The CEO as the cultural epicenter
As a former CEO and senior executive, there was a time when I did not quite understand the profound impact a CEO has on the culture of a company, even though I always knew culture was important.

The organization reflects the behavior and characteristics of the CEO and that establishes the culture. Foster an environment of open communication and the organization inherits a culture of open communication. Operationally detailed? The organization becomes operationally detailed. Political? The organization becomes political. Curse a lot? The organization curses. Angry? The organization gets angry. Have a big office? Everyone wants a big office. It doesn’t matter what’s written on a coffee mug or on a “culture” slide, what you do as a CEO, day in and day out, and how you behave will define your company’s culture.

Despite the best intentions, companies often become culturally dysfunctional. This occurs when leadership has a perception about the culture that conflicts with reality, or leadership behaves differently than what might be written down. 

One of the most studied examples of cultural dysfunction occurred at Enron, the former energy-trading giant. The CEO (Ken Skilling) and several top executives were arrested for a pattern of deceit, dishonesty and illegal financial practices. They promoted a culture of dishonesty, self-dealing and self-enrichment that destroyed the company. Ironically, the Enron code of ethics outlined four key principles: communication, respect, integrity and excellence… So, yes, culture matters and the CEO defines it.

Cultural dysfunction is not limited to large companies. When I arrived at XenSource, which was a 50-person company, the culture was dysfunctional, despite the fact that the founding team believed the culture to be awesome and supportive of innovation and collaborative thinking. There were two telltale signs: 1. Employees painted a very different cultural view from the founders and 2. The responses were inconsistent with each other, indicating that the culture was a free-for-all with very little leadership. One clear example of the inconsistency resulted in the organization having two engineering efforts that competed with each other. Here was a company with a supposedly “collaborative, non-political” culture that had engineering teams pitted against each other to see who would win. The competitive activity turned out to be corrosive and undermined the intended culture of the company.

Stemming dysfunction
I often talk about CEO self-awareness as one of the key attributes of corporate success. In the case of XenSource, the leadership espoused and verbalized cultural “intent”, but practiced and allowed something very different in the company. The company almost failed due to a highly dysfunctional culture. Make sure what you believe is what is truly happening in the company. 

Stemming dysfunction requires leadership and taking some simple but important actions: proactively define cultural attributes important to the organization—write them down and let people know what they are, and “walk the talk”. You must practice and exemplify your culture, and have a mechanism to review culture deep within the organization. Ask the following questions:

  • Is the organization’s culture consistent with the defined attributes?
  • Where are the differences?
  • What are we doing right or wrong to keep a strong and consistent cultural backbone in the company?

The Cultural Paradox: I can’t change the culture because that’s not part of our culture
Culture is formed—whether intentionally or not—in the early days of a company’s life. Activities and behaviors are repeated and these become the elements that shape the culture of the company.  Examples of such early practices might be: 1) The founding team always interviews all new people applying to the company; or 2) a product-oriented focus in everything the company does. The accepted and repeated practices become the culture and define how the company operates.

However, what has worked in the early days might not be as effective as the company grows up. As a result, you might be forced to choose between two conflicting cultural attributes.

Take the attribute “the founding team must always interview new people”—a great cultural practice intended to ensure new employees are a perfect fit. Is there a point where growth is hampered because the company can’t interview fast enough and candidates go elsewhere? What part of the culture do you change? Limit growth or change your hiring practice? Changing either impacts culture.

One of the most difficult aspects for technical founders is hiring outside the comfort zone of the founding team. This is evident when hiring sales, marketing and finance people. A good example of this is how a technical founder might apply engineering hiring techniques to a sales organization, which my partner Ben Horowitz recently blogged about here. The fear here is that bringing on non-technical people will destroy the company culture. Do you put engineers in all the non-engineering functions and continue to only hire technical people, or do you augment the culture and integrate new and different organizations into the company? Here again, sticking to the past practice/culture of only hiring technical people might be counter to building a great finance or sales organization.

