, what our project manager's often trying to do, which is like, what does next week look like for Harv? Are we gonna be overworked or not? You know, are we

Harv Nagra:

Yeah.

Marcel Petitpas:

resources to do the design thing? That's such a different question and a different context, and that's why the tool should be so different for that.

Harv Nagra:

Really, really good point. Let's jump forward to that kind of next threshold, which was I think 30 plus is what you were saying. What kinds of things do agencies of this size, experience, and face, and what are some of the problems or pitfalls that you kind of see?

Marcel Petitpas:

this is where we start to get into installing, real tight framework within the organization. There's two big problems that I see when you get past 30, you're, you're starting to approach C-suite land,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

gonna have C-level executives. So problem number one is there isn't a common language within the organization. So I can't tell you the number of times that we're like meeting with teams. We're starting in a discussion and we're talking about, let's call it utilization rate. We'll just pick a random metric, but this happens with all metrics and we start talking about utilization there's a moment where I have to be like, okay, pause. Project manager define utilization for me, and they're like, oh, you know, billable hours over capacity. I'm like, great. What is someone's capacity? Exactly? What does it

Harv Nagra:

include what does it not include? They give me an answer. Then I ask the CFO

Marcel Petitpas:

what's your definition of capacity? Oh, it's different. Fascinating. Then I ask the CEO, what's your definition of capacity? Oh, it's also different. Fascinating. Okay, so we're all talking about a metric that we think we're speaking the same language. We don't understand each other.

Harv Nagra:

Right.

Marcel Petitpas:

problem number one is you often have. Just you don't have a common language across the organization, and that is extremely problematic because people are interpreting things differently. You end up with these very material skews. You have these very different expectations. The CEO thinks we should be at 95% utilization, whereas the CFO thinks we should be at 60%, but that's not because they have difference of opinion. They're just calculating it differently. Those implicit problems need to become explicit. And we do that by saying, let's decide on a framework. What are the

Harv Nagra:

Hmm

Marcel Petitpas:

track? What are we going to call them,

Harv Nagra:

mm-hmm.

Marcel Petitpas:

how are they going to be defined, and exactly

Harv Nagra:

Hmm.

Marcel Petitpas:

are they going to be measured? And all the data's gonna be collected. So when we sit down to have a conversation, it's productive because we're all speaking the same language. that's problem number one. because so much of the importance when you get to that level is you have these layers of management. You have to work through people. And so you need some rails for that communication to live on. And when it comes

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

that is like how we define the actual, metrics. And the other piece of that is I. How we understand the relationship between the metrics and the business, right? So if this number goes up or down, what does that actually mean? Why

Harv Nagra:

Hmm.

Marcel Petitpas:

happen? How do

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

influence it? And of course, the way we define the nuances of someone's capacity or billable hour changes the answers to those questions, right? So it's so important to understand and agree on those things. So that to

Harv Nagra:

Hmm.

Marcel Petitpas:

the first problem and the first major thing to make sure is really crystal clear.

Harv Nagra:

It sounds so obvious when you're saying it, but it's so true that we all just go through the motions of kind of growing up through agencies and learning these terms. And everybody has a slightly different definition and every agency has a slightly different approach. What kinds of things do you think beyond the language and the stuff that you just spoke about, should agencies be prioritizing at this size that could have the biggest impact?

Marcel Petitpas:

so once they agree on the language and stuff, and, the last thing I'll say on this is I think one of the reasons that this happens is generally at that scale, you start bringing experts in that have a lot of experience from other firms also, they were an MD at another firm. You hire them and then everybody wants to bring their institutional knowledge to the organization. Oh, well we used to do it this way at my firm but the CFO did it different way at their, their firm. And that's

Harv Nagra:

Yeah.

Marcel Petitpas:

But like, we gotta get on the same page. We can't have everybody bringing their own playbook, from there, once we've agreed on the framework, then I think this is where we have to be really considerate of system architecture. And there's some

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

kind of ideas that are important. Precision versus accuracy is the first the executive level visibility into the business is often gonna be very, very different than what middle managers and kind of direct delivery managers are paying attention to and how they're thinking. And so the next step is saying, what are the questions we need to answer or the decisions that we need to make?

Harv Nagra:

Yeah.

Marcel Petitpas:

do we need to make them? that should start to inform what does the report look like for the different departments at the executive level, at the middle level, what's the time increment that we need to be able to calculate things at? And what's the level of sort of fidelity of the data and where is it going to come from? And what I found is that absent that discussion, what you end up with often is a system where we fall into a whole bunch of precision traps. So I'll give you a few examples of this. somebody on the executive team is like, oh, we should be measuring this on a daily basis. But they have that meeting twice a month,

Harv Nagra:

Right.

