Optimiser les coûts ne signifie pas réduire les budgets sans réfléchir. Il s’agit d’améliorer la visibilité, la gouvernance et la discipline opérationnelle dans les environnements TI et infonuagiques.
Dans cet épisode de la série Conversations avec CGI, Derek Marinos accueille Mark van Engelen et Scott Stanley pour discuter des pratiques FinOps et de quelles façons elles peuvent renforcer la maturité des organisations dans les opérations infonuagiques en améliorant la gouvernance, les prévisions et la transparence des coûts.
La discussion explique également comment les organisations peuvent passer d’une gestion réactive des coûts infonuagiques à un modèle de gouvernance opérationnelle proactive qui soutient l’évolutivité et l’innovation à long terme.
Bien des organisations ont déjà mis en place des outils infonuagiques et FinOps. Leur plus grand défi est généralement de savoir si elles ont le modèle de gouvernance, la matrice de responsabilité des coûts et les processus opérationnels nécessaires pour utiliser efficacement ces outils.
Les pratiques FinOps donnent de meilleurs résultats lorsque les organisations vont au-delà de la simple plateforme infonuagique pour développer une vue d’ensemble complète des coûts informatiques liés au nuage, aux logiciels services (SaaS), aux licences et aux opérations. Mark souligne ainsi l’importance de s’éloigner des tableaux de bord fragmentés et de la gestion cloisonnée.
Scott Stanley rappelle que, une fois que les organisations ont accès aux bonnes données, la responsabilité des coûts devient plus claire. Grâce à cette meilleure visibilité, les responsables de produit, l’équipe des profits et pertes, les leaders de l’entreprise et les équipes de prestation de services peuvent comprendre les répercussions des décisions sur les dépenses, les prévisions et la durabilité à long terme.
Pendant la discussion, il apparaît clairement que les cibles d’économies en pourcentage ne sont qu’une partie de FinOps. Pour Mark, l’objectif doit être plus vaste : il faut aider les organisations à construire un système répétable qui s’adapte aux technologies, aux plateformes et aux changements de priorités d’entreprise.
Pour Scott, les pratiques FinOps sont liées à des disciplines opérationnelles plus larges comme l’ingénierie de fiabilité de site, l’observabilité et les opérations alimentées par l’IA. Dès lors que les organisations peuvent prédire et éviter des coûts avant qu’ils se concrétisent, FinOps s’intègre à un modèle opérationnel plus mature.
- Chapter 1: Establishing the FinOps foundation
-
Derek Marinos
Welcome to a CGI Conversation, a place where we explore the trends, challenges, and opportunities shaping the future of business and technology. I'm your host, Derek Marinos. IT and cloud cost optimization isn't about slashing budgets. It's about building visibility and discipline into how IT environments are run, including cloud. When that foundation is in place, savings follow. Today, we're unpacking that approach and how one Canadian investment firm applied FinOps principles to regain control, improve forecasting, and significantly reduce cloud spend without slowing delivery. Joining me are Mark van Engelen, who leads emerging technology at CGI in Western Canada, and Scott Stanley, a longtime cloud and FinOps leader. Gentlemen, welcome to the program.
Mark van Engelen and Scott Stanley
Thank you. Great to be here.
Derek Marinos
Let's start with the pattern. And Mark, when you walk into a cloud-heavy organization and assess it through a FinOps lens, what are you really looking for?
Mark van Engelen
I really look for their maturity level, first of all, because... A lot of time around this is spent on kind of governance and process related issues. We find it's not really the tool sets that are the problem. There's many different tools available that can help in the FinOps space. So that's not really the issue. It's really about kind of adoption and implementation of processes around FinOps. That is usually the gap. So I look very quickly beyond the cloud, if they're heavy on cloud and really kind of how is their IT cost governance structured across the board. Do they have a single pane of glass where they can see all of this together? Do they have a definition for total cost of ownership? All those type of things. How is the roles and responsibilities organized? That we find is a key thing. For example, I came into a client where they were into their cloud journey for about three, four years. but they were still very fragmented on how they were organized. We helped implement some of the FinOps processes. We helped configure some of the existing tools. I get them a single pane of glass, and with that, they were able to find about 25% cost savings through that process. But it's really around governance and process and maybe some configuration of the tools, but that's not typically the issue.
Derek Marinos
So this isn't usually reckless cloud usage. It's maturity, like in many respects. Cloud scales quickly. Operating discipline doesn't always scale with it. How common is that pattern across organizations you're seeing?
