Org Design

Org structure might be causing your company's dysfunction


Have you worked for a company that constantly reorganizes itself ("re-orging")?

  • In one year, your company might centralize to cut costs and increase control, only to decentralize four years later to boost adaptability and innovation. Bayer, for example, just announced their decentralized re-org.
  • Or one year, your company shifts its power structures to the Sales team to be more revenue-driven. Four years later, the company shifts its power structures to the Product team to be more value-driven. This is similar to what Bob Iger did when he returned to lead a struggling Disney.

Organizational structures are crucial to an operating model, yet constant re-orging significantly drains motivation and performance. So what's an executive team to do?

This article unpacks five critical first principles for building a motivating, adaptive, and high-performing org structure.

  1. Form follows function (i.e., vision and strategy come first)
  2. Reporting lines are powerful enough
  3. Strengthen the horizontals
  4. Focus on feedback loops
  5. Structure is a part of a system

Principle 1: Form follows function

Imagine swimming in a river with a strong current; going with the flow is easy, but swimming across or against it is challenging. Organizations have currents too.  

Created by power and communication structures, this current makes organizations adept at solving aligned problems and inept at addressing opposing ones.

As one CEO we work closely with put it, "My company solves problems well with one silo involved. With two, the odds drop. With three, we don't succeed at all."

This CEO's point is that problems "against the current" are less likely to be solved. This is why "Conway's Law" is shockingly accurate.

Conway's law states that an organization's problem-solving resembles its structures, power dynamics, and communication flows.

  • For example, company websites often reflect the company's organization, not user preferences.
  • Another example is the early Microsoft Office, where different applications—Word, Excel, PowerPoint—had varied interfaces and compatibility issues because teams were organized by application, not function.
  • The P&L dynamics at Xerox and Kodak contributed to their inability to exploit their innovations in personal computing and digital photography.

Many organizations view Conway's Law as a cautionary tale for diagnosis. However, Conway's Law should serve as a design principle for competitive advantage.

Think about it this way: if your company and a larger competitor share the same org structure, you're likely to lose because you're solving problems in the same way.

To take advantage of Conway's Law, first start with your vision and strategy.

  • Consider your vision as a description of your organization in five years.
  • Think about your strategy as the critical problems you need to solve to get there.

Then, ask yourself if your org structure mirrors your problem solving structure. Are these critical problems "with your organization's current," or "against the current," significantly lowering your odds of success?

Instead, more often than not, we find companies whose org structures look almost exactly the same as their competitors (often driven by poaching talent). Teams are often designed by carving up products or processes, regardless of how much effort is warranted. For example:

Company situation From-state To-state
A large industrials company is realizing that scripted catalog sales is obsolete as customer buying behavior changes. They believe they need solutions-based selling in an era of hybrid work. Regionally oriented sellers focused on in-person local coverage. National industry oriented sales teams focused on building deep industry expertise to provide consultative support during sales calls.
A fintech org is realizing that authentic content-based marketing is a critical part of their customer acquisition strategy. Strategy, Product, Implementation, and Marketing teams all had independent content research teams. Single content research team responsible for value proposition and content-marketing research. 
A major financial institution is seeking to build a generative AI-first user interface for mobile banking. Teams are aligned around product categories (checking, savings, wealth, credit cards), each with incremental AI projects. Single AI interface team working across all product categories, with smaller product category teams (with a pause on additional user interface improvements).

A good litmus test is whether an expert can predict your success drivers by looking at your org structure. If not, Conway's Law becomes a curse, not an advantage.

Principle 2: Reporting lines are powerful enough

We once led a cultural transformation in a major financial institution's product and technology functions. While observing a tech team, I witnessed the product manager and technology lead get into a vicious argument, personally attacking each other. Why, when in theory they should have been working as a team? In this case, the technology function and product function each had separate OKRs, which led to these two colleagues having different priorities. Those OKRs were enforced with a high-pressure performance review system. These two colleagues were thus a team in name only.

Reporting lines already create a strong current based on how much your boss can influence your career.