Marcel Petitpas:

right? So you spend 10 times more on the operational execution, the tooling, like it's so much more complex and expensive to measure that metric on a daily basis. It's all waste, it's all complete waste, and in most cases it's not even feasible because let's say like you're having to do something based off of time tracking data. Well, everybody on the team's not gonna put their time in every single day If you build a system and a report that relies on that kind of behavior, you are setting that

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

and system up for failure because that behavior is gonna be almost impossible to police and actually

Harv Nagra:

Yeah.

Marcel Petitpas:

to do on a consistent basis. So it really starts to become a question of saying. What is the actual requirement for this dataset? Then what would need to be true for us to meet that requirement, and is it operationally feasible? And really trying to avoid the trap of building systems that are based on assumptions that are never true. So

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

we're going to take data straight from time tracking and put it into a report.

Harv Nagra:

Yeah.

Marcel Petitpas:

that does that, in my opinion, is at risk because people make mistakes in their time tracking data.

Harv Nagra:

Mm.

Marcel Petitpas:

you have a middle step where somebody's reviewing that, cleaning up the data, making sure it's accurate, you're

Harv Nagra:

Yeah.

Marcel Petitpas:

end up with reports that you learn to not trust. Because it's constantly full of the 99 hour time entry that Johnny logged while he was on vacation and left his

Harv Nagra:

Yeah,

Marcel Petitpas:

running

Harv Nagra:

exactly.

Marcel Petitpas:

the 17 variations of the client's name that are being misspelled all over the place, the engineer that accidentally logged time to strategy, right? Because I don't know some, something was going on, so I. that's the other big thing is you really have to start to be considerate about how are we architecting our data systems and start to become intentional about it and really start adopting what I would call data operations,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

of the, the formal language for this job that is common in a lot of other industries, but we're just learning about it in the agency industry, which is the team that is responsible for managing data in the organization and creating accurate reports for everybody in the organization, all the different stakeholders that

Harv Nagra:

Mm-hmm a role we had in my past agency was project management officer. this woman was absolutely brilliant, but she had a monthly process where she would dip into each account manager's projects and just do a spot check with, several things that she would check and correct if things were off. So I think a really good point that's important. I wanna bring it back to that kinda top down forecasting. So you were alluding to like how those agencies that were 10 to 30, and now we were talking about the 30 to 50. how does it differ?

Marcel Petitpas:

So this kind of gets us into a little bit more tactically when you're going from like 30 to 50, the feedback loops become really important. from, the 10 to 30, we talked about the model, we talked about pricing, we talked about forecasting. All of that is really about having an understanding. And a way to structure data around our assumptions,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

being really explicit about these are the assumptions that we make about our people, how much they get paid, how much time they work, when we sell a project, how much time we're gonna invest, how much money we're going to make, and having a way to project those assumptions that into the future. And basically just validate like, are we playing to win here?

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

decisions and assumptions that theoretically should lead us to success? When you're below 30, management accounting, and a couple of simple, operational numbers are enough to kind of help you identify if there's a major gap between your expectations and reality. But as you start to scale,

Harv Nagra:

That becomes

Marcel Petitpas:

increasingly important. And so that's, I. Really about like three operational numbers.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

utilization rate. How busy did we

Harv Nagra:

think the

Marcel Petitpas:

team was going to be? How busy were they actually?

Harv Nagra:

Right.

Marcel Petitpas:

how that's defined becomes really important. The second is average billable rate. How efficient did we expect to be at earning revenue? How efficient were we actually?

Harv Nagra:

Mm-hmm. Third

Marcel Petitpas:

is average cost per hour. How much did we expect to spend on labor? How much do we actually spend on labor? And those can all be measured operationally without. The need for finance. And if you have a good grip on those three metrics, you will never be surprised by your P&L again. So I find

Harv Nagra:

Hmm.

Marcel Petitpas:

really important. and those are mostly just about matching up sets of data to time tracking for utilization and average cost per hour. That's people data, who are the people? How much do they get paid and how much time did we expect them to have available and to spend on client work versus what actually happened. then the second for average billable rate is projects. what is the project? How much did we sell it for? How much time did we expect to spend on it versus what actually got spent? And by who? That's really the, the math, right? It's, I'm radically simplifying that because as we just said, the systems to get all this data together and measure it is that's non-trivial, but. are kind of the key feedback loops there. And then at the finance level, it's quite simple. It's like how much money did we collect from clients? How much of that is actually ours? That would be what we call agency gross income. did it cost for us to earn that money? So that's all of the money that got spent on delivery. I. Which is mostly gonna be payroll and softwares and that ultimately gives you what we call delivery margin, what should be called a gross margin. But I don't like to call it gross margin'cause I'm sick of arguing with accountants over

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

issues. That's really important. And is that really above 50%? Are we keeping more than 50% of every dollar that a client. Gives us after we've

Harv Nagra:

Hmm.

