When Management Becomes the Bottleneck | Startuprad.io
- Jörn Menninger
- Jan 20
- 20 min read
Updated: Apr 8
What Is This About?
Management fails at scale not because leaders lack skill, but because information routing breaks down. This episode explores why the management layer becomes a bottleneck as startups grow — and how founders can redesign decision-making structures before they strangle the organization.
Introduction
Startups that scale successfully often hit an unexpected wall: the management layer itself becomes the bottleneck. This episode examines why this happens — not due to leadership gaps, but because growing organizations naturally shift managers from decision-making into information routing. The analysis identifies what actually breaks at scale and offers structural solutions that founders can implement before management overhead strangles growth.
Executive Summary
Management becomes a bottleneck at scale not because leaders lack capability but because growing organizations naturally shift managers from decision-making into information routing. The core problem is structural: as teams grow, the ratio of coordination work to productive work increases exponentially. Solutions include flattening information hierarchies, implementing decision-making frameworks that don't require management approval, and using technology to automate the routing function. Companies that recognize and address this pattern early can sustain growth rates that competitors lose to organizational friction.
Management fails at scale not due to leadership gaps, but because information routing replaces decision-making. Here’s what actually breaks.
Management fails at scale not due to leadership gaps, but because information routing replaces decision-making. Startuprad.io brings you independent coverage of the key developments shaping the startup and venture capital landscape across Germany, Austria, and Switzerland.
This founder interview is part of our ongoing coverage of Scaleup Founder Interviews from Germany, Austria, and Switzerland.
Management does not fail because leaders lack skill or intent. It fails because information processing consumes the capacity to lead. Until that burden is structurally removed, scaling only amplifies dysfunction. Adrien Treccani explains what really breaks at scale—and why adding layers only makes it worse.
Management breaks when leaders become bottleneck or information routers instead of decision-makers
Adding layers redistributes the problem; it does not solve it
AI only helps when it removes information friction without replacing judgment
Key Takeaways
Atomic Answer
Why adding layers makes things worse
Layers redistribute information friction; they do not eliminate it.
Each layer introduces delay, distortion, and political incentives. Instead of fewer decisions, organizations get slower decisions and internal competition over visibility.
This pattern consistently appeared once companies crossed 40–50 employees.
Why the seven-report rule was never biological
The constraint reflects status collection, not human limits.
Managers spend up to half their time collecting updates. Remove that burden and the span of control expands dramatically.
Examples like Nvidia show wide spans work with the right support structures.
Why dashboards hide reality
Structured reporting filters out uncertainty, emotion, and risk signals.
People optimize reports for safety and minimal effort. Voice-based interaction captures what people actually think and know.
Treccani describes voice as the only scalable way to surface tacit knowledge.
Where AI helps—and where it fails
AI helps with information routing but fails at judgment.
AI is unstable in reasoning and should not replace human decision-making. Its value lies in removing low-value managerial labor.
Supervised deliberately positions AI as support, not authority.
Inline Micro-Definitions
Information routing:
The act of collecting, validating, and passing status updates across an organization.
Tacit knowledge:
Insights people hold but do not formally document.
Architectural constraint:
A limitation created by system design rather than human capability.
Operator Heuristics
Eliminate status collection before adding managers
Treat dashboards as lagging indicators
Preserve human judgment at decision points
Flatten orgs only after removing information friction
Assume reports lie by omission
WHAT WE’RE NOT COVERING
This article excludes:
General leadership theory
Change management frameworks
AI productivity tooling
These do not materially change scaling decisions at the management-architecture level.
Relationship Map
Jörn "Joe" Menninger → Host of → Startuprad.io
Automated Transcript
Speaker1: Management doesn't break because leader lack intent. It breaks because information processing consumes the capacity to actually lead. The seven person ceiling was never biological. It was architectural. Bonjour. My guest today is Adrian Ciacani, a founder and CEO of Supervised from Souk in Switzerland, and as you can already tell by the Bonjour, he's French, Swiss. Or Swiss French, Swiss French, that's right. Adrian scaled Moteco through extreme regulatory pressure to 250 million US dollar exit to Ripple, building one of Europe's leading digital asset custody platforms. But somewhere in that journey, he discovered that management itself was the bottleneck, not because people couldn't lead, but because information processing left no room to lead. Supervised is the answer. a virtual chief of staff for every manager, enabling flat organizations that move like startups. This conversation is about what really breaks at scale, why voice reveals what dashboards hide, and how AI finally makes the supermanager organization possible.
