Ep. 160: Karl Sahyoun, Bregal Milestone | Building Data Infrastructure and Commercial Analytics

0:00Hey everybody, before we get started with today's episode, I want to take an opportunity to tell you about my new book, AI Marketing Blueprint. It just hit number one bestseller status across multiple categories on Amazon, and it walks you through a proven framework that we are taking our clients through here at How to SaaS and how to stay relevant and visible in a world with AI platforms. One of the biggest problems that we have seen across clients and companies is that the traditional channels that they relied on, like paid search or SEO, have declined significantly or are getting more expensive and competitive.

0:33The reason for this is that a lot of buyers are starting to self-educate themselves and research on AI platforms. So a lot of that volume that was previously going to those channels is now moving to these new AI platforms. So to stay relevant, the book walks you through seven new rules that companies in all kinds of verticals and industries need to consider so that they can continue growing their pipeline and revenue and ultimately their enterprise value. So if this applies to your business or one of the companies that you've invested in, I definitely recommend that you pick up a copy. It's available on Amazon, on Apple Books, on Barnes & Noble, and even share it with your teams.

1:10Grab a bunch of copies so that you can kind of consume the content together and start to apply it to your business. And with that said, let's move on to the episode. Welcome to the Private Equity Value Creation Podcast, where we interview leading investors, operators, bankers, and advisors to help you answer one question. How do we increase the enterprise value of our companies? My name is Shiv Narayanan, and each episode, I will dive deep with a guest to help you become a better value creator and capital allocator. So with that said, let's jump right in and let's get started with today's episode.

1:45My guest today is Carl Sayun, and he is the operating director and head of data and analytics at Briegel Milestone. And what I really loved about this episode is almost every PE firms, portcodes that we come into, we see data as being a major issue. And that's Carl's entire role at Briegel Milestone. And he walks us through how they approach data infrastructure, tracking, and how they leverage analytics to make critical decisions across customer success, marketing and sales, lead scoring, and even on the product side. And there's a ton of great takeaways when it comes to designing the right infrastructure, when you do it, what you should be tracking, what are the best practices.

2:21And we just also talk about on the value creation side and strategically how Briegel goes about approaching this with their portfolio company. So I think there's a ton to learn for all PE investors, especially if data is one of those ongoing issues that you see with your portfolio companies, which I imagine is a lot of you and the investments that you're making. So with that said, I'll leave you to it. Enjoy the episode.

2:51All right, Carl, welcome to the show. How's it going? Good, good. And here? Good. Excited to have you on. So why don't we start with your background and Briegel milestone and let's go from there. Yeah, sounds good. Sounds good. So I studied industrial engineering, specializing in data and AI, actually. So this is where I really got into data science. And I got my first job at Walmart in Bentonville, Arkansas, doing data science in the real estate division of Walmart. Before moving back to Paris, where I'm from, where I worked for Deloitte for a few years. At Deloitte, my responsibility was to lead and to basically set up a data team in the financial services division of Deloitte.

3:33I spent a few years there before going to London. That was in 2019. I spent then three and a half years at BCG Gamma. As you know, BCG Gamma is the tech arm of BCG. And what we were doing, our team was to build BI and data science applications for big retailers globally. So my role there was a mix of really data engineering, data science, but also, you know, kind of customer success in some weird way and product management because I was working closely with the strategy team of BCG plus obviously the management team of the clients. And then I got a chance to join Briggle Mindstone in 2021. So I've been there for about five years.

4:16And my roommate has always been to lead our data analytics capabilities in the firm, which at the beginning meant two things. Number one was to work with management teams to drive enterprise value, really, across several types of initiatives, data-driven initiatives, based on my background. And number two, build an AI-powered sourcing engine for the investment team to obviously accelerate, help the investment team accelerate the pipeline velocity and make sure we get quality deals in IC effectively. And so that has been part of my remit. Obviously, things have evolved over time, but that's how I got to Brigham and Maxson in the first place.