Marcel Petitpas:

the work, and then what are we spending on overhead and are those ratios healthy? That's like really kind of the simple set of feedback loops that we want to have

Harv Nagra:

Hmm.

Marcel Petitpas:

when we're doing all the, the first things we talked about, understanding our model and forecasting, we should know where we expect to land, and then we're able to identify where we're actually landing. And then that really allows us to identify what is actually the problem, right? If we're not reaching our potential, what is that related to? And it's pretty much always related to. One of two things. Either our utilization's not as high as we expect it to be,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

our average billable rate is lower than we expected it to be, or a

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

of those two things. And then it becomes very, very clear what we need to focus on. And that at that scale is so important because you have so many people, you've gotta be clear and focused because it

Harv Nagra:

Yeah.

Marcel Petitpas:

time for that initiative to make its way all. The way down to the execution of the business,

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

needs to be fairly consistent for it to actually start to take effect.

Harv Nagra:

Really good advice, Marcel. we're coming towards the end, but quickly time tracking. It sounds like you're an advocate. Tell me good I important or not so important.

Marcel Petitpas:

I'm a huge advocate for time tracking. I'll Go as far as to say that I think it's irresponsible to run a service-based business without it. To me it would be the same as running a restaurant and not really having any idea what your food costs are. the misconception, I think, around time tracking is that that has to be time sheets, I would abstract the definition of time tracking to be having a record or at least some kind of a model of where time is being invested. Because it is your biggest variable cost. You can't understand gross margins, you can't understand client, Contribution margins. you can't really understand anything just your raw P&L without time tracking information. And there are ways to collect and analyze that data that don't involve time sheets. but they're not available to everyone. So, for example, if you are an organization that doesn't spread your team across many projects at a time, most people

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

on one, two, maybe three things at a time. your resource plan, if it's kept up to date by your project management team. It is probably a pretty accurate summation of where your time is going, right? It might not

Harv Nagra:

Absolutely. Mm-hmm.

Marcel Petitpas:

export that to a CSV. What does it look like? Well, it looks a lot like a time sheet project,

Harv Nagra:

person

Marcel Petitpas:

time, date, right? It's all the same metadata. your qualm is with time sheets and you just don't like the experience of filling in time sheets, well, there's ways to create models and records of where time is going to be used for this internal analysis and empowering yourself and your team that

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

you to necessarily be filling in time sheets and the technology for that's getting. Better all the time. We've

Harv Nagra:

Absolutely.

Marcel Petitpas:

tools that help us fill in

Harv Nagra:

Yeah.

Marcel Petitpas:

sheets. We have resource, plan informed, calendar informed, kind of time sheets. So there's a lot of hybrids, there's a lot

Harv Nagra:

Yeah.

Marcel Petitpas:

right, to start to take the friction out. But fundamental belief, Harv, is that if you made a team 100% accountable to the financial outcomes, if they only got paid based on the profit every project that they worked on,

Harv Nagra:

Mm.

Marcel Petitpas:

said to them, you could choose to track time or not. Most of those teams after a long enough period of time, I believe, would organically gravitate towards tracking their time.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

Right. So if you're struggling with this culturally it has to do with how it's being discussed, how it's being

Harv Nagra:

absolutely.

Marcel Petitpas:

it's being enforced. There's a deeper problem because inherently time tracking is only healthy for the organization

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

done. That's my belief.

Harv Nagra:

Absolutely, completely agree with you, couple of examples that you gave that I think are just so effective. The resource plan informed it makes it so easy for somebody that has their schedule planned to just say. Was that accurate or does it need to be adjusted slightly and there's no timers involved. There's no time sheet tables involved. It's just a tweak to the schedule. And I think for those of us that kind of plan our calendars, that's another great way if you can, just tag it to a project or, a task, something like that.

Marcel Petitpas:

The thing I'll say on time tracking is this is another place that I see people fall into a precision trap all the

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

I go in and I'm like, Hey, how's your time tracking? They're like, oh man, our compliance is so bad. You know, people

Harv Nagra:

Hmm.