Steering change
Existing culture can get in the way of future growth and company leadership must steer the transition. Changes to practices and culture should be done by first asking why something is done a certain way and what’s the intended outcome. Preserving the intended outcome should trump the practice. 

Let’s go back to the example, “the founding team shall interview all new applicants”. The intended outcome is to make sure that all new employees are of the acceptable caliber and intelligence, and understand the culture and origins of the organization. The problem is the system does not scale, particularly as candidates are hired around the world and at a pace that far outstrips the capacity for the founders to handle.

A change to the practice might be to empower key employee “ambassadors” who act as a proxy for the founding team. Alternatively, maybe just one of the founders meets all new candidates as opposed to all founders meeting all candidates. If part of the intended outcome is for a candidate to meet the founders and get a feel for the company, then have all new employees meet the founders at a lunch or dinner after they join the company. Developing a strong and scalable interview process and on-boarding/mentoring system will ensure that the intended culture is preserved while steering change from an operational perspective.

Managing culture
The concept of managing culture may seem a bit heavy-handed, particularly in tech companies that pride themselves on being free from overbearing rules and bureaucracy. However, not managing culture can be likened to not managing growth, or not managing expenses, or simply not managing and certainly not leading…


  1. Self-awareness. If you can’t accept self-awareness, you should not be CEO.
  2. What are you trying to accomplish? What’s the end game?
  3. Translate energy to the areas you are least comfortable understanding.

A strong culture is the backbone of any organization and the CEO is the standard bearer and the agent of change. In a recent Fast Company article, GitHub Co-founder and CEO Tom Preston-Werner shares his perspective on how he and his cofounders have thought about and managed the company culture from 10 people to 160. Regardless of age, background and experience, culture is something that evolves with the CEO and the process of creating a great culture requires leadership to routinely and consistently assess and exemplify the core values of the organization.

Last month, as CEO of SpiderNet, you were faced with the issue of firing your VP of Engineering, who initially seemed like a rock star but proved to lack motivation, didn’t gain his team’s respect, and was ultimately ineffective.  You talked with your board and everyone unanimously agreed that you should terminate the VP immediately and ask your co-founder to manage the Engineering team in the interim.  It was also agreed that the company would search for a new VP of Engineering after the new year.

It is now the new year, the old VP is gone and you are faced with the rather undesirable task of hiring a new VP.  Upon reflection, you realize that the SpiderNet hiring process failed.  Clearly, the issues surrounding the VP you just fired had to do with laziness and you realize now that the SpiderNet process (or lack thereof) did not exist to properly vet out a prospective job candidate.

Before you move on to hiring again, you take an inventory of the current process in an effort to better understand how to correct things in the future.  Here’s what you find:

  1. You identified the requirements of the position and developed a comprehensive job specification.  You feel pretty good about the list, which you developed on your own.  Given that this position reported to you, you did not see the need to ask for other input.
  2. You hired a search firm and saw a number of good candidates, so you felt that the candidate pool was adequate but not exceptional.
  3. The interview meetings were always with you, your co-founder and then one or two other people from the executive team.  Everyone exchanged emails on the candidate after their meetings.
  4. The search firm did the bulk of the reference checks, which significantly helped your schedule and time.

Most hiring mistakes occur when the interview process does not effectively assess candidate fit, motivation, and expectation.  You don’t want to make the mistake again and you need to figure out what to do next.  What mistakes do you think you made and how can you better go about interviewing future candidates?  You are very nervous about screwing up again.

What now?

I have often considered the ability to hire great people as being analogous to a batting average in baseball.  A player who hits .300 is near the top of their game.  While I have certainly done much better than 30% success rate in my hiring overall, I’m not sure that I’ve topped 30% for truly exceptional hires.  Hiring phenomenal employees, like hitting in baseball, is difficult.  I must admit that only recently did I come to realize that there is a set of steps that a company can follow that significantly increase the hit rate.

