Bayer’s Bill Anderson: Turning a 168 Year-Old Tanker Like a Speedboat
In a Nutshell
Bayer CEO Bill Anderson flattened 11 management layers to 6-7, boosted spans of control to 14-90 direct reports (called "span of coaching"), and scrapped annual budgets for 90-day cycles with dynamic resource reallocation, enabling a 100,000-employee giant to act like a nimble speedboat amid patent cliffs and regulations. Key to avoiding bureaucracy: hire missionary zealots over professional managers, use peer feedback decoupled from compensation for performance reviews, prioritize outcomes over outputs in jazz-like teams that form/dissolve rapidly, and resist hierarchical "orchestrification" when scaling from 100 to 1,000 people. Advice for CEOs/boards: external hires should aggressively reboot orgs; aspiring leaders admit knowledge gaps and seek peer input to build trust.
These notes were generated by AI and may contain inaccuracies.
In a 30-person organization, you see an opportunity and just do it. Almost every large organization has a bureaucracy problem, but things often don't happen due to too much overhead.
Bill Anderson runs Bayer, a company founded in 1863 that invented aspirin. It's a 160-year-old pharmaceutical and agricultural giant with 100,000 employees worldwide when he joined. Since becoming CEO 2 years ago, he has torn down and rebuilt operations: flattened 11 layers of management to 6-7 layers, expanded some managers' direct reports from 60 to 90, and eliminated the entire budgeting and planning process.
This is a real-world case study on scaling without becoming sclerotic, staying agile at massive size, and avoiding the bureaucratic death spiral. Key lessons: bureaucracy arises from the organization's composition, not as an infecting virus; staying nimble requires hiring the right people; earning trust means admitting when you don't know what you're doing. Bill navigates this amid patent cliffs, regulatory challenges, and shareholder pressure to turn around the stock.
Many companies have specific problems, but large organizations universally suffer from bureaucracy. Bill noticed this 25 years ago as a new manager in 1999, fighting bureaucracy by simplifying and prioritizing. 15 years later, despite efforts, bureaucracy worsened—like treating a disease that gets worse with treatment, requiring a strategy rethink.
At HubSpot, growth from smaller sizes led to less getting done. Bill experienced this scaling from 30-person to 700-person enterprise, merging to 5,000, then to 10,000 and 100,000. In small teams, opportunities lead to immediate action over lunch. In larger ones (e.g., 700 to 5,000), it requires talking to 7 people, manager approvals, follow-up meetings—overhead prevents action.
At Bayer, eliminated 11-12 layers to 6-7 layers. CEO has 12-14 direct reports. Average span of control increased from 6.5 to 14 (still rising), with some managers at 20, 30, or 90 direct reports. They call it span of coaching.
Bureaucracy isn't a virus infecting a healthy organization—it's created by the structure: 10-12 layers, functional org charts instead of customer/product focus, signoffs needed from layers removed from product/customer. Must remove system parts causing it.
Typical reorgs remove 2 layers from 12, achieving little if hierarchical functions, annual budgets trapping money in thousands of cost centers persist.
With wide spans (e.g., 5 to 90 direct reports), managers can't do one-on-ones, annual performance plans, hiring/firing as before—job changes fundamentally. Command and control impossible in flat structure. People own the business (engineers, agronomists, doctors, lawyers, salespeople); managers grease skids, remove bottlenecks.
Eliminated annual budgets; now 90-day cycles organization-wide, including leadership. Every 90 days: evaluate last 90 days, plan next 90 days (one day spent), assess if fewer people needed, reallocate 10-15% of organization to other teams. Teams collapse/reform: durable ones (e.g., 5-year molecule launch) persist; others form for 90-180 days then dissolve.
Businesses lack a fixed score like an orchestra—more like jazz than Beethoven. Investors care about results (e.g., earnings calls), not process. Large organizations gravitate to orchestra-like structure: layers, long planning, strict budgeting—Bayer broke this.
A well-run hierarchical bureaucracy can deliver okay performance. Analogy: Hierarchy Hill—two 10,000-person orgs where top 100 call shots, rest are order-takers (unmotivated, no innovation due to layers). One with execs focused on golf performs worse than one obsessed with customers/products/technology (missionaries vs. mercenaries).
Early HubSpot employees (first 10) joined for mission/colleagues, worked intensely; later hires differ, more mercenary with life priorities—natural progression.
At 100,000-person Bayer, nearly everyone capable of missionary zeal about something. Example: Bill's brother-in-law, a cop with free time, organized church teams to rebuild tornado-destroyed houses in 2 weeks with 15 people. Create environment where all-in on Bayer mission is normal, not weird (fighting city hall).
For CEO scaling from 100 to 1,000 employees: Don't hire professional managers. VCs push to "uplevel team" post-Series B—resist. Amazon kept same people from 5 to 50,000, avoiding dull mediocrity. Professional managers trained in bureaucracy implement it—it's systemic.