Marcel Petitpas:

in. And then I go and look and it's like, okay, well I. You're asking your team to track time against the subtask, within the task, within the milestone, within the deliverable, within the phase, within the project. It's too much detail

Harv Nagra:

Yeah.

Marcel Petitpas:

to make 19 decisions just to log a time entry.

Harv Nagra:

Hmm.

Marcel Petitpas:

them like, Hey, what's the most important thing you're trying to measure with this data? They're like, oh, we just wanna know how profitable our clients are. And it's like, okay, great. Right. This comes back to the thing we were discussing earlier, like what's the actual question that's being answered? And then I'm like, oh, well that's great. out of the 18 pieces of information you're asking for, you actually only need like four.

Harv Nagra:

Yeah.

Marcel Petitpas:

we could radically simplify this. And so I, I think part of the problem too that people experience is they make it way too hard for their team to track time.

Harv Nagra:

Mm-hmm.

Marcel Petitpas:

you really double click on it, the vast majority of the metadata that they're going after isn't necessary. They're like tracking it just in case.

Harv Nagra:

Yeah.

Marcel Petitpas:

to tracking on a purpose. So that's another like example of trying to be too precise and it

Harv Nagra:

Yeah.

Marcel Petitpas:

the cost of accuracy because now you don't have enough information to answer the

Harv Nagra:

Hmm.

Marcel Petitpas:

in the first place,

Harv Nagra:

we've covered a lot today. So if an ops person is listening to this and feel like they've got a load to learn or a lot to fix and they want some more advice, where can they reach out to you or your team, and, and get some more information?

Marcel Petitpas:

Yeah. Well the first thing that I would encourage everyone to do, in particular if that segment about the framework really resonated with you of like,

Harv Nagra:

Hmm.

Marcel Petitpas:

know, we don't agree on metrics to track or how to calculate them. We have a

Harv Nagra:

Yeah.

Marcel Petitpas:

called the Agency Profitability Toolkit, and it just basically walks you through our framework at no cost. it's very important for me to have this out in the world at no cost'cause I remember how hard it was to get clear

Harv Nagra:

Yeah.

Marcel Petitpas:

to these questions when I was, in that seat. So you can get that at parakeeto.com/toolkit. And if you wanna connect with me, you can find me on LinkedIn. I'm very active there. I can't help but Chat with people in the dms. I think that's maybe where we ended up, deciding we were gonna do this podcast together. So connect with me there, reach out to me, and if you wanna learn more about our services at Per Keto or just get more free content, like this than head to per keto.com. I also want to give a shout out to the team at S Coro, great people at Koro. So thank you for having me. This is a lot of fun.

Harv Nagra:

Absolutely. Thank you very much. I've just noticed,, you're wearing a hat that says data as well, which is amazing. I love it. Marcel, it's been an absolute pleasure. Thank you so much for coming on today and looking forward to speaking to you again.

Marcel Petitpas:

Thank you Harv.

All right. That's a wrap with Marcel. If your brain is spinning a little, a good opportunity for me to remind you that you can sign up for the Handbook newsletter to get the key takeaways in your inbox. The link is in the episode notes, but here are three things that stuck with me. Number one, most agencies struggle not because they're messy, but because they're too detailed in the wrong places. Precision isn't the goal. Accuracy is number two. Forecasting doesn't need to be perfect. To be powerful. Even a simple top-down model can give you the clarity you need to make smart decisions. And number three, as you grow, alignment matters more than ambition. You can't scale decision making if your team's not speaking the same language when it comes to the numbers. Now we can't resist mentioning that some of the stuff that Marcel was talking about there is doable in Scoro. For example, top-down bookings versus granular, bottom-up resourcing. Next simplified time tracking by just confirming or adjusting what's in the resource plan for the day, or assigning meetings in your calendar to projects or tasks. And finally, financial visibility at a glance so you can see how much capacity you have left next month, or how you're tracking against your target margins. A huge thank you to Marcel for dropping so much wisdom. You can check out his agency Profitability Toolkit on Parakeeto's website. It's a free resource that walks you through the frameworks we talked about today and make sure you give the Agency Profit Podcast a listen as well. We'll put links to both in the episode notes. If there's a question you want me to put to Marcel to share in a follow up, please email me at podcast@score.com. And if you've enjoyed today's episode, please share the episode with someone who would enjoy it. Join the conversation when you see Marcel or I posting about it on LinkedIn. if you've not done so yet, then please rate the podcast on Apple or Spotify. That's it for me for this week. Thanks so much.

People on this episode

Head Shot

Harv Nagra

Host