4:56And so can you expand on both those things? Expand on your mandate at Brigham and what you're kind of focused on and maybe give some more detail on how you're implementing that. So Bruegel Manson is a London-based software private equity firm with$2.3 billion in capital raise since inception across three funds. We're pan-European and we invest in software, AI, and cybersecurity. We're quite proud of that. So that's just context on Bruegel. I guess in terms of my remit, And so, as I said, 80 % of my time is spent on value creation initiatives with the portfolio companies. And within that, I break down the type of work I do in two buckets.

5:38Number one is making sure by year one, end of year one, that we've built, you know, very robust information management, if you will, right? Robust data infrastructure, financial and operational reporting. That's the priority number one for every single deal we make, you know, for me. And then from year one to year five, then we move on to commercial analytics use cases, which is things around trend analytics, lead scoring, pricing packaging. So effectively, anything that falls under really customer acquisition, customer retention and expansion, which is quite exciting work. And I think it's always very enabled by the work we do on year one.

6:22If you have good data in the first place, that is accessible, it's reliable, etc., then it makes everything easier to drive those insights and drive those actions that then drive outcomes. This is something that we've seen as a problem in almost every business. Data is such a huge problem. And the right data infrastructure is not in place. The right analytics platforms are not connected. The data is in one system here and in another system here, and they don't really talk to each other. So can you expand on that piece? Like explain year one a little bit further, because I know a lot of PE firms and their portfolio companies struggle with this.

6:59So what does that work actually involve? And maybe dive into different functions and what way you stand up there on the data side. So it all starts pre-deal, right? So when we are reading exclusivity, I will run some data workshops where I'm effectively going to assess the maturity of not only the data, but the systems, of course, the corporate systems that they have, the data that they have as well. and the quality of the reporting, right? It all starts there. And then, you know, we literally spend, you know, a full day, sometimes a day and a half with the relevant stakeholders and that, you know, management teams, members, CRO, CFO.

7:33Of course, if there's a head of data already, we get these guys in for this workshop. And basically, before the deal is even made, we already have a sense of, from an investment perspective as well, what it is going to require on year one, right? And that's important. Often, more often than not, the companies we invest in, you know, we will have, you know, very good set of corporate systems, right? So rarely we need to do, you know, much changes there. It's minor changes. What needs to be done often, however, is to start integrating those systems together, right? Sometimes those systems don't speak to each other.

8:07And that's the first issue you get, right? Because then, you know, you don't have the same, for example, unique identifier across all your systems. And so when you start to pull your data warehouse solution, it's a bit of a more difficult job to do right um so this is so we effectively are based on what happened pre-deal and our findings we have built a data roadmap with like a budget and timelines and everything right and you know it will always look similar in terms of like the end outcome which is to have built this data infrastructure but data infrastructure i mean like obviously the tools your data stack but also what kind of reporting do you need and uh it often starts you know in terms of like you know integrating the systems together, you know, making sure we have a unique identifier, aligning on the tech stack that we want to, the data stack we want to implement at the portfolio company.

8:55And then we get on the implementation, right? You know, I'm very much present and involved in me and my team on this, but we need, you know, often we have a CFO typically kind of accountable for this piece of work. And we work together on this, we work with the deal team very closely to first, once we've aligned on, once we've fixed all the systems issue and integration, by the way, I didn't speak about systems hygiene. That's the other big issue we tend to have, right, is, you know, we may have a very good CRM system, relatively good processes, but actually there's a lot of inconsistencies in your CRM, which is causing obviously the data to be wrong and unreliable.

9:34So that's another thing that typically needs to be resolved with the right processes in place that need to be robust, right? But then we actually go as deep before we get our hands dirty as mocking up what we want to build, really, like, you know, what exact dashboards we want to build, what exact views, what kind of questions we want to be able to answer with the data. So we come up with those mock-ups. We'll keep refining them with the management team, with the deal team, until we're all very satisfied. And then we get on the building, right? and then of course requirements change over time but that's typically how it works and the great thing about also AI today is that what used to take us maybe 7-8 months now takes us frankly 3 months to do if not less so now building kind of end-to-end data infrastructure from A to Z is much faster thanks to AI Yeah, a lot of what you said resonates so much like when we're coming into companies' feedback portcos and we look at their data and we're trying to sort out their marketing strategy and where they should invest their limited budget or what the ROI looks like, they don't have a lot of these systems.