If early finance lead can't scale from 10- to 1,000-person org, find dynamic non-bureaucratic replacement, not professional manager.
Both speakers are MBAs from Sloan School; Bill loves education/lifelong learners, doesn't focus on MBAs. Bayer's 6-person executive team: he has MBA, unsure about others—irrelevant for senior hires.
VCs push annual budgeting cycles (plan September for January year), but world changes by July—doesn't work.
Bureaucratic fix: pin down every dollar in advance by department/activity—logical but fails as world changes.
Bayer's two-tier dynamic resource flow:
- Tier 1: High-level commitment (e.g., spend $20 million next 12 months, deliver big outcomes)—nailed down for investors.
- Tier 2: Resource pool (e.g., $20 million across 5 program teams + 1 admin team)—rough starting allocation, not tied to bonuses. Every 90 days, check in: treat money as own, drive products/prospects. Reallocate flexibly (e.g., Team A stuck, shift resources).
Team A hits a roadblock waiting for regulator feedback that takes six months. Shift all people from Team A to other teams and spread their money across those teams. Team B identifies a new opportunity to do more; allocate some Team A people there. Teams sort this out themselves, not the boss deciding.
At year-end, no one is a hero for spending less than allocated (e.g., 5 million budgeted, spent 4 million). Performance evaluated in peer review sessions.
Tip 1: Careful of Hiring Professional Managers
Tip 2: 90-Day Budgeting Cycles
90-day cycles, not yearly. Not basis for performance management. Funds available, but aim to spend less and make more progress. Evaluate later.
For 100-person company, avoid complexity creep. By 1000 people, bureaucracy creeps in. Going from 100 to 1000, key decision: make managers responsible for performance management or draw on power of peers and doers? This decides startup feel vs. hierarchical fork.
Usually, scaling to 1000 brings in professional managers around 500 people. Up to 300-500, informal works.
Performance reviews every 90 days. With 1000 people, 15-40 managers, each with 25-50 direct reports. Managers not reviewing/approving everything; hire smart people who don't need babysitting, all on teams.
Every 90 days, feedback from peers worked with most:
- 1: Great job, love working with you, great impact.
- 2: Off the charts, clearly performing above expectations.
- 0: Needs improvement, not good enough.
Two questions: 1) Major impact last 90 days? 2) What could they have done better?
People care more about peers' opinions than boss's.
Do not use peer ratings directly as bonus multiplier (e.g., average 1.2 vs. 1.3 vs. 0.9 over 40 ratings). Weaponizes feedback; leads to reciprocity (I give you 1 if you give me 1) or bias (they need the money).
98/100 say peers have better view of impact than boss, but most say boss should evaluate. Humans not logic machines.
Implementing peer assessment worldwide at Bayer (90,000 employees, down from 100,000+; most eliminated positions were management). Encourage deep breaths, peers know better than managers. Goal: grow and improve via peer feedback. Hard break between peer feedback and comp.
Manager (e.g., of 50) uses eyeballs, peer feedback, qualitative/quantitative data, team deliveries to decide comp.
Titles not a big topic at Bayer. If need title to get job done, talk to manager. Spending time on titles doesn't advance science or customers.
Tried no titles at HubSpot (~50 employees); lobbied back due to external perception (e.g., family asking about promotions) and sales hierarchy signaling. Napoleon: amazing what soldier does for color ribbon.
Bayer has org chart but puts it "at back of room"—not important. Org chart doesn't determine pay.
Tried influence chart at HubSpot (email analysis); most influential was first intern Brad Coffee, not CEO.
Regret getting talked out of counterintuitive org ideas.
Titles/no titles both CEO control; better: titles don't matter much, trust people. Not "hire great people and turn loose" (fails, leads to chaos).
Every 90 days, team retro: How did last 90 days go? Achieved 3 outcomes? Why not? What went wrong? (1 hour, not navel-gazing.)
Next: Most important for next 90 days? Still comfortable with long-term vision (e.g., launch product X with features by date)?
Outcomes, not outputs: Not "sell more," but "improve product performance by X%" or customer need.
Everyone on same 90-day cycle; if team dissolves, people go to forming teams.
Not control vs. decentralization; stiff backbone, arms/legs go wild.
Over a year ago, made massive changes quickly. Right speed: fast due to existential threats (patent expiries, technical challenges), but Goldilocks (baby bear's porridge).
Bayer people passionate about mission/science/people, but "can't get anything done, too bureaucratic." Bureaucracy unacceptable amid challenges (people tolerate if sales grow 10%). Couldn't go slower.
Few turnarounds like this; most similar orgs built from ground up or early.
Specialty: established orgs. At Genentech (CEO second half 2016), bureaucratic despite innovation rep. Mechanics broken (levels, governance). 10 years prior non-stop anti-bureaucracy efforts failed. Insight: body designed bureaucratic.