However, organizations add "power-ups" to this reporting line, further strengthening the company's primary current. For example:

  • Siloed OKRs
  • High-pressure ratings
  • Promising executives "fiefdoms" when you hire them
  • P&L ownership

Implementing "power-ups" strengthens the current to a point where cross-functional problem-solving becomes overly challenging. It is best to be cautious when pulling these levers.

Principle 3: Strengthen the horizontals

Are there groups in your organization who consider themselves "second-class citizens"? If so, you likely have an imbalanced org structure.

Even if you follow principles 1 and 2, there will still always be critical issues that are "against the current" of your organization. For example:

  • Imagine if Amazon's warehouses are primarily run by General Managers that have near complete control over their domains. If the HR team wanted to implement a compensation change, that would be "horizontal".
  • Imagine if JP Morgan Chase were primarily organized by their main lines of business (e.g., Retail Banking, Wealth Management, Private Banking) but they also wanted to implement a common rewards program across all businesses. That program would be "horizontal".
  • Apple made a big bet to insource their own CPU (i.e., the main microchip) design across many of their product categories. This team is horizontal to the main structure of the organization.

In many organizations, these "horizontals" are more important than their influence suggests.

In 1973, sociologist Mark Granovetter published an influential paper introducing the Weak Ties theory. It suggests that the loose connections we have with acquaintances (i.e., "weak ties") can be more valuable for accessing new information, opportunities, and ideas than the strong ties we share with close friends and family. Unlike strong ties, which often circulate similar information within a tightly knit group, weak ties serve as bridges to different social circles and networks, thereby facilitating the flow of novel information and opportunities.

A similar concept can be found in corporate valuation through the capital asset pricing model (CAPM). The model demonstrates that companies whose stock prices are correlated with the market (i.e., with the current) have higher beta and should be worth less than comparable companies whose stock prices are not correlated with the market (low beta or against the current).

Both these theories are implications of a deeper principle: information has value that is related to its information beta.

  • Information that comes from the current of your organization (i.e., aligned to your primary structures) has high information beta and is thus less valuable. This information tends to be easy to get and is often not that surprising. A boat that needs to go with the current isn't that hard to operate!
  • Information that is not aligned to the current of your organization (i.e., patterns that are horizontal) have low information beta and are thus more valuable. This information tends to be hard to get and is often more insightful. A boat that needs to go against the current is much harder to operate.

This is why horizontal structures are critical in your organization, often being the source of your next innovation wave when vertical structures (your current) are depleted.

Many ways exist to strengthen these structures, but the easiest ensures they have their own currents, equally important to the organization. Steve Jobs applied this principle when he returned to lead a near-bankrupt Apple. As the Harvard Business Review put it:

"When Steve Jobs returned to Apple, in 1997, it had a conventional structure for a company of its size and scope. It was divided into business units, each with its own P&L responsibilities. Believing that conventional management had stifled innovation, Jobs laid off the general managers of all the business units (in a single day), put the entire company under one P&L, and combined the disparate functional departments of the business units into one functional organization. Although such a structure is common for small entrepreneurial firms, Apple—remarkably—retains it today, even though the company is nearly 40 times as large in terms of revenue and far more complex than it was in 1997."

Principle 4: Focus on feedback speed and quality

I once asked a very successful tech founder if the company's founding team was more or less skilled than his average product team is today.

He said, "We hire the best and brightest in our region. Our average product team is far more skilled than our founding team."

I replied, "So then, why is it that your founding team created billions in value, and they have not?"

Of course, there are a few answers to this question that should not be ignored just for the sake of good rhetoric:

  • Survivorship bias - We're not taking into account the sheer number of founding teams that have failed.
  • Motivation - For many reasons, this founding team was more motivated. We won't cover this now, but to learn more about the science of motivation and performance, read Primed to Perform.

But a critical part of this answer is often neglected in org design: feedback loops.

Imagine you’re driving a car down a winding road. On both sides of the road are sheer 100-meter cliffs. Now imagine if you’re only allowed to touch the steering wheel once every ten minutes. How fast would you drive the car? Obviously, you'd drive really slowly.

