Organisation Structure and employees

A business's organisation structure rarely stays the same as it grows. The structure that works for 5 people is wrong for 250; the structure right for 250 would have suffocated the start-up. Understanding the typical evolution helps founders and managers anticipate when structural change is needed.

Stage 1 — Start-up (5 people)

Completely flat. The founder talks directly to everyone. No formal hierarchy. Decision-making is immediate; communication is verbal and informal. Each person typically does several different jobs (sales + customer service + admin in one role). No specialisation, no managers — just the founder coordinating the team.

Strengths: Fast, flexible, energetic. Weaknesses: No backup if the founder is ill; relies entirely on the founder's capacity.

Stage 2 — Small business (15-30 people)

The founder can no longer talk to everyone every day. First management layer emerges — typically 2-4 team leaders or department heads (one for sales, one for operations, perhaps one for finance/admin). Span of control 5-8. Founder now manages managers as well as some direct relationships.

Strengths: Still fast; founder remains close to the action; specialisation begins. Weaknesses: Founder still a bottleneck for many decisions; first signs of inconsistency between teams.

Stage 3 — Medium-small business (50-100 people)

Three-layer structure emerges: founder/CEO → 4-6 department managers → team leaders → workers. Departments fully specialise (sales, operations, finance, HR, marketing). HR function appears as a separate role (recruitment, contracts, payroll). Founder rarely manages workers directly.

Strengths: Coordinated functions; founder freed for strategy. Weaknesses: Departments start to silo; communication formality rises; cultural informality erodes.

Stage 4 — Medium business (100-250 people)

Four-layer structure: CEO → directors → managers → team leaders → workers. Formal HR, finance, IT, legal functions. Written processes, performance reviews, structured training, formal communication channels (intranet, all-hands meetings). Spans of control 5-10 in most layers. Centralisation of some decisions (pricing, technology, compliance) while operational decisions decentralise to managers.

Strengths: Can scale to handle complex business; supports specialisation and quality. Weaknesses: Slower decisions; can feel bureaucratic; founder may struggle to let go.

Drivers of structural change

The triggers for adding a new layer are predictable:

• A manager's span of control reaches 10-12+; quality of supervision falls.
• The founder/CEO can't track what's happening.
• Communication between layers starts to fail (junior staff don't know senior strategy; senior staff don't know front-line reality).
• Inconsistency between teams grows.
• Customer or supplier feedback reveals operational gaps.

Edexcel A* insight

The most common structural mistake is adding layers too slowly (founder tries to manage everyone; chaos builds; quality suffers). The second most common is adding layers too quickly (becomes bureaucratic and slow). The right cadence is to add a layer when you can clearly identify the work it will do and the decisions it will make — not because 'we're getting bigger and should have managers'.

Conclusion. Structure evolves from flat (5 people) → emerging hierarchy (30) → multi-layer (100-250). The driver is span of control: when managers can no longer effectively manage the people reporting to them, a new layer must be added. The art is in timing — adding layers as the business needs them, not as a status signal. Founders who refuse to add structure cap their firm's growth; firms that over-structure too early lose the agility that made them successful.

Flattening a six-layer hierarchy to four is one of the most consequential decisions a CEO can make. The financial saving (£12m) is genuine and visible; the operational implications (good and bad) are complex and only fully visible 2-3 years later. The right answer depends on what those two middle layers were actually doing — and whether the firm can absorb their work into remaining layers.

Arguments FOR flattening

• Speed of decisions. Six layers means decisions get filtered up and down through many people, slowing response time and distorting information. Four layers significantly speeds decision-making.
• Cost saving. £12m of recurring savings on management salaries is material. This funds investment elsewhere or improves margins.
• Communication quality. Each layer adds distortion. Removing layers brings senior leaders closer to front-line reality and brings front-line concerns up to leadership.
• Empowerment. Removing middle-management layers typically pushes decisions DOWN — managers nearer the work make the decisions, increasing autonomy and motivation (Herzberg).
• Removing 'bloat'. Many large firms have layers that exist for status / political reasons rather than operational ones. These can be removed with little operational cost.

Arguments AGAINST flattening

• Span of control increases. Removing 2 layers means remaining managers each have many more direct reports. Quality of supervision may fall.
• Removed managers were doing something. Even if some were 'bloat', many were coordinating across teams, managing day-to-day issues, mentoring juniors. That work doesn't disappear — it lands on the remaining managers or on team leaders not ready for it.
• Loss of career paths. Two fewer levels means fewer promotion steps for ambitious staff. Some may leave.
• Loss of expertise. Experienced managers carry institutional knowledge. Redundancy programmes typically lose more knowledge than expected.
• Short-term disruption. A major restructure consumes management attention for 12-18 months, distracts from operations, demoralises survivors who fear they're next.
• Risk of over-flattening. Manufacturing with 1,500 people genuinely needs management. Cutting too aggressively can leave the firm under-managed.