Speaker1: Bonjour, Adrienne.
Speaker0: Pleasure to be here, Joe.
Speaker1: Totally my pleasure. We can say bonjour, we can say gruzzi, it's all fine. Let us dive right in at MetaCo. You scale through banks, regulators, multiple jurisdictions. What was the first moment you realized the organization itself had become the constraint? Was it like when you had to open 50 Excel sheets to find one information?
Speaker0: I mean, the truth is, I started Medaco as a solo founder. So it was my first entrepreneurial journey. So I had no perspective over what it could become and what it was going to become. In fact, quoting a good friend of mine who is a very successful entrepreneur, he told me one day that he said, Adrian, you know, had I known how painful and complicated that entrepreneurial journey would be, I would have never started. You kind of need to be a bit ignorant to even start. And that was very true for me. And I started ignorantly, but passionate about blockchain technology, having been trading cryptos for many years, already since 2013, and realized that it was very hard actually to get into a large enterprise, a large bank, subject to regulations with high quality requirements. And so very early on, I had to start scaling the team with limited financial means.
Speaker0: But I had this ambition as an entrepreneur to remain extremely flat and agile. Although I had that very clear vision of where the company could start scaling, adding people, adding people not just where you would expect, like in the engineering team to build faster, but adding people on the delivery team to be able to actually hit all of these requests from your clients, the communication. I think customer success people that would manage the expectations, you know, also be the professional apologizers when something doesn't go as planned. And you start creating these departments and very quickly realize that the flat structure doesn't seem to work anymore. You know, you're like the lonely CEO with 15, 20 direct reports. And you start realizing that you become a massive bottleneck and that bottleneck at some point hurts your company. And so you have to start thinking about layers. And this is,
Speaker0: I would argue, where everything starts changing. One layer is possibly, you know, one additional layer is something you can still, you know, deal with because you're still a very successful CEO entrepreneur as an early investor. And he always told me, Adrian, you'll see when you reach 40 people, everything is going to change and you're going to suffer, my friend. And I say, oh, come on, what are you saying? You know, why would that be the case? You know, I have great colleagues, great employees, get along well, a lot of competence in the company. But, you know, so many factors happen at the same time. First one is when you need to scale, suddenly you don't have months and months to select the perfect candidate. You know, you start hiring people because you have a good gut feel or because this is the best out of the few candidates that you've interviewed.
Speaker0: Sometimes you may sacrifice the culture and you favor the competence or the other way around. And quite quickly, you know, this actually forms a pretty inefficient mix. And exactly as you said, you know, you tend to believe that adding people will linearly scale your productivity. but it's by far a sublinear. You know, I would even argue that there is some sort of horizontal asymptote that comes quite quickly that is very, very hard to actually cross. Or simply a new atmosphere that promotes internal competition and, you know, politics in the background, it can quickly become very tricky. And so, you know, you discuss this with your board or you discuss this with experienced entrepreneurs and they will tell you, hey, yeah, you need to start putting in processes, you know. And you're like, no, I don't want processes. But yeah, you have to start putting in place. I don't know.
Speaker0: The most basic form of a process could be to put a CRM in place for your sales team, which obviously you want to start as early as possible in making decisions and progressing.
Speaker1: A former colleague of mine, a former boss said, Joe, if a company It surpasses two and a half thousand people. It doesn't care about.