With respect to the SpiderNet situation, there were several mistakes made with the current process.  First, there was no input from either the board or the other executives on the job specifications.  It is particularly important to gain agreement on the position and have the interviewing team help develop the requirements.  Second, there wasn’t a defined interview team and the results of the interviews felt a bit ad hoc.  Having a defined interview team should be part of the process.  Finally, the CEO left reference checking to the search firm, which was a huge error in judgment and probably the biggest mistake of the entire process.

I took the liberty of asking Jeff Stump, our partner responsible for Executive Talent at Andreessen Horowitz, and he outlined the following steps that should be taken when hiring an executive:

  1. Develop the candidate profile and expected qualifications
  2. Lay out the compensation framework
  3. Craft a set of questions to be used in the interview process
  4. Identify your interview and recruiting methodology
  5. Perform reference checks

Let’s look at each of these in more detail.

Candidate Profile and Qualifications.  There should be a universal understanding of what the profile is and why the company is looking for this new hire.  What is the charter of this hire over the next six months to one year?  What does the company/hiring manager expect this person to deliver during this timeframe and what does success look like for this hire one year out?

The focus in this phase should be on developing a set of non-negotiable attributes as input to the interview process.  Make sure to include input from anyone who is going to have a say in the interview process including the Board and the management team.  There does not need to be universal alignment, but this process will help identify any differences of opinion that can be addressed up front.  This first step will drive consistency in the interview process and candidates will take notice.

Compensation and Reporting Framework.  Determine the title, reporting structure and the compensation of the position.  Not doing this upfront can often lead to misalignment late in the recruiting process.

Interview Questions.  Once the profile and qualifications have been identified and agreed upon, develop a set of questions that are going to be used in the interview process.  These can also be used by the search firm to screen candidates.  Questions should be crafted to gauge motivation, fit and expectation.

During this phase, the hiring manager also should identify the interviewing team and start the selection of a search firm.  In most cases, the company’s internal network will not be sufficient to identify the absolute best candidate and a search firm is recommended.  There are exceptions to this, but in most cases, thinking that the company has enough contacts often results in a substantial delay.

Interview and Recruiting Methodology.  High-level items to consider:

  • Decide on the appropriate team members to be involved in the search process;
  • Determine who is running point on the management team.  Tasks may include interview coordination, candidate follow-up, feedback collection, reference coordination, compensation/negotiation, etc.
  • Who from the board of directors will be involved, if at all?
  • Define the sequence of interviews and meetings.  Who needs to meet when and what is the strategy/objective during a first round vs. a second round of interviews?  I also recommend spending a great deal of the first meeting on determining “fit.”  Can you work with this person?  Nothing else matters if you can’t check this box.
  • Have the team interview for different attributes to create a better experience for the candidate and to avoid overlapping questions in the interview process.
  • Define the feedback loop: What info gets captured, recorded and how/when is it shared?

Reference Checking.  One of the most important—and overlooked—parts of the recruiting process is reference checking.  When you make an offer to a candidate, you should feel as if you really know the person, warts and all.  Reference checking is the best way to really understand what a candidate is like to work with.

Often your search firm will offer to do reference checks.  While it is fine for them to do some investigation, do not leave all of it to them.  You need to do a large part of the reference checking yourself.  Compare notes with the search firm but handing it off to someone else is a huge mistake.

In general, you should do between 10 to 15 reference checks and they should have a 360-degree approach (i.e., 1/3 bosses, 1/3 peers, 1/3 subordinates).  At least one-third of these queries should be backchannel references.

Finally, when you have settled on the final two candidates, you should have the finalists come in to present a 100-day plan with your executive team at your staff meeting.  This will give you a very good sense of how the person responds to an assignment and the nature of their work.

Hiring takes planning and time and the process is often ad hoc or simply non-existent.  Don’t make hiring a batting average.  Take the time, develop a clear process, and you will see much better results in your overall hit rate.