10:43They don't have enough data to make the right decisions. So oftentimes we find ourselves helping them organize their data or organize the insights so we can have the right conversations at a board level and connect it to finances. So a lot of what you're saying is resonating. Help me understand, because in putting all this stuff in place, there's a lot of infrastructure work required. And it's like front loaded investment in terms of time, people, the costs associated with, let's say, adding another system. And meanwhile, that infrastructure work isn't really driving revenue at the same time. Right.

11:15So how do you balance that in your world? And how do you look at that upfront investment at Briegel? Yeah. So I think we have a great understanding at Briegel that this is foundational work, enablement work. And we understand that it's quite difficult to kind of allocate or assign a value to it, right? Like that's the billion-dollar question. How valuable is it to build robust data infrastructure? No one has an answer to this. And so obviously you can argue you can generate revenue uplift, cost savings. You can argue in many ways, right? But it's very difficult to put a number. Thankfully, I think in our organization, everyone recognizes the importance of it and what it does moving forward.

11:58And, you know, just to be clear, while we do this, it's not like we stay idle on the value creation plan, right? A lot of things happen in parallel, whether that's kind of more, you know, whatever that is, right? Like FP &A work, is it talent, you know, hiring people, is it, you know, go-to-market stuff? Things happen in parallel. And even on the data side of things, we get busy, you know, building this infrastructure. But there are often quick wins you can get in parallel from a go-to-market standpoint, for example, a revenue standpoint, that we make sure to unlock. So I'll give you an example.

12:30For one portfolio company, we saw pre-deal again that there was this need of building robust infrastructure, but we also saw other needs in the go-to-market side that had kind of data aspect to it. And that was on pricing packaging. So that portfolio company had more of an opportunity to basically get a lot of, like, transfer a lot of their reoccurring revenue into recurring revenue. So how do you kind of package your existing products and services accordingly to maximize your percentage of recurring revenue? But they also had a short-term opportunity to increase prices, frankly. And so this required, you know, just in only a few weeks after the deal was made, we ran some basic data analysis to see, okay, you know, from the customers who have inflationary clause in their contract, which one are being applied?

13:16from out of our, you know, several various customer cohorts, which customer cohorts are stagnant, which one haven't had a price increase for more than, you know, let's say three years, right? Out of our customer base, which one have seemed to be underpriced based on what peers are being, you know, charged, right? And so based on that, we were able to come up with like, you know, a plan on, you know, each customer, how much you can increase prices by. and that was implemented in the first quarter, right? And that happened in parallel of the building that's infrastructure, right? If that makes sense.

13:51So, you know, so we don't, it's not like we stay idle on other topics during the first year. We'll get back to the show in just a moment, but before we do, one of the most common and important value creation levers that we hear about on the show from private equity investors and our own PE partners is go-to-market. Yet when these same PE partners bring us into their portfolio companies or new target investments that they're exploring, we find that the marketing function is quite immature, underutilized, and under-optimized. And so that's a huge opportunity that we see inside these companies. And if you have a portfolio company that you feel like it's scaled a lot faster to drive more pipeline and revenue, or you're looking at a new investment where you feel like that could be core to your investment thesis, well, we'd love to explore that with you and figure out how we can partner with you to drive more enterprise value creation on the marketing side, similar to the way that we've done with major PE firms like TA, Updata, HG, SDG, and many more.

14:50At this point, we've done hundreds of engagements across hundreds of industries and verticals, and we have a ton of benchmarks and frameworks that we bring to these engagements to help you drive as much enterprise value as quickly as possible. So if that sounds like something that you might be interested in, you can just email me directly at shiv at howtosass.com or go to our website and schedule a demo. And we'd love to speak with you about it further. And now with that said, let's get back to the show. Yeah, 100%. That makes so much sense. I think having this longer term view on the investment on data as almost an enablement function so that you can long-term grow the value of the business faster or better is the right way to look at things.

