Taking the Mystery Out of Scaling a Company

I recently came across one of Ben Horowitz’s classic posts:

Taking the Mystery Out of Scaling a Company

Ben argues that the key scaling challenges are driven three core component becoming much more difficult as the organization grows in size:

  • Communication
  • Common Knowledge
  • Decision Making

Avoiding their degradation altogether is impossible, so what we’re really trying to do is “give ground grudgingly”.- try to slow them down as much as possible using three key levers. Because they all include a trade-off of increased complexity, “giving ground grudgingly” is the right strategy here, and they should be applied with their impact on the three core components in mind.  


It is typically necessary to apply this level first, but it’s also the one with the most challenging side-effects: hand-offs, conflicting agendas, etc. The next two levers aim to mitigate these negative effects.

Org Structure 

There is no perfect org design since there is no way to completely eliminate the negative side-effects of specialization. Organizational design has substantial impact on the company’s communication architecture, both internally and externally – and this is the key to effectively utilizing it, using the following steps:

  1. Figure out what needs to be communicated – key pieces of knowledge and who needs to have it
  2. Figure out what needs to be decided – try to minimize the number of decision makers that need to be involved in making the most frequent and critical decisions
  3. Prioritize the most important communication and decision paths – every org design represents a trade-off…
  4. Decide who’s going to run each group
  5. Identify the paths that you did not optimize
  6. Build plans for mitigating the issues identified in step 5 – typically by applying the next lever:


The purpose of process is communication. It’s a formal, well-structured communication vehicle, meant to ensure that:

  • Communication happens
  • It happens with quality

The people who are already doing the work are the ones who are in the best position to design the necessary process, keeping a few best practices in mind:

  1. Focus on the output first
  2. Figure out how you’ll know if you are getting what you want in each step – usually via some form of measurement
  3. Engineer accountability into the system – which organization/individual is responsible for each step. Make their performance visible.
Taking the Mystery Out of Scaling a Company

The Vice President of Business & People Operations

This post builds on a previous post in this blog, where I made the case for having architects for organizational systems. Under this paradigm, business outcomes hinge on the interaction between 5 sub-systems: org structure, incentive systems, work systems, collaboration systems, and people systems.

The Table GroupPatrick Lencioni’s consulting group, argues that “organizational smarts” – strategy, technology, marketing, finance, etc. – is no longer the driver of competitive advantage, since intellectual ideas are not a sustainable differentiator in an age when information is ubiquitous. Therefore, what truly matters is “organizational health” – minimal politics, minimal confusion, high morale, high productivity, low turnover, etc. I’d argue that even “healthy” is not enough. What we really need are organizations that are “fit” to deliver on their mission/business outcomes. All athletes/top performers need to be healthy. But a weight-lifter and a cyclist need to be fit in different ways in order to be the best in what they do.

Which brings us back to the 5 systems. Organizational fitness is reflected in the way these systems work in concert to drive the specific business outcomes an organization is trying to accomplish. And, without falling into a recursive trap, building and maintaining organizational fitness is then a business outcome in and of its own. Therefore, we must ask ourselves: since org design is a key system that enables business outcomes, are businesses organized in a way that enables them to build and maintain fitness?

The good news for the Table Group, and the bad news for almost everyone else, is that the short answer is: no. We know from Conway’s Law that the way we divide and assign responsibilities in the organization has a material impact on its function. A typical organization has departmental heads for all the “organizational smarts” disciplines: a head of marketing, a head of technology (R&D/CTO), a head of finance, and so on. Yet “organizational fitness” is not expressed in the organizational structure in any material way. Responsibility over the 5 systems is typically no-one and everyone’s job, so it’s no surprise that many organizations are not organically poised to build and maintain a high level of organizational fitness. The one exception to this rule is people systems, which have a well defined owner, at least in theory – the head of people. But that’s a story for a different blog post altogether.

So perhaps it’s time for organizations to revisit the traditional structure of the executive team, and consider whether an alternative structure can better express the growing importance of organizational fitness as a critical discipline driving competitive advantage.