Then CEO Roche Pharma (Genentech parent), 6-7 years practice before Bayer.
Insights from others, not solo. Book Reinventing Organizations by Frederic Laloux (teal orgs). Buurtzorg (Netherlands, 16,000 employees, 2 managers, 30 coaches, 30 back-office; rest individual contributors).
Buurtzorg shock: performance problems, team splits, evaluations done by employees, not managers. Thousands of similar home health care teams, somewhat independent.
Adapted to Bayer: complex pharma needs coordination across 2-3k people over 12-15 years, interdependencies. Not 2 managers, but more support/coordination.
Admired CEOs (Jobs, Gates, Bezos, Musk, Huang): thick skin, alligator-like.
Those are product innovator types: great leaders but weak org managers sometimes (needed others to make org work). Lifelong learners, compulsive/obsessed, improved over time on CEO craft.
Decide what to be: Bill is nerd/engineer (chemical engineering masters), loves product, but not modeling product visionaries. For 90,000 people, wants 1000 product visionaries.
Need product visionaries coming through. Not talking about being a professional manager. Must have a picture for what you're going to be. Thick skin but internally referenced while socially sensitive.
Internally referenced vs. externally referenced: Externally referenced people reflect those around them—if everyone else is happy, they're happy. Internally referenced people have their own internal compass.
Internally referenced versus externally referenced. Some people... ask them how are you doing today? They say well how are you?
From neuro-linguistic programming, can understand people better. Example: time constant—what's your time constant? For some, long time is a century; for others, six months or a month.
Being externally referenced good for empathy, not missing birthdays, not offending. But hard for making hard decisions or changing things. Speaker is internally referenced but cares if somebody says "I hate you." Got hate mail from customer venting about a product, wrote back politely, they apologized.
Hey, I'm really sorry about this, but I don't think what you said is quite fair.
Ethics: cares what people think; personality: pretty internally referenced.
Advice for CEOs in place for decades thinking of retiring, and boards hiring: internal or external?
Trend in Silicon Valley: internal (Google, Microsoft); some external worked (Uber). Speaker became CEO at young age, grew with company; has done almost every level job—Bayer had 12 layers, worked at each. Asset, but can age out.
Nine out of 10 large companies hire inside; external hire signals crisis, poor succession planning. Speaker got first public CEO job at 56, was ready earlier but didn't get it. Seen big companies hire internal who were settling.
You got to hire someone who's awesome.
CEOs often pick number two who thrived in their shadow, assume they'll maintain what was built—but no maintenance mode. Either changing or dead.
New external CEO: get to know the company and business.
External hires should tear house down and rebuild org structure/management—something's wrong. Be aggressive, make large change, not incremental. Examples: speaker at Bayer; Dara Khosrowshahi at Uber—changed culture right out of the gate.
Playbook same for internal: Yamin Rangan took over HubSpot from speaker, did great but advice now would be make big change right away. Caretaker CEOs are death.
Almost every organization needs reboot:
- In crisis: quickly know people, trust who to trust, assess, high-level conclusions, move fast bringing organization along.
- After successful 10-year CEO: never think just keep going. Organization reflects predecessor's flaws (inaction, swept issues, bloated, slowed decisions).
Employee engagement high but can't get things done = content organization, dangerous. Signs look good but needs reboot.
Talk to lots of people, ask experienced leaders for time—most later-career people glad to help.
Trust people who work for you to shape and train you. Early mistakes: thinking supposed to have all answers, be invulnerable. Be confident but open/approachable.
This company's really important. Our mission is great. Our technology is great, but boy, I could really use your advice. I don't think I'm that good of a leader, and I'd really appreciate any input you have on how I can be more effective at helping the company.
Pulling bureaucracy from 100,000-person org; goal: scale from 10 to 1,000 employees without bureaucracy slowing pace.
Things that slow companies:
- Annual planning/budgeting: do shorter cycles, six-month increments (esp. with AI).
- Layers: director layer slowed HubSpot; push off as long as possible, span of coaching 1:10 min (not 1:3).
- Mercenaries: manage missionary-to-mercenary ratio, push out as long as possible but will need some.
- Org charts/titles: tried banning at 50 employees for 6-9 months, went back—time spent exhausting but upside outweighs downside.
- One-on-ones: jam CEO calendar; prefer Jensen Huang style—no 1:1s, 50 direct reports, feedback in public/large meetings. Regret switching to traditional 1:1s at HubSpot.
- Compensation/bands: eventually need; decouple from org position (e.g., Zuckerberg hires $100M+ devs > VPs; ICs can > VPs value).
Professional managers: push to post-100 employees, build with early crew/homegrown first. Eventually need (VP Sales ~100 employees, GC ~500, CFO for IPO). Avoid from ginormous companies—impedance mismatch.
Keep Sequoia Capital in your library
Save the videos and channels worth coming back to, and find them again in one place.