Critical questions before deciding

Before approving the flattening, the firm should answer:

1. Why did those layers exist? If they exist because work needed coordinating, removing them creates a coordination gap. If they exist because of status / political accumulation, removing them is fine.
2. What is the post-restructure span of control? Going from 5 direct reports to 10 is sustainable; going to 20 is not.
3. Can technology absorb the lost coordination? Modern tools (data dashboards, project management software) can replace some middle-management coordination functions.
4. What is the talent quality of remaining layers? Younger managers may need more support — flattening assumes existing managers can step up.

Justified judgement

The flattening can be the right move — but only with three conditions:

1. Audit which layers actually do what. Don't cut layers; cut roles where the work is duplicative, low-value or has been automated. Some of the 'middle management' may be doing critical coordinating; removing them creates problems. Other middle managers may be doing little of value; removing them is genuine improvement.

2. Invest in remaining managers. Expand training, give better tools (dashboards, decision-support software), reduce span of control to manageable levels (8-12, not 20+).

3. Phase carefully. Don't flatten everywhere at once. Pick one division, restructure, learn, then apply lessons elsewhere. Allow 12-18 months for full implementation.

Conclusion. The flattening can deliver real value — £12m savings, faster decisions, better employee engagement — but only if executed selectively and supported with investment in the remaining layers. A crude 'remove two layers everywhere' approach is the most common failure mode; the firm ends up with an under-managed organisation, decision quality drops, productivity falls, and the savings are eaten by hidden costs within 18-24 months. The strategic discipline is to cut deliberately, not uniformly.

The deeper insight is that organisational structure is load-bearing, not just symbolic. The layers carry real work, even when they look bureaucratic from outside. Successful flattening rebuilds the load distribution; failed flattening leaves the same work to be done with fewer people, leading to overwork and quality decline.

The founder's situation is one of the most common — and limiting — patterns in growing businesses. The recognition that 'I should delegate more' is the easy part; actually building the capability to delegate well is harder, and most founders never fully crack it. This evaluation must address WHY her past delegation attempts failed and provide a structured path to genuine delegation, not just more attempts at the same broken pattern.

Diagnose why delegation has been failing

The founder's complaint ('the work isn't done well enough') reveals several possible root causes:

1. Wrong task choice. She may be delegating tasks that are too senior or too critical for the people she's delegating to.
2. Unclear handover. She may be telling people WHAT to do without explaining the standards, context, or decision-rights.
3. Wrong people. The people she has aren't yet capable of the level of work she expects.
4. 'My way' expectations. She may judge delegated work against 'how I would have done it' rather than 'whether the outcome was achieved'.
5. Insufficient time investment. She may not be spending time training and developing the people she delegates to.
6. Reverse delegation. When something looks like it might fail, she takes the work back rather than coaching the person through.

Most founders have ALL of these issues simultaneously. The fix has to address all of them.

Strategic recommendations

1. Start with the right tasks.

• Delegate ROUTINE, OBSERVABLE work first (scheduling, basic client communication, standard project tasks).
• Build up to JUDGEMENT work over time (client strategy, hiring decisions).
• NEVER delegate the firm's core strategic decisions or relationships with key clients — these stay with her.

2. Match task to person.

• Senior consultants can handle complex judgement work; junior staff need clearer instructions.
• Be explicit about the difficulty level of each task — don't give a juniorlevel task to a senior person ('boring') or a senior task to a junior ('out of depth').

3. Be explicit about authority AND outcomes.

• Say: 'You have authority to spend up to £5k on this. You don't have authority on hiring or pricing. The outcome I want is X by Y date, measured by Z.'
• Vague delegations ('look after this client') fail; specific ones with clear boundaries succeed.

4. Train + coach.

• Spend time WITH the person initially — walk through a similar problem together. The investment pays back many times over.
• Have regular check-ins (weekly 30-min one-to-ones) where they can ask questions safely.

5. Accept different ways of doing things.

• The hardest delegation skill: let people do it their way, not your way. If the outcome is met, the route doesn't matter.
• Critique outcomes, not approaches.

6. Don't take the work back when it's wobbling.

• If a delegated task is going off-track, coach the person through, don't grab the keyboard.
• 'Doing it for them' confirms to everyone that delegation doesn't really work here — and prevents the person from learning.

7. Build a small layer of leaders.

• Promote 3-5 people to a 'leadership team' that meets weekly with her. They take responsibility for departments; she sets direction and supports them.
• This creates the structure she has been resisting (no middle layer = founder is the only manager).