Mark van Engelen
It's pretty common, even if they have it under control for, let's say, cloud, you can still see gaps or issues maybe with their SaaS or licensing costs or new AI costs that are coming up, if they haven't kind of extrapolated those processes and governance to all those areas, there's always an area to improve, let's say, in that space. So my go-to is thinking about FinOps really at an enterprise level and have it across all systems, make sure your processes are consistently. Otherwise, it looks good on the cloud. You got it under control, but then there's a gap somewhere else. That's appearing because technology is constantly changing, right? And therefore, if you don't have the processes and governance in place to be able to adapt to that, there's always going to be a gap somewhere.
Derek Marinos
Yeah, and once you identify that maturity gap, the next question becomes how you engineer the foundation to close it. And Scott, once you understand that maturity gap, what did you change first or what would you change first?
- Chapter 2: Closing maturity gaps through data and accountability
-
Scott Stanley
Yeah, I mean, so there are... are industry standards that are generally found within a software delivery lifecycle or a full delivery of any IT process. And in turn, a lot of things that we look at is how are you addressing it today? As Mark called it, maturity gaps are different from client to client. We very much focus on what's going to get the most value back to the client. What are the biggest indicators of anti-patterns that we're seeing? In a lot of cases, We try to make sure that we're collecting data and observing that data and providing the right KPIs to the right stakeholders to make sure that we're feeding back at the beginning of the process for P&L owners or product owners so they can address those governance gaps that Mark was describing.
Derek Marinos
That's interesting. And once that visibility was kind of in place, how did it change the way teams forecasted, owned, or acted on the cloud spend?
Scott Stanley
Yeah, you'd be really surprised that, a lot of our clients, they, once they see this data, it becomes very apparent to them and revealing that, they in turn need to take more accountability for this spend, especially when you can identify that teams that are more, but I'll say responsible on the RACI as opposed to accountable, you hand it back to the person who's accountable. They immediately want to kind of start to change this. And again, what you'll see is, you know, the prediction of cost really benefits individuals like P&L owners, right? Because they need to be more accountable for than just the revenue of what they're building, but the overall cost and sustainability of the products that they're developing. So that's where it really starts to have a direct impact on that cloud spend. A lot more guardrails will be put in place and processes will be enhanced to ensure that, again, there is a predictable nature of the cost prior to receiving the bill.
Derek Marinos
and when we opened up the show, we talked about that Canadian firm. And I'd love to circle back to use that example that we're bringing in here. And can you, Scott, just elaborate a little bit more about how we're engineering accountability here?
Scott Stanley
Yeah, so in a lot of cases, we go in, we'll find the biggest opportunity for cost savings in here, right? Once we've identified that, then there's obvious questions around, well, how did we get here? What are those anti-patterns that we've developed? helped over time and a lot of companies struggle with that because they feel they're very mature in their governance process. But when you reveal the data, it's really hard to argue with that data. So we leverage that data to show them that, okay, in the firm that we're specifically talking about in that scenario, there's a very high spend on their AWS costs or their cloud costs specifically. And, you know, we immediately saw, well, if you produce these instances as an example, you can find immediate cost savings. But the challenge there lied with If we make these changes, how does that impact the customer? So in turn, it starts to kick off a whole initiative to make sure that we are understanding, well, if we make this change, what is the real return on investment we're going to see? Do we need to take away from our revenue drivers to achieve this, to focus on legacy cost savings? And what's going to be more valuable? Is it going to be more valuable to do this in year and achieve these savings? Obviously, in my opinion, the best time to start saving is now. But in the grand scheme, you have to move at the pace that the customer can move at. We don't want to get involved in competing priorities and take away from the revenue. We want to try and focus on what is the most optimal move out of all the permutations that's going to get them the most in your savings.
Derek Marinos
And the cost reduction, the value that you're bringing here and you're talking about is really important. Mark, from an executive standpoint, why does control matter more than percentage?
Mark van Engelen
Because from my perspective, percentages change, right, as I was saying earlier. the technology changes over time. So maybe initially you can get a 20% savings on cloud, but then I have clients that switch from their current hyperscaler to another hyperscaler. So that 20% savings you had all of a sudden changes dramatically because you have to start from scratch with somebody else, right, with a different tool set and all those things. So it's really not about... the percentage, and it's around having the system in place that you can capture those changes and adapt accordingly. So it's more about the control of the process, more having the accountability clear, and understanding there's a raft of tools that probably need to work together as well to give you that comprehensive picture. So that's the The main focus is control, but it often starts with a percentage, right? You come in, well, help me save 10% on my cloud cost or 20%. But very quickly, they turn around and they basically say, well, how can you make this sustainable?
Derek Marinos
No, it's true. It's very much true. And savings are measurable. Control is strategic. Let's zoom out for a minute. And Mark, at what point does FinOps stop being about cost, start becoming about operational maturity?