15:31I think sometimes PE firms or just companies get impatient about this type of work and then they kind of rush through it and then they don't have the right data to make the right kinds of decisions. So I guess that's really the trade-off. You said something earlier that also resonated with me, which is you said you get the executive team to really own this. You're not the ultimate person with the accountability there. Can you expand on that further? Yeah, so I think we've had experiences in my five years at Braggle White where we push for something to happen, but there's no one on the exec team accountable for the work, no one that's sponsoring it really.

16:07and so then you don't get the outcome that you want. I think we've learned that to maximize obviously the results from a piece of work, we need to make sure there is accountability on the C-level side. So anything that we do is sponsored today by an exec. That's very important. That's not going to change. So every initiative that our value creation team does goes through the board really. right and that's where a lot of the i guess projects come from right these discussions in the board things that we want to recommend to do sometimes it's inbound sometimes it's outbound based on us saying guys i think we can support you on this right um so yeah that's that's extremely important to have executive sponsorship which areas do you see most lacking when it comes to the data side like obviously we get we get the viewpoint on the marketing side but i'm curious like from your perspective where are the biggest gaps inside these organizations yeah actually interestingly uh quite i have a strong opinion on this um well there are two gaps i can think of one is really people right so um you know for we invest in companies as i may have said earlier that are between 20 million you know ar companies to to 80 to 90 right so you know typical porco might have you know 30 million ar um and sometimes quite often actually we don't have a dedicated data person, right?

17:30So that's the first thing, right? So you usually have data-driven people in the organization that have taken some of the work on data, but no dedicated resources there, right? So that's something I try to bring as well. So once we build all that infrastructure and all of that stuff, we need someone to maintain it. We're not going to be maintaining it. We're not going to be enhancing it ourselves. We're going to typically bring in someone to do that. and in some cases actually we do get we do allocate that piece that that role to some existing you know resources they have in the company right sometimes you have engineers in tech in the engineering organization that you know are very happy to take on this role they get excited by doing that so but we make sure that every portfolio company has someone who can maintain this moving forward right so that's that's the number one gap just people frankly people yeah yeah the second gap that I find interesting is, you know, quite often I see those software companies not necessarily tracking a lot of product telemetry, right?

18:32There's a lot of value, as you know, in, you know, in the kind of understanding, you know, user behavior in your application. And often there is, you know, basic tracking that's not truly being used, right, for, again, call it commercial use cases, right? So this is not being used by customer success enough to really understand the health of a customer, for example. It's not being used enough by product organization to understand, you know, what works, what doesn't, you know, what kind of A-B test you should be running, you know, to basically, you know, improve the experience of the users. And I think that's a gap that I was, you know, surprised to see.

19:12And, you know, we try to support on this as well to make sure that every product and customer a success organization get smarter about the usage of the product. Yeah, totally. On the product side, there's just so much data in terms of metrics and improving retention and things like that. You mentioned a couple of areas earlier. You said churn analytics is a big area and also lead scoring. Can you expand on both of those? Yeah, churn analytics, obviously, all our organizations have customer success teams that have typically the responsibility to save customers and to get them to stay and to also expand as well, right?

19:51Upsell and cross-sell. But again, because some of our brokers don't have, you know, big data teams they can rely on to actually go deep in the data and understand patterns and everything, we get involved quite often on this type of work, right? So, you know, no matter how good your GR is, you can always get better. And, you know, there's a lot, as you know,$1 saved from a customer is worth a lot. And so typically what we do in probably a period of about three weeks, right, is that we go very deep in understanding, first of all, basic questions. What is the magnitude of your churn? Who's churning?

20:31Why are they churning? When are they churning? And trying to identify patterns and kind of leading indicators of churn, right? and again this is also where usage comes in as well so then once you identify the leading indicators of churn then of course you have everything you need to build a predictive model to basically score each customer in terms of risk of churning and so this is one of the things we implement very frequently for our portfolio companies and now that goes all the way you know, to kind of, once you have your scores, you have, you have also the reasons why you scored them, you know, well or badly, right.

21:15Then you can actually, you know, using clothes, for example, you can start automating agents to, to basically notify customers success teams every Monday, send an email at 8am saying, you know, Hey, Shiv, those are the five customers you should, you should be taking care of this week, right? Based on, and of course, then you share why this is the case. You actually create a battle card out of this and kind of what should be your play when you speak to the customer based on what you see in the data, right? So this kind of end-to-end, you know, churn analytics project is very valuable. You know, it's a mix, of course, of, you know, getting insights and getting the customer success smarter about their customers.