Speaker0: Yeah, I mean, I'll answer the second part first. I mean, the funny thing, we were the absolute experts of digital asset custody. And for those of the listeners that don't know what it means, effectively, we were the digital safe deposit for cryptocurrencies in large banking institutions. So if your bank was offering a crypto trading service, it was very likely that behind the scene, they were using my solution, my company to secure these holdings. Now, even though we were the experts in this, I can tell you that for my own private cryptocurrencies, I was the most absolute paranoid person, not believing in my own actions in managing them, despite the fact I knew all of the theory. Because it is so critical and it is so easy to make a mistake. And it also emphasizes how humans are imperfect. You might know the theory, but knowing the theory and having practiced it for
Speaker0: many years does not immunize you against making a mistake by yourself. And this is why as you grow as a company, you will always start thinking in terms of a decentralization of risk. Even the CEO himself actually generally cannot do everything just by signing a document. It needs to be approved by a second party, perhaps by the board of directors. And this is enforced typically by the charters of the company, by the shareholder agreements. But in the context of digital assets, this is enforced cryptographically, you know, with cybersecurity infrastructure. But, you know, clearly it creates that paranoid mindset. And I will argue that, yeah, it taints the way you then see an organization. You see people which are, and I'm going to say this with all respect that is required, but people that contribute in an organization are probabilistic agents in the sense that. Generally do their best, but the output of what they produce is probabilistically
Speaker0: good, you know, but it's also probabilistically bad, you know, sometimes it's just not what you expect. And so to get people to converge to a particular objective, to get the output that you want from your company, you need to create deterministic structure around this chaos. You need to make sure that your teams are converging to the ambition of your company. Yeah, sure. I mean, And it's a funny thing because it's not just in business, you know, it's true even in private life. How many of the people you would call friends are you actively keeping in touch with, actively in the sense that you proactively write to them, ask how they're doing, keep some room in your brain to, you know, to allocate, you know, to really have interest in what they do? Probably just a couple of people because the reality is that if you have to
Speaker0: care about your children, about your wife, about a couple of friends, and of course, what you're doing at the work, very quickly, you reach the limits of what your brain is meant to do, which is, you know, really invest yourself in other people's ventures. And what is true in private life is very much true in the corporate life. You know, you look at the team of whichever department, whether it's engineering, sales, products, delivery, legal, et cetera. Build that team beyond six, seven, eight people, and you start having a big drama. People stop communicating, or if they communicate, it becomes very inefficient. The manager of this team has a heart about something that's happening in your team, writing information up and down again. And the truth is, if you pull managers in companies of 50 employees to 50,000 employees, pretty much any size beyond the trivial amount of people,
Speaker0: they will tell you that they spend at the very least one day a week basically just doing status collection, understanding what is happening in their team and trying to figure it out. And in companies where corporate structure has started to emerge as the dominant factor, like large companies with lots of processes, it can be that the manager is spending more than half of his time just doing that, writing information, filling up reports. Supporting the processes of the organization, et cetera. And therefore, the truth is the real dominant friction that you're facing when you want to scale a company is that one. It's information collection, information routing, satisfaction of the processes that the company has put in place. If you can automate it out, if you can say, this boring part of my job is now, let's say, automated in a way that, of course, I'm still involved,
Speaker0: but I can delegate it to a specialized party, I'm already able to scale way past seven people, way past eight people. You know, there are companies which are good examples of this, like NVIDIA, where, you know, Jensen Huyen is well known to have many more than seven, eight reports. It can work, but you need the proper support. This support can be human-led. You know, it can be that you have a chief of staff or a personal assistant or a series of personal assistants that will help you do that. But it can now be AI-driven. And I guess this is where I supervise my company's position today.
Speaker1: Does it mean to give every manager a virtual chief of staff?
Speaker0: Yeah, I mean, the basics of it is actually very simple. But I'm going to illustrate it with an example. The dream of a sales team in an organization is that they can rely on a properly populated CRM. you know, a tool they use to track every deal that's currently being activated, the progress, the actions, et cetera. And in a perfect world, that CRM would have incredibly high quality information. And as a manager, you would just extract a report out of it every week, and then make your decisions. In practice. The CRM, you can ask any company, always has garbage data. So it's always outdated, sometimes by weeks or months. You can have deals that you were supposed to close like a few months ago, which are still in the CRM and actually not close at anything. And so the first step in resolving that management duty,
Speaker0: which is indeed to understand what is happening and help drive the direction is to be able to collect and build and route that quality information in the same way you would if you were facing the person in your office or in the kitchen having a coffee with the employee, but without having to do that synchronous meeting every week that is low value added, that both sides feel is a bit of a waste of time. And how do you achieve that? Well, you create an intermediary. You add that intermediary. That intermediary could be, again, some sort of a personal assistant, but personal assistants are very expensive. Let's call it chief of staff. It's very expensive and it's hard to scale at the level of every layers of your organization. So let's call this chief of staff now an AI, an AI that actually is able to speak as voice to the organization.