The Vice President of Business & People Operations


Henrik Kinberg‘s seminal work at Spotify has made real waves in the community when Scaling Agile at Spotify started making the rounds. More recently, Henrik put out these two videos outlining the key components of Spotify’s engineering culture:

Spotify Engineering Culture – Part 1

Spotify Engineering culture – Part 2

The thing I found most impressing about these videos is the connection between the high level, abstract culture, and the more pragmatic set of business practices that it was broken down to.

As Henrik mentions in a different post, there is no right or wrong answer here. Every cultural decision has a trade-off built into it:


What’s really unique about the Spotify case, is the way they acknowledge and own up to the trade-offs that their culture embodies, and fully align their business processes with it. This tight alignment between culture and process/structure is rare, and in my opinion, one of the key ingredients in what makes Spotify so successful.


Book: Reinventing Organizations

Reinventing Organizations by Frederick Laloux is so far the best book I’ve read in 2014. Which is worth noting given that I average about 20-30 non-fiction books a year.

Laloux applies a developmental psychology lens to looking at organizations. As it turns out, every major shift in the way we, humans, viewed the world, was accompanied by major innovations in the way we collaborate and work together, as summarized by the (modified) table taken from the book below:


The vast majority of organizations today operate under an “achievement” paradigm, but a shift from “achievement” to “pluralistic” is already a noticeable trend in the industry. Laloux, however, chooses to focus on the next paradigm shift: from “pluralistic” to “evolutionary”. He studied an impressive set of organizations who have already made or are in the process of making the paradigm shift, and distilled the commonalities into three major innovations: self management, striving for wholeness, and evolutionary purpose.

But not stopping there is what turns this book from an average to a very good one. For each of these innovations, Laloux provides detailed descriptions of how the core organizational structures, processes and practices change as a result. From decision-making, through hiring, firing and promoting, all the way to the squishy topic of organizational culture (I’ll cover the latter in a separate post).

For example, he provides some of the most inspiring and detailed descriptions of self management, a term that often gets aimlessly thrown around in agile environments. In the process, he debunks, through real examples rather than theory, some of the most common misconceptions around this concept, like the belief that the lack of hierarchy automatically means that there’s no structure, no management and no leadership.

The book does have its weaknesses. The forward, by Ken Wilson, is the worst part of the book and almost made me miss out on a phenomenal read. Referencing the organizational paradigms by their colors rather than their one-word descriptions was distracting at best. But most importantly, organizations with long and deep value chains, like software companies, are not getting a lot of attention in the study and the book. Laloux acknowledges in the appendix that they require certain adaptations, but Holacracy, the one known “operating system” for such organizations, is only covered anecdotally in the book.

As an aspiring evangelist of Laloux’s thesis, the things that I’m missing the most are shorter materials that I can use to pique the interest of a boarder yet-to-be-engaged audience, for which reading a 300+ page book is asking too much. A 10-page HBR-like article or a 20-min TED-like video talk will be great.

Get it. Read it. And tell me what you think.

Book: Reinventing Organizations

Functional vs. Divisional

About a year ago, Ben Thompson wrote a great blog post comparing and contrasting the strengths and weaknesses of divisional organizations and functional organizations:

Why Microsoft’s Reorganization is a Bad Idea

As you can probably tell from its title, the context was Microsoft’s reorg from a divisional organization to a functional one, but that’s outside the point for this post.

I keep going back to this post every time we’re considering a reorg, asking myself whether this moves us closer to the divisional or functional pole of the spectrum (most companies, especially in lower tiers, are hybrids) and therefore, how we should mitigate the risk in the trade-off.

Massive, enterprise-scale companies, truly have the option of choosing one over the other. But smaller companies don’t. Most small to mid-size companies are organized functionally, since at that scale, it is clearly the more effective way to organize. But it still has its shortcomings.

As Ben suggests, accountability in functional organizations is less clear. It’s also more likely that despite the overall structure, some people will be holding more divisional responsibilities (Product Managers are a great example).

More on one approach for dealing with this challenge in the next post.

Functional vs. Divisional