21:53But of course, the goal is to then take action. So we identify actions that need to be taken that we recommend customer success to take or product in some cases to take. And of course, then that predictive model is then productized and used by the customer success team on a daily basis to basically track and assess the health of their customers. So yeah, this is a big piece of work that we tend to do. And as I said, it can take two to three weeks to build, right? So I think there's a lot of value in doing this. And then lead scoring is another one. So obviously the first one I mentioned, Chain Analytics, is focused on customer retention.

22:30It's going on customer acquisition, right? And so one of the things we do is really assess the dynamic on the customer acquisition side and really why is growth slowing down, right? And being able to diagnose that and understand which customer segments are we underperforming on, why that's the case, what's happening in the funnel, the conversion rates, or the volume of them create. So we really sort of diagnose that. And often one low-hanging fruit indeed is to score kind of the opportunities in the middle of the pipeline in terms of probability of converting. And, of course, you know, based also on the size, you want to prioritize the ones that will bring you more money, you know, that have the highest kind of average selling price target.

23:16But, yeah, I think there's a big opportunity there that we tend to work on. sometimes it's not necessarily so sometimes it's also a free trial to pay conversion rate that we try to give Intel on and kind of scoring every free trial and saying these ones are the most likely to convert as you can imagine that's very much driven by usage you know people sign up to free trials and you know they try the tool for 20 minutes and they never come back and you know usually the ones that at least use it for two days in a row that's already a massive signal for example right so those are the things we the type of things we support indeed both on the customer acquisition part and the customer retention.

23:55How has this stuff evolved in an AI world? Because as an example, retention and churn, right? Like in product companies where they have per-seat pricing, they're changing to usage-based pricing or deploying different types of pricing models to capture value from customers. And then on the lead scoring side, it's harder to score leads because they're interacting with AI platforms or they're coming from different sources. So I'm curious, like how the data analysis here has evolved, or are you more focused on like optimizing already what's working or where where a company is currently focused? Yeah, yeah.

24:28So I think I think it's more of the latter. So, you know, we you have to implement the right processes, first of all, to be able to obviously, again, to have the right data that you can track, that you can that you can you can measure, that you can predict. right um i guess for now i haven't seen much changes i guess in that process with that new world and to your point on the kind of the trends we're seeing in ai pricing and packaging uh that's not impacting a lot if anything you know what we do is you know when you do your segmentation that you take that into account right so uh you would sort of know what kind of customers would be wanting what kind of you know pricing and packaging right where they would fit in terms of tiers etc So we do try to leverage that as an input to the models, but it hasn't changed dramatically how we think about it today.

25:18That might evolve in the future. How has that affected the data models that you're building for these companies or on the finance side when you look at forecasting revenue? Has that changed anything on that side for you guys? No, it hasn't changed anything because when we think about forecasting, usually we don't take into account, most of our brokers at least, don't take into account prospects in the final that are too early stage, right? So in any case, you know, it's only prospects that have a likelihood of 50 % plus of converting that are already part of the forecast, if that makes sense. So that hasn't changed much, right?

25:51I don't believe, we tend not to believe that your early stage MQLs should be really integrated in your forecast because the probability of converting is still low. But yeah, to your point, this will evolve, right? And as potential customers get smarter and smarter about using those LLMs and really when they reach out to you, it's quite, you're already maybe at the 50 % mark, then that changes things, right? But I don't think we're at that stage yet, at least from what I've seen. I understood. Okay, that's helpful. And then coming back to something you opened earlier, you said year one, we've set up the infrastructure, but then year two onwards, we use it to drive decisions.

26:27Can you touch on that? Like, what does that cycle look like? How are you evolving the data over time or how is it helping you make decisions on an ongoing basis? Yeah, of course. So I think what we especially lately, right, what we've done is we've changed the way we think about how to implement and develop this data infrastructure. So before it used to was a classic modern data stack. So you would use a data ingestion tool like Fivetrend, data warehouse like Snowflake. Then you would use DBT for transformation, and then you would use a Power BI or Tableau to do the dashboard. Things have evolved now, and I think what we've been doing is getting more and more common, which is to build custom BI applications, effectively, for Percos that are obviously all secure, hosted on Snowflake, for example.