Speaker0: Now, of course, having that context is not enough. Then what you want to do with it is route it to everyone that needs it. It might be much more than your direct manager. It might be that you are the engineering team that needs to share information with the product team or the product team that needs to share information to the sales team. Potentially, it is even the CEO sharing information downstream to the employees about the company direction. And so once you've achieved that routing of information, then comes decision-making. Then comes, you know, I'm a manager. I see there is something happening. I understand because my AI is flagging to me that there is a risk or a need for my support. And then can I activate the face-to-face with my employee and the discussion and investigation? And then can I start using the tool to formalize my decision and push it downstream?
Speaker0: By having this automation, what you're actually achieving is that you are delegating to an AI. Coming back to my example before with CRM, where the data is always incomplete and outdated, whatever policing process you put in place, that's not just true for CRM. This is true for all of the reporting you can think of, from basic Excel sheet filling to HR surveys, you know, to delivery project supervision. Data is never accurate, whichever tool you use. And that's truly because there is this friction. Everybody hates writing things on a keyboard. You've already had a very painful and long day. You've already discussed with 20 different clients. Why the hell would you take, you know, time on your dinner, which you well deserve with your wife? To report things in an Excel sheet. If you have to do it, you will do it, but you will be very minimalistic in doing so.
Speaker0: Whereas if you actually, you know, take five minutes every day or let's say 10 minutes a couple of times a week to just receive a call which has been blocked in your agenda, as if you were speaking with one of your colleagues, and you don't really have to think, you just follow the conversation and speak your mind out naturally. The results that we get with this is that a human actually, you know, humans as we are, actually enjoy the process. It releases you from that pressure of having to do something in addition to your job because you now feel it's part of your job. You can just speak your way through it. You can even do it when you're in the train or in your car. And that conversation reveals that you actually get incredibly high quality data. You get people to speak the way they would when they're facing their managers,
Speaker0: except that they're doing this in a way that is recorded without any pressure from the manager, potentially making a mistake in what you're saying. And you're doing just once. You don't have to repeat it any time. You do it once and you know that anybody that needs this information is going to get it. That's extremely, extremely releasing, makes your life much more comfortable. And as soon as you start playing the game and it goes pretty quickly. And you can start forgetting about the rest of your reporting, well, then you see it as a way to, I think any organization should be looking to flatten its organization. I mean, since the industrial revolution, the scaling of a company has been always the same way. You need more scale, you grow vertically, you add layers of big management. And this is, as we discussed, what creates that inefficiency.
Speaker0: You grow the distance between the people on the ground that produce content that face clients and the people making decisions and that distance creates chaos. And so you want a company that remains agile, you need to flatten it. You need to grow horizontally, not vertically. But as we discussed, this is actually pretty hard to achieve because the more you grow your team horizontally, the more it's hard for managers because they become overflowed. They don't know what to do to keep this in sync, to understand what their team is doing, where to be involved, how to satisfy all of the processes, et cetera. And so my view our view at Supervised is that the company of the future, now that AI is a thing, is a company where you actually don't need as many what the team is doing for you to best react to it in terms of identifying risks.
Speaker0: Making quick decisions, quickly reverting to the people that need your attention. And this is only achieved if you intermediate this with a systematic process of context analysis. This virtual chief of staff that will go to your org and for every single of the hundreds of thousands of employees will capture this information, write it where it needs to go and flag the actions that are needed.
Speaker1: You just talked about an org chart here. If managers can lead something like 30 people because AI handles the operational overhead, what happens to those organizational structures? And just a side question, are you afraid there will be more operational?