- Chapter 3: Moving from cost savings to operational maturity
-
Mark van Engelen
I think it's as soon as you've produced some initial results is what I find, because then that justifies doing more efforts, right? If they try and go down the path and they can't find any savings, then it's a dead end. There's very little executive leadership sponsorship and all those things, right? So it's very quickly getting some initial savings, but then turning around the conversation to make this kind of enterprise ready, controlled, and across the board, not just for clients. Cloud specifically.
Derek Marinos
What changes structurally, Mark, when this approach truly sticks?
Mark van Engelen
It's more of an integrated mindset, I would say, and integrated processes. People typically have siloed teams. I may have a data analytics team. I've got an on-premise infrastructure networking team. I've got a cloud team. And these FinOps processes kind of bring that all together to see how these processes link together, where the touch points are, and therefore gives that better kind of end-to-end enterprise view of that. So I think that's the big value for them is having that whole end-to-end enterprise view on their costs.
Derek Marinos
And Scott, this is where, site reliability engineering intersects. How do FinOps and SRE reinforce each other?
Scott Stanley
Yeah, so I mean, if you consider that a highly available product means you're going to generate more revenue, site reliability engineering is really about, you know, high availability and resilience. It's about making sure that, you know, you're putting practices in place to be proactive in responses to incidents as opposed to reactive, which therefore means that you're going to have less spend on the human side. You know, if you consider things like a 15-hour outage, And the types of silos of teams that Mark described, often production support is at the end of the SDLC. They're the team that has to go and address all these problems. They have, you know, wake up 2:00 in the morning and address all these. And that overtime cost and those outages have a high impact on your costs. So site liability engineering is really about observability. It's really about capturing, you know, data metrics to help you enhance and improve your delivery process to ensure that you aren't delivering bad code production to ensure that, you know, you have that high availability and resilience. But coincidentally, a lot of these tools that are best-in-class solutions out there are now adding in things like cloud cost management modules into their services. which allows you to have predictive cost analysis on your cost so that as you're building and developing things, before you receive the bill, you can see those costs. So what we're really trying to do is help our clients have cost avoidance and predictive nature of cost together in a single pane of glass. So site reliability engineering definitely lends itself and is very symbiotic in the relationship with FinOps. And once that foundation is in place, I'm curious here, where does AI start to enhance optimization?
Derek Marinos
Absolutely.
Scott Stanley
So, I mean, the dream is that your services will self-heal and will automatically stand themselves back up. And again, a lot of these tools offer those types of AI capabilities so that in the event that it detects an outage, it can actually go in and execute processes to ensure that the services do stand themselves back up without, you know, alerting, you know, obviously it'll create an incident, but without waking up people at 2 A.m. because the issue can actually be resolved. Now, again, this is a very advanced state of reliability engineering and AI ops. It's challenging to achieve, and a lot of our clients aren't close to that maturity level, but that's entirely part of the process of what we do here at CGI. We help our clients continue to evolve their practice, teaching them how to fish and adopt these future-facing approaches.
Derek Marinos
That's interesting. And Mark, the old saying is, if the tools aren't the problem, then what is?
Mark van Engelen
Yeah, we talked a bit about the governance and process, but maybe what I wanted to highlight here is the division, right? So there's, for example, many different measurement frameworks people could be using, anywhere from technology business management to COBIT, FinOps Foundation, if you haven't selected a proper kind of cost measurement framework. then there's your first problem already, right? And the second bit, after you selected it, now it's like, well, how am I going to do this? So what's the process going to look like? What's the governance going to look like? And which mixture of tools do I need to pull that all together to have FinOps in my organization across the enterprise? So I think that often when we come into clients, it's like they haven't really decided on that measurement framework even yet. They're just starting to adopt tools And without that vision. So I think that for me is the first part, having a clear vision of, hey, I want to run this across the enterprise. This is the framework I want to use, which is aligned to the rest of my, how I run my IT. And that's a really good starting point to actually filter out which tools you would need and which processes you need to change and what governance needs to change as well.
- Chapter 4: Breaking down silos and reducing OpEx
-
Derek Marinos
And most organizations already have cloud platforms and observability systems. So Why are they still struggling?
Mark van Engelen
Because that's very silo-based, right? So let's say I have a multi-cloud environment and I've got some AWS and I've got some Azure and some Google, they all have their own cost management tools, but now I'm looking at three different dashboards. then I can't really kind of stitch it together. And how does that relate to a business function? How does that relate to a service I'm providing as an IT organization? So that's typically the challenge is, okay, well I have all these bits and pieces of siloed information, but how does that actually stitch together? And how am I going to govern that? Because fragmented teams, fragmented processes and fragmented tools do not give you that end-to-end view of your enterprise and where you can actually save the most.
Derek Marinos
Yeah. I know, that's true. Scott, a little perspective on that as well.