27:17that fit directly from your data warehouse. There is much more importance, by the way, now, as everyone is reading about this these days, but on the semantic layer and making sure that your semantic layer is, you know, you spend a lot of time on that to make sure that once you interact with your data, you know, you and I are getting the same answer when we ask Claude some questions, right, effectively. And so, you know, you work hard on your semantic layer, you build your custom web application that's now your BI tool. and then more importantly you also create an mcp out of uh out of your you know your data warehouse in this case or semantic layer so in this case on snowflake and then you you enable you know business users to interact with your data via cloud for example right so you create your mcp you add your connector on cloud and you start you know diagnosing your data asking questions basic and complex data you know questions on cloud so that's that's first of all how kind kind of at a high level how we think about that data infrastructure I've been speaking about.

28:17And then, of course, so as I said, so year one, you get to that point where you've built that, sometimes obviously earlier, but year one is sort of the latest. Then moving forward, you are now equipped with all of that data to drive decisions. So we make sure to kind of incorporate a data-driven culture in the organization, right? So whether that's the sales team on their weekly sales calls to make sure that, you know, they have built and they have their own sort of dashboards that they track, that they look at every single week to understand, you know, how the pipeline is looking, to also make people accountable for the system's hygiene as well, you know, all of that stuff.

28:52Customer success is now much smarter about their customers because you can create Customer 360 very easily. You know, before, if you don't have a proper data warehouse, you know, Customer 360 is painful. You know, you get data from siloed systems and, you know, you have errors, things don't make sense. if you've done your job well on year one, then this is customer 360s are very easy and they really can be very powerful. They're very powerful, not only for customer success, but also for sales team to understand how the ICP customers are doing, how things are evolving, et cetera. Yeah, I think there's a fine balance between getting insights so you can make decisions sooner versus just setting up the whole infrastructure.

29:31So that's a bit of a trade-off. Yeah, exactly. Exactly. And usually, you know, like when we invest in a company, of course like the board is uh wants to make sure that you know in the first board meeting in whatever month three of the investment you have something good to show and you have a very good you know you're able to give good visibility to the board in terms of you know the forecast what's going to come in etc yeah but often at you know at month three if you do this correctly you know there's less to show right so you know we need to have that patience we need to be patient enough. I think you mentioned patience earlier.

30:04You need to be patient enough, you know, to wait for, you know, this to be done properly. Because again, if you take shortcuts, you're going to regret it later. You're going to pay for it long term for sure. Yeah, totally. Yeah. Carl, we're coming up on time before we close things off. If people want to learn more about you or Briegel Milestone, what's the best way for them to do that? Yeah, of course. I'm happy to share my work email, happy to go, you know, have a chat over coffee, lunch, etc. Always glad to exchange notes and also my LinkedIn profile. I'm happy to share it as well, Shiv. Maybe we can do that via your platform.

30:37Excellent. We'll be sure to include all those links in the show notes. And with that said, Carl, thanks for coming on and sharing your wisdom on the importance of data. It's something that we see as a problem in so many companies. And I think more PE firms should be taking the type of approach that you are at Real Milestone. So I appreciate you coming on and sharing your wisdom. Sounds good. Thanks, Shiv. Good to see you. Thanks for listening to today's episode. Before we close off, if you haven't already, go grab a copy of my new book, AI Marketing Blueprint. It just hit number one bestseller status on Amazon across multiple categories, and it walks you through a proven framework on how to scale revenue and pipeline in a world with AI platforms where customers are self-educating and learning a lot more off your website and where companies are struggling with declining inbound traffic volumes from paid search, paid social, and even organic and SEO channels.

31:26So if that sounds like you or a company that you're invested in, go grab a copy for yourself, your team or your portfolio company. And I'm sure it's going to help you out on your journey with that business. And with that said, thanks for listening today and we'll see you guys next time.