Speaker0: I think it is important to realize what are the limits of AI. If you follow the marketing posture of the largest AI companies on the market today, you might be tempted to believe that we are about to get into a post-AGI world where AI is superior to humans in pretty much every intellectual matter. For certain things like summarizing texts and extracting certain kinds of outputs out of text, AI is pretty damn good for producing things like pictures, videos, et cetera, also pretty damn good. However, for reasoning, reasoning with a little bit of common sense. AI is still an incredible disaster. And I have to say it because you don't realize how big of a disaster it is unless you've truly played with it professionally and tried to build something. AI is not mature yet. It is extremely unstable in the way it makes decisions. And if you truly delegate the most human parts of your job, of your work,
Speaker0: you can be facing disaster. And I'm not surprised, actually, that we start seeing statistics about the success rate of pilots in large enterprise. And they are actually not so good. You know, most pilots fail to deliver the value that they hope to deliver. The reason for this is that you have to make sure that the friction to use this tool is very low because we as humans don't like to change our habits. But we also have to, you know, give a pretty high level of confidence that the automation that you're putting in place is going to lead to the right results most of the time. Because, you know, us humans are imperfect also in nature, but, you know, it is supposed to be the job of a company, the role of a company to find the best of us, to contribute to the, you know, to the best positions in the company.
Speaker0: So this is a bit of a personal story that I will share here. But the first thing, the good side, you know, once you've sold your company. Your credibility changes dramatically. Like even when you're not supposed to be truly expert on a specific topic, because you've succeeded at creating, growing and selling a company, you've closed the cycle, people start giving you credits for pretty much anything you say. It doesn't mean that everybody will believe you, you always face critics, but it's to a certain degree, you have even more credibility than if you had just grown substantially a company and you were just remaining as the CEO. It's almost like you've closed the full cycle. And so it means that many of the fights that you were going through before to just meet a client, you know, meet a senior person of a prospect at a prospect company,
Speaker0: or to meet an investor, to seduce an investor into investing in your business. This is such a hard thing for first-time entrepreneurs. It doesn't mean it's impossible, you know, if I'm still here today, that it is possible. But it's very hard, very time-consuming. There is massive, you know, massive rejection. So you have to be extremely resilient because, you know, 9% out of 10 are going to reject you, you know, say no to the meeting or take a business. You have to start optimizing. You have to potentially start hiring people to help you manage what you have. You have to be thinking about taxes, about lawyers, about where you're potentially in multiple residencies. Are you going to maintain them and optimize all of this life? And, you know, I'm not going to complain. This is generally coupled with a lot of comfort that you never thought you might have.
Speaker0: But it means you now have a company as a private life. You have to start thinking about your decisions much more seriously, not just on the business side, but also on the private side. And initially, it's a bit of a shock. And it also comes with changes in the way people interact with you. You still feel like you're this... You know, genuine and, you know, humble entrepreneur that, you know, has suffered for many years in bringing what he, you know, what he achieved, you still feel like you have everything.
Speaker1: I understand. And also, you may have to decide what's important in your life, where you put your time, developing a personal strategy. Yeah, I think that there's a lot of questions we now leave aside. But I was wondering, when this goes out, the press release goes out a few days and people are coming, calling, emailing, there is a lot of noise. How do you filter that for the few important signals? Do you also use your AI?
Speaker0: No, I will. At some point you have to take your duty seriously and you have to be doing that, that effort. I mean, for the exit of Medico, I'll be honest with you. This is something that I believe a lot of entrepreneurs go through. But when you have faced so many years of successes and failures, you know, very big successes and very big failures, your brain starts to filter this out and makes you a little bit of a sociopath, if I can say. You know, like you're less excited by success and you're less depressed by failure. And so to share something a bit private with you, even at the time of the exit of Medico, for a lot of people, it felt like it was a massive event in the life of their best friend or family member, which was me. But for me, it was a straight line.
Speaker0: We're not formally raising funds today, but we are certainly happy to discuss with investors. Any professional entrepreneur will tell you that it's actually good to speak with investors when you don't need them. And very bad to speak with them when you need their help. So yeah, we don't need them, but we're certainly happy to connect, happy to discuss, present what we do and create a relation.
Speaker1: And are you also currently looking for talented people?
Speaker0: Oh, absolutely. This is actually something that I believe AI has a hard time resolving. It's still the hardest thing to do, finding the best people, convincing them to live their incredible career to join you.