8. Track the firm's progress without micromanaging.

• A simple dashboard: weekly revenue, pipeline, active client issues, team headcount. Five metrics, fifteen minutes a week. Replaces 'I need to know everything'.

Trade-offs and risks

• Speed of transition. Going too fast (delegating everything at once) means low quality and brand risk. Going too slow (still trying to do everything herself) means continued bottleneck and burnout. Right pace is 'delegate one new category every 2-4 weeks'.
• Quality variance. Delegated work will be imperfect at first. Acceptable IF the firm tolerates some variance in exchange for the founder's freed time. The math usually works out positively.
• Loss of perfectionism. The founder must accept that 80% of how she'd do it is better than 100% of nothing because she's overloaded.
• Founder's identity. Some founders are emotionally attached to being 'the person who does it all'. Delegation requires letting go of that identity.

Justified judgement and recommended path

The founder should pursue a structured 12-month delegation transformation:

Months 1-2: Identify and promote. Choose 3 senior people to be the first leadership team. Have explicit conversations: 'I'm going to delegate X to you with this authority and these expected outcomes.'

Months 3-6: Practise and coach. Delegate one new category every 2-4 weeks. Spend explicit time training. Run weekly 1-to-1s. Resist taking work back.

Months 7-9: Expand. Add a second tier of delegates (5-10 more people). Build a clearer organisation chart with named roles.

Months 10-12: Step up. The founder moves into more strategic roles — partnerships, key clients, hiring direction, growth strategy. Operational management is genuinely delegated.

Expected outcomes (12 months)

• Founder works 40-50 hours a week (not 80).
• Decision speed across the firm improves (decisions don't all wait for her).
• Two or three 'next generation' leaders emerge.
• Quality may dip 10-20% in some areas — recoverable with experience.
• Firm capacity to take on new clients rises significantly.

Conclusion. Delegation is not a skill that 'comes naturally' — it must be deliberately built. The founder's history of failed delegation has likely been caused by a combination of unclear handovers, wrong tasks, wrong people and reverse-delegation. A 12-month structured programme that starts with the right tasks and the right people, invests in coaching, and accepts imperfect outcomes can transform her role from bottleneck to true leader. The biggest barrier is the founder's own identity as 'the person who does everything' — letting go of that identity is the central work of the transition.

The deeper insight is that founder-driven firms cap at the founder's personal capacity. Beyond 30-50 people, the founder MUST delegate or the firm stops growing — there is no third option. The firms that scale past this threshold are the ones whose founders learned to delegate; the firms that stay stuck are the ones whose founders didn't.

Matrix structures are one of the most ambitious — and most controversial — organisational designs in modern business. The intent is to capture both the depth benefits of functional organisation (deep technical expertise) AND the responsiveness benefits of product/regional organisation (focus on specific outcomes). The execution is brutally difficult, and many matrix transitions have failed catastrophically. For a global pharmaceutical company, the stakes are particularly high: drug development is too complex for any single dimension of organisation to handle alone, but dual reporting creates conflict that can paralyse decision-making. The decision must be made with eyes wide open.

Why matrix structures are tempting

The classic functional organisation (R&D / manufacturing / marketing / sales) has powerful strengths:

• Deep technical expertise — chemists with chemists, statisticians with statisticians.
• Career paths within functions.
• Knowledge management within disciplines.
• Cost efficiency through specialisation.

But it has equally serious weaknesses:

• Poor product focus — each function works on many products; no single owner of any one.
• Slow decisions on cross-functional issues (a drug launch needs R&D + manufacturing + marketing + sales).
• Conflict between functions (R&D wants to perfect the molecule; marketing wants to launch yesterday).
• Customer/market intelligence trapped in sales, not flowing into R&D.

A product-based organisation (drug X team, drug Y team) would solve the cross-functional issues but lose the deep functional expertise.

A matrix attempts to capture BOTH: employees have BOTH a functional home AND a product/regional team.

The full theoretical case FOR a matrix in this firm

• Cross-functional execution. A new drug launch requires coordinated R&D, manufacturing scale-up, regulatory filing, marketing, sales training. A product manager who has all four functions reporting to them (via matrix) can drive integrated decision-making.
• Regional adaptation. Pharma faces different regulatory environments, healthcare systems and pricing across regions. A regional manager (matrix-second-line) ensures local adaptation while the functional line maintains global standards.
• Better resource allocation. Functional managers can flex specialists between product teams as needs change — better utilisation of scarce skilled labour.
• Career flexibility. Employees can develop both functional depth (their functional home) and product/regional breadth (their assignments).
• Modern industry practice. Most large pharmaceutical companies (Pfizer, GSK, Novartis, Roche) operate variations of matrix structures. Functional-only organisations are increasingly rare at this scale.