A fundamental principle is that an entity's velocity—whether a person, machine, team, or organization—depends on the frequency and quality of its feedback loops. When feedback loops are low frequency or low quality, teams will make costly mistakes, waste their effort, or just move super slowly out of risk aversion and bureaucracy.

When his organization was just the founding team, their feedback loops were amazing. They all shared their information with each other every day. They were all close to their customers. They reacted quickly.

But now, the organization with tens of thousands of people has low-frequency and low-quality feedback loops.

  • The best feedback often lies with the sales and implementation teams (who often don't know what to look for).
  • They share their learnings with their boss, who has to share it with their boss, eventually leading to the head of sales.
  • The head of sales shares these learnings with the head of product, who now has to cascade them down to their frontline colleagues.

This chain leads to feedback that is so slow that the whole organization is slow. It's a massive game of telephone. Solving these problems often requires cross-functional teams at the frontlines who have direct access to their customers.

Principle 5: Align to strengths and passions

Toyota built what we would describe as an organizational miracle—a highly effective, and motivating, continuous improvement ecosystem in production lines. Apple and Samsung built another organizational miracle—a product apparatus that is able to simultaneously build great standalone products while also enhancing all their other products.

But to do so required deep domain expertise in each type of work. Org structures must consider this in their design.

There are four primary types of work where deep domain expertise is required for high-performance leadership:

Principle 6: Structure is part of a system

As the world has become more volatile, and as AI is increasingly used to automate tactical work, human work will inevitably become more adaptive. Many companies, like Bayer and Apple, have realized this and have embarked on their own structural transformations to build more adaptive organizations. Then why do most of these kinds of transformations fail?

Changing only structure without altering other elements will likely worsen performance, not improve it. Operating models are like ecosystems. If their pieces are inconsistent with each other, they will each diminish each other's impact. For example, here are common inconsistencies we see when organizations aim to build adaptive structures:

Operating model element How this element goes wrong when building adaptive org structures
1. Vision & strategy
  • Without a well-understood five-year vision, adaptive teams will feel chaotic and rudderless.
  • Annual, finance-led, top-down strategy cadences are not compatible with high-performing, adaptive organizations.
2. Structure &  roles
  • Decentralization typically requires more cross-functional teams, which means that the role expectations of leaders at every level will need to change.
3. Performance management / performance cadences
  • The typical quarterly review and annual ratings performance management system is not effective at managing faster-paced adaptive teams.
4. Supportive leadership
  • Decentralized structures need leaders who learn how to be supportive problem-solving coaches, not taskmasters.
5. Talent system
  • The talent systems of decentralized orgs typically need to support fewer but higher-skilled colleagues.
  • Compensation systems often hold leaders accountable to legacy behaviors, not new ones.
  • Often talent systems prime employees to think the most powerful (and thus valued) roles are in centralized enterprise functions.

Don't wait...

If accomplishing tasks feels like swimming against the current, your org structure might be to blame. One study found that "83% of employees indicated that silos exist within their companies, and 97% saw these conditions as having a negative effect on performance."

If you feel stuck or want an expert opinion, don't hesitate to reach out. We're happy to help you think through how to build a high-performing culture.

Originally published at:

Neel Doshi

Neel is the co-founder of Vega Factor and co-author of bestselling book, Primed to Perform: How to Build the Highest Performing Cultures Through the Science of Total Motivation. Previously, Neel was a Partner at McKinsey & Company, CTO and founding member of an award-winning tech startup, and employee of several mega-institutions. He studied engineering at MIT and received his MBA from Wharton. In his spare time, he’s an avid yet mediocre woodworker and photographer.

Read full bio.

Lindsay McGregor

Lindsay is the co-founder of Vega Factor and co-author of bestselling book, Primed to Perform: How to Build the Highest Performing Cultures Through the Science of Total Motivation. Previously, Lindsay led projects at McKinsey & Company, working with large fortune 500 companies, nonprofits, universities and school systems. She received her B.A. from Princeton and an MBA from Harvard. In her spare time she loves investigating and sharing great stories.

Read full bio.