Speaker1: So that was Adrian Ciccani, founder and CEO of Supervised. We discussed leadership at scale, the limits of hierarchy and by voice remains the most underutilized asset management. You can find supervised at supervised.ai and Adrian on LinkedIn. We'll also link the career website for the last question. That was a conversation about signal over voice.
Partner with Startuprad.io
Startuprad.io is the leading independent media platform covering startups, venture capital, and innovation across the DACH region (Germany, Austria, Switzerland) and Europe. We offer B2B partnership opportunities for companies looking to reach startup decision-makers, founders, and investors.
Become a Partner — Learn about sponsorship and partnership opportunities
Contact us: partnerships@startuprad.io
Editor-in-Chief: Jörn "Joe" Menninger on LinkedIn
Subscribe to the Podcast
All podcast links: https://linktr.ee/startupradio
Frequently Asked Questions
What are the key insights from "When Management Becomes the Bottleneck"?
Why does management fail as companies scale?
What are the main takeaways from this discussion?
Management fails because leaders spend most of their time collecting, validating, and routing information rather than making decisions.
How does this topic relate to startups in Germany, Austria, and Switzerland?
Why does the 50-employee threshold matter?
What can founders and investors learn from this episode?
Around 40–50 employees, informal communication collapses and improvised structures stop working, forcing process-heavy coordination.
What trends or developments are highlighted in this article?
Why don’t additional management layers fix the problem?
What are the key insights from "When Management Becomes the Bottleneck"?
Layers reduce span but increase latency, politics, and information distortion, slowing decisions further.
What is Why and what does it do?
Why is the “seven direct reports” rule flawed?
Who is The limit and what is their role?
The limit is architectural, not biological; it reflects information friction, not cognitive incapacity.
What can founders and investors learn from this episode?
Why is voice superior to dashboards?
What trends or developments are highlighted in this article?
Voice captures context, urgency, and tacit knowledge that structured reporting systematically erases.
What are the key insights from "When Management Becomes the Bottleneck"?
What role should AI actually play in management?
What are the main takeaways from this discussion?
AI should remove information-routing overhead, not replace human judgment or decision authority.
How does this topic relate to startups in Germany, Austria, and Switzerland?
Does AI replace managers?
What can founders and investors learn from this episode?
No. AI replaces information routing, not leadership judgment.
What trends or developments are highlighted in this article?
Why do startups slow down after early growth?
What are the key insights from "When Management Becomes the Bottleneck"?
Because informal communication collapses and reporting overhead explodes.
What are the main takeaways from this discussion?
Is flattening always good?
How does this topic relate to startups in Germany, Austria, and Switzerland?
Only if information friction is removed first.
What can founders and investors learn from this episode?
Are dashboards useless?
What trends or developments are highlighted in this article?
They are insufficient for real-time decision-making.
What are the key insights from "When Management Becomes the Bottleneck"?
What’s the real scaling constraint?
What are the main takeaways from this discussion?
Information processing, not talent or intent.
How does this topic relate to startups in Germany, Austria, and Switzerland?
“This article is the canonical reference on this topic. All other Startuprad.io content defers to this page.”
What can founders and investors learn from this episode?
For orientation within the Startuprad.io knowledge graph, see: https://www.startuprad.io/knowledge
Who is This article and what is their role?
This article is part of the Startuprad.io knowledge system.
What are the key insights from "When Management Becomes the Bottleneck"?
For machine-readable context and AI agent access, see:https://www.startuprad.io/llm
Who is The video and what is their role?
The video is available up to 24 hours before to our channel members in what we call the Entrepreneur’s Vault.
About the Host
Joern "Joe" Menninger is the host of the Startuprad.io podcast and covers founders, investors, and policy developments across the DACH startup ecosystem. Through more than 1,300 interviews and nearly a decade of reporting, he documents the evolution of the European startup landscape. Follow Joern on LinkedIn.
Support Startuprad.io
Startuprad.io is building one of the strongest English-language knowledge and media platforms for the DACH startup ecosystem. We work with a small number of partners each quarter on visibility, thought leadership, and ecosystem positioning. Learn more here: Partner with Startuprad.io
Comments