Org Charts for Startups: When to Add Layers and When to Resist
Early-stage founders hear two contradictory pieces of advice. The first: stay flat, move fast, avoid bureaucracy. The second: build structure before you scale, or chaos will catch up with you.
Both are right, and both are dangerous when followed blindly. A startup that adds management layers too early becomes slow, expensive, and political before it has found its footing. A startup that resists them too long ends up with an exhausted founder, stalled decisions, and good people leaving because nobody has time to lead them.
The real question isn’t whether to add layers. It’s when, why, and what kind.
What a “layer” actually costs
Every management layer you add has a price, and it’s more than the manager’s salary.
Speed. Each layer adds a step between a decision and the people who carry it out. Information travels up more slowly, and instructions come down more diluted.
Distance from the customer. Founders who sit three layers above the front line often learn about problems late, and secondhand.
Cost. Managers are typically among your more expensive hires, and in the early years a manager often produces less direct output than a strong individual contributor.
Politics. More layers create more positions to compete for, more territory to protect, and more room for misaligned incentives.
Layers aren’t free, so they need to earn their place.
What a missing layer costs
Too few layers have a price too, though it’s harder to see on a spreadsheet.
A founder bottleneck. When everything routes through one or two people, decisions queue up, and the company can only move as fast as the founder’s calendar allows.
Neglected people. Employees with no real manager get little feedback, no development, and no one advocating for them. They tend to leave quietly.
Inconsistent execution. Without someone owning a function, standards vary from person to person and quality depends on who happened to do the work.
Burnout at the top. Founders managing 12 to 15 direct reports while also running strategy, fundraising, and sales are rarely doing any of those things well.
A cautionary tale from both directions
Two well-known experiments show what happens at the extremes.
Google tried removing managers. In its early years, Google experimented with eliminating engineering managers entirely, on the theory that they slowed engineers down. The experiment was short-lived. People came to their founders with questions about everything from expense reports to interpersonal conflicts, and the company reversed course. Years later, Google’s internal research programme, known as Project Oxygen, found that good managers measurably improved team performance and retention.
Zappos removed traditional hierarchy altogether. In 2015, the company adopted Holacracy, a self-management system that replaced managers with self-organising roles and “circles.” Employees uncomfortable with the change were offered a buyout, and a notable share of the workforce took it. The system was later reported to have been loosened considerably.
The lesson from both isn’t that hierarchy is good or bad. Structure solves real problems, and removing it without solving those problems another way simply moves the pain elsewhere.
On the other side, there’s an equally instructive counter-trend. In 2024, Paul Graham’s essay on “founder mode,” prompted by a talk from Airbnb’s Brian Chesky, sparked wide debate about founders who had over-delegated, letting layers of management insulate them from the details of their own businesses. Too many layers can cost a company as much as too few.
The stages most startups move through
Every company is different, but the structural challenges tend to arrive in a predictable order.
1 to 10 people: flat by necessity.Everyone reports to a founder. Communication happens naturally. Structure here should be almost entirely about clarity of roles, not hierarchy. Who owns what?
10 to 25 people: the first leads.The founder’s span starts to break. This is usually when the first “lead” roles appear: experienced people who still do the work but also guide two to five others. These are player-coaches, not full-time managers.
25 to 60 people: functional heads.Distinct functions such as sales, operations, product, and finance now need dedicated owners. The founder moves from managing work to managing the people who manage work. This transition is harder than it sounds, and many founders resist it longer than they should.
60 to 150 people: middle management arrives.Functions grow large enough to need their own sub-teams. A second layer of managers forms beneath functional heads. Communication can no longer rely on everyone knowing everyone, so documented processes, regular rituals, and clear decision rights become essential.
150+ people: formal structure.At this size, informal coordination breaks down entirely. The company needs defined reporting lines, consistent role levels, documented policies, and deliberate internal communication.
Treat these as patterns, not rules. A 20-person company running complex operations across several cities may need structure earlier than a 50-person software company where everyone sits in one room.
Signals it’s time to add a layer
Add structure when you see these signals, not when you hit a headcount number.
A manager has more direct reports than they can genuinely support. There’s no universal rule, but once regular one-on-ones start getting skipped and people wait days for decisions, the span is too wide. Many practitioners treat roughly six to eight reports as comfortable for a manager who also has their own work, though highly experienced, independent teams can handle more.