Scott Stanley
Yeah, I mean, if we see that, a cacophony of logos with our clients as they select different technologies, we often see that product owners have their own, cost center or budget, and their team will advise them that this is the best tool we should use. But then a lot of these tools are redundant. They have overlapping capabilities. The licensing costs continue to expand as opposed to leveraging the opportunities to amalgamate these costs and get different tiered licensing, which will actually give you reduced costs, as well as doing different things like purchasing these through your enterprise cloud agreement, as an example, which gives you different advantages. These are all things that we take into consideration with our clients to help them map out what the best approach would be, and so it's not just purely about cloud utilization. It's about the opportunities to save dollars in all areas, including software licensing and human efforts. Because when it comes down to it, I mean, we're talking about reducing maintenance cost. Because every CIO who's out there, if you can reduce maintenance costs, it frees up budget and important resources to get after R&D, to get after other things that can help.
Derek Marinos
Is that as much a priority? I'll open up to both of you very quickly when going in the door, because quite most of the time you hear some of those maintenance budgets can be 70, 75%. I'm handcuffed, help me here, is often the case, right?
Mark van Engelen
Yeah, I would say. It's really about OpEx reduction, right? Because as people have been switching more to cloud and getting more SaaS products, their OpEx has increased, while from a business lens, probably the CIO hasn't added a lot of services, right? So that's been the main question we've gotten from clients is help me reduce my OpEx, because that then frees up other things. You're right around the maintenance budget though, because like a CIO typically spends like 70% of their budget on like maintenance and technical debt type of activity. activities. But those are typically project costs. But the main focus, I think, around FinOps is getting a handle of your OPEX, because people have just been adding in tools and not having the control around that. And therefore, they get a lot of questions around how do I reduce my OPEX. And these FinOps processes help give that transparency, help give back control where OPEX may have skyrocketed without those controls, proper controls in place. So, this is for me the right area of getting your OPEX under control properly and consistently.
Scott Stanley
and just to color a little bit more on what Mark's saying, it's... the transition into a scalable solution like cloud, right? It's now, it's on demand. The ability to observe these things, it's happening much more rapid fire. And specifically staying on OPEX, there are actually creative ways to amortize in the cloud and actually make it more of a CapEx spend. But that's a very unique situation, right? There's very specific scenarios where that works and where that's beneficial. So really what you need to focus on is How do you truly build the scale? The promise of cloud is that you only pay for what you use. But the truth of the matter is a lot of organizations really have adopted cloud in a very experimental way or not a consistent way. Their teams go, they learn, they stand things up, they structure it. Is it well architected? Is it aligning to best practices? And is it truly scaling properly? Because in a lot of cases, the demand, the customer demand invokes its own anti-patterns such as, we'll see right shifting and Conway law, start to apply within the teams. And that's when the teams start to bloat out and be created based on the needs of the product, as opposed to ensuring that you're doing things correctly and building an intelligent and right-sized product where you can keep your teams lean and efficient and reduce the human cost, which is exactly the intention of what these services are intended to do. But in a lot of cases, people are treating the cloud like a data center. And that's why you start to see objects cost start to skyrocket.
- Chapter 5: Key takeaways for leaders
-
Derek Marinos
That's really interesting. Thank you for your perspective on that. As we come to part of the program. I'd like to get some key takeaways from each of you, rapid fire, and some really good short bites. Mark, I'll start with you. What's the one decision leaders should make this quarter? I think making sure that they are consciously taking control of their costs across the enterprise.
Mark van Engelen
So really looking at it holistically and not just leave it in the silos where it's typically at. Okay, and Scott, what's the one discipline teams tighten immediately? I think it's really accountability and it's ensuring that making sure that each service as you develop it, everyone's aware of what they're accountable for. That's one thing we start to see. When you introduce FinOps into most organizations, what you immediately start to see is people cleaning up the house. Like, yes, they're coming over.
Derek Marinos
Let's go and actually start to decommission those things that we stood up and orphaned and left sitting there for some time, which was just consuming dollars. The notion of FinOps really starts to invoke that sense of accountability within teams. And then in turn, because it's all visible to everyone, helps create that overall cultural shift that you're looking for. And again, starts to immediately reduce spend.
Scott Stanley
IT and cloud spend isn't just a finance problem. It's an engineering signal. FinOps works when it's practical and repeatable. And when it evolves into operational maturity, that's when transformation happens.
Derek Marinos
My thanks to Mark van Engelen and Scott Stanley for joining the conversation and sharing their insights. And thank you for listening to a CGI Conversation, a place where we explore the trends, challenges, and opportunities shaping the future of business and technology. I'm your host, Derek Marinos. Bye for now.