The full theoretical case AGAINST a matrix

• Dual loyalty / conflicting priorities. An R&D scientist reports to both the head of R&D (priority: scientific excellence) and the head of Drug X (priority: launch date). When these conflict — which they will — the employee is caught.
• Decision paralysis. Both bosses must agree, or the matrix manager must escalate. Many matrix organisations spend more time on internal coordination than on actual work.
• Accountability ambiguity. Who is responsible when something goes wrong? Both bosses? Neither? In a regulated industry like pharma, ambiguous accountability is dangerous.
• Increased management overhead. Matrix structures typically have 1.5-2x the management headcount of functional structures.
• Cultural and political complexity. Matrix politics — managing the two-bosses dynamic — consumes vast amounts of employee energy.
• Implementation risk. Transitions from functional to matrix are 18-36 month change programmes. Many fail or partially fail.
• Real-world examples of failure. ABB famously moved to matrix in the 1990s and dismantled it 10 years later. Many other firms have had similar experiences.

The crucial nuance — matrix is not binary

The choice is not 'matrix vs functional'. There are several variations:

• Weak matrix: Product/regional managers exist but have less formal authority than functional managers. Easier to manage but limited benefit.
• Balanced matrix: Roughly equal authority. Highest tension and complexity, but maximum integration potential.
• Strong matrix: Product/regional managers have MORE authority than functional managers. Closer to project-team structure.
• Hybrid: Some teams in matrix, others functional. Lets the firm experiment without all-in commitment.

For a global pharma, a weak-to-balanced matrix is usually the right starting point — capturing the cross-functional integration benefits while preserving most of the functional clarity.

Implementation considerations

A pharma matrix transition needs:

1. Crystal clarity on decision rights. WRITE DOWN what each manager (functional vs product) decides. Don't let it emerge organically — it never does.
2. Strong matrix managers. Need people who can navigate ambiguity, build coalitions, manage upward in two directions. Many existing managers won't have these skills.
3. HR / performance systems redesign. Performance reviews involve BOTH bosses; promotion decisions need joint agreement.
4. Culture work. Matrix succeeds where employees see collaboration as core; fails in 'win at all costs' cultures.
5. Decision-escalation paths. Pre-defined how conflicts get resolved (often: matrix manager has tiebreaker on schedule, functional manager has tiebreaker on quality).
6. Phased rollout. Start with one drug team or one region as a pilot; learn; expand.

Justified judgement and recommendation

The firm should adopt a weak-to-balanced matrix structure, but phase it carefully over 18-24 months and only after specific preparation:

Year 1, Months 1-6 — Prepare. Define decision rights in writing; train current managers in matrix skills; redesign performance management; identify the first pilot team.

Year 1, Months 7-12 — Pilot. Run matrix structure for ONE major drug team. Track decisions, friction, results. Learn what works and what doesn't.

Year 2 — Expand. Roll out to other product teams based on pilot learnings. By end of Year 2, full matrix in place across major drug portfolios and major regions.

Expected outcomes (positive)

• Cross-functional drug launches accelerate by 6-12 months on average.
• Regional revenue grows as local adaptation improves.
• Functional expertise preserved through retained functional reporting line.
• Employee development broader (both functional depth and product breadth).

Expected costs / negatives

• Management overhead rises 15-25% — additional product/regional managers + matrix management capacity.
• 18-24 months of significant change-management cost and disruption.
• Some senior employees leave because they don't like dual reporting.
• Decision speed initially SLOWS as the new dynamic is learned; recovers and exceeds prior speed by Year 2.

What NOT to do

• Don't move to a 'strong matrix' (product managers dominant) immediately — too disruptive for a 30-year functional culture.
• Don't pretend the dual-reporting tension doesn't exist — manage it explicitly with clear decision rights.
• Don't expect the transition to be cheap or fast — budget £20-50m and 24 months minimum.
• Don't roll out everywhere at once — pilot first.

Conclusion. A matrix structure can deliver significant benefits for a global pharma (faster drug launches, better regional adaptation, integrated decision-making) — but only if implemented carefully with explicit decision rights, strong matrix managers, phased rollout and explicit conflict-resolution mechanisms. A clumsy matrix transition is one of the most damaging organisational moves possible — many firms have lost years of productivity and many senior people through poor execution. The firm's 30-year functional culture is both an asset (deep expertise) and a barrier (resistance to ambiguity). The recommendation is YES to matrix, but treated as a multi-year strategic transformation, not an organisational chart change.

The deeper insight is that organisational structure is the result of a TRADE-OFF between depth (functional) and integration (product/regional). No structure resolves the trade-off; matrix attempts to manage it explicitly rather than choose. This explicit management of trade-offs is the source of both matrix's power and its fragility. Done well, matrix accepts ambiguity and builds skills to navigate it; done badly, matrix pretends the ambiguity doesn't exist and creates paralysis.