Decisions are queuing. Work stops while people wait for approval from someone who’s overloaded.
The founder is the only link between functions. If sales and operations can only coordinate through the CEO, you’re missing a layer or a process.
New hires aren’t being onboarded properly. If nobody has time to train new joiners, quality drops and attrition rises.
Specialist work needs specialist leadership. A finance or engineering team led by someone without that expertise often struggles with quality and hiring.
Good people are leaving because they have no growth path. Sometimes a layer is needed not just for operations but to give strong performers somewhere to grow.
Signals you should resist
Some of the most common reasons startups add layers are the wrong ones.
To give someone a promotion. Rewarding a strong performer by making them a manager, when there’s no real management need, often loses you a great individual contributor and gains you a reluctant manager. Create senior individual contributor roles instead.
To solve a performance problem. Putting a manager above an underperformer rarely fixes the underperformance. Address it directly.
To hire a big-company executive too early. A VP from a large organisation often expects a team, a budget, and processes that a 30-person startup doesn’t have yet. The mismatch frustrates everyone.
To create managers with one or two reports. A manager with a single direct report is usually a sign that the layer isn’t needed yet, or that the role is really a senior individual contributor position with a management title.
To insulate the founder from difficult details. Delegation is necessary. Losing touch with customers, product quality, or team morale is not.
Because the org chart “looks small.” Structure should be designed for how work flows, not for how it looks to investors or candidates.
Alternatives to adding a layer
Before adding a manager, check whether one of these would solve the problem more cheaply:
Clear decision rights. Many bottlenecks disappear once it’s written down who decides what, who must be consulted, and who just needs to be informed. Frameworks like RACI or DACI help.
Player-coach leads. A senior team member who guides others while still doing the work can bridge the gap for a long time.
Documented processes. If the same questions keep reaching the founder, standard operating procedures may be the fix, not another manager.
Operating rhythms. Weekly leadership meetings, monthly reviews, and shared dashboards can replace a surprising amount of ad-hoc coordination.
Better tools. Sometimes the coordination problem is really an information problem.
Titles: the hidden structural problem
Startups, especially in fast-growing markets like India, often hand out senior titles early to attract talent or compensate for lower salaries. A “VP” or “Head of” at 25 people may feel like a cheap concession at the time.
It becomes expensive later. When the company grows and needs to hire an experienced leader above that person, the conversation is painful. Titles also set expectations about pay, reporting lines, and authority that the company may not be able to meet.
Some designations also carry legal weight. In India, roles such as CEO, CFO, company secretary, and whole-time director can count as Key Managerial Personnel under the Companies Act, 2013, with specific statutory responsibilities. These designations shouldn’t be handed out casually.
A better approach: Define a simple levelling framework early, even a rough one, and keep titles tied to scope and responsibility, not to negotiation.
Designing your org chart well
When you do restructure, a few principles help:
Design around the work, not the people. Decide what roles the business needs first, then match people to them. Building structures around individuals creates fragile organisations.
Write role definitions. Every box on the chart should have a clear mandate, decision rights, and measures of success.
Plan one stage ahead. Know roughly what your structure will need to look like at the next stage of growth, so today’s hires and titles don’t block tomorrow’s.
Communicate changes carefully. Restructuring affects people’s sense of status and security. Explain the reasoning, address concerns individually, and update contracts and role documents to match.
Review regularly. An org chart that worked at 30 people may be wrong at 60. Revisit it at least once a year, or whenever the business changes significantly.
A quick self-check
Ask yourself:
How many people report directly to me, and when did I last have a meaningful conversation with each of them?
Where in the company are decisions waiting the longest?
Do we have anyone with a management title who manages one person or none?
Could every team member describe what they own and who they report to?
Would our current titles make it difficult to hire a senior leader next year?
If those questions are uncomfortable to answer, your org chart may need attention, whether that means adding a layer or removing one.
The bottom line
The best startup org charts aren’t the flattest or the most structured. They’re the ones where every layer exists for a reason the founder can explain. Add structure when the business genuinely needs it, resist it when the motivation is status or appearance, and revisit the decision as the company grows.
Comments