Below are the possible outcomes to the hypothetical situation I posed in my previous blog post, “Founder Re-org”.

When I arrived as the CEO of XenSource in February 2006, I faced a similar (though not exact) situation to what SpiderNet is facing.  At XenSource, we had an incredibly talented founding team, located in Cambridge, UK.  There was another group of talented engineers in Palo Alto.  The problem was that the Cambridge and Palo Alto groups pretty much hated each other.  There was a lack of cohesive leadership and there was risk that either side (or both) would implode, leaving the company with no founders and few engineers.  Not an optimal outcome for a new CEO.

There was no question in my mind that allowing any of the founding team to leave would be a disaster for the company.  There was also no question in my mind that scaling the organization was going to require hiring a VP engineering.  Before the XenSource plane could take off, I needed to switch around a few parts.  In hindsight, it all seems so simple.

SpiderNet (like XenSource) needs to keep its founder and hire a VP of engineering.  In both cases, a change of this sort needs to be done thoughtfully.  Before I made any changes at XenSource, I needed to know more about the organization I was running.  To help better make decisions, I did the following and would recommend this to SpiderNet:

  1. I engaged myself in the cadence of the business: product, sales, strategy, finance.  I met with all our employees, asked questions, and tried to identify areas of weakness.  But instead of stopping after I met with all our employees for the first time, I kept going, every day and every month afterwards, such that I “felt” what was happening in the organization.  I understood our “DNA” and as a result I could make changes that were credible and respected.
  2. I held everyone accountable to measurable quarterly objectives.  While this may sound bureaucratic, I had each member of the team come up with 3-5 quarterly objectives.  It became one of the easiest and most effective ways of managing and reporting on our business.  The result was that I could objectively determine areas of strength and weakness and make necessary changes.

In the case of SpiderNet (and to some extent at XenSource), there are several possible organizational outcomes.  They are:

  1. The co-founder moves to an individual contributor role (CTO, Architect, Evangelist).  He may find this role more acceptable and will appreciate the suggestion that he move to a different position.  This change augments and leverages his strengths as an architect or CTO.  It is important for the co-founder to understand that he will continue to have a key role in shaping the company on a go-forward basis.  This moves allows the company to hire a VP of Engineering, reporting to the CEO.
  2. Keep in the current role and hire an operational engineering VP under the co-founder.  This organizational structure might work, though you need to be certain that the company can recruit a senior enough VP of Engineering or whether the co-founder is capable of continuing to manage the entire engineering organization (even with the help).  I would have the conversation with the co-founder about how together, you will manage this structure and determine if in the future this situation is right for the company.  The new engineering VP will need to feel comfortable working under the co-founder.  I would have the engineering VP sit on the executive team, in any case.
  3. Hire a VP of Engineering above the co-founder.  Probably not a great outcome since the co-founder should have a place at the executive table and continue to have a direct reporting line to the CEO.  I would consider this structure if the co-founder suggested it, though I could not imagine this situation working in the long term.
  4. Leave things as they are.  Bad outcome.  This is the short-term easy way out and not a long-term solution.  It simply moves the problem into the future.  Product quality and scale in the engineering organization continue to suffer.
  5. Co-founder leaves company.  Really bad outcome.  This is the least desirable outcome, given that the company was created by the vision and team dynamics of the founding team.  Sometimes it is necessary to replace a co-founder, but only as an absolute last resort and only if the co-founder is disruptive to the growth of the company or has become a negative influence on the organization.  The SpiderNet facts do not warrant co-founder replacement.

Once you have determined the optimal structure, having an open dialogue with the co-founder will be more helpful than not.  Ideally, you want the co-founder to be excited about the change and your leadership is key in making this transition seamless.  Many times, a person who is struggling in a role may actually feel relieved that you’ve identified a new and better position.

Regardless of the new organizational structure, it would be advisable to ask your co-founder to help hire the new VP and stay in the current role until the new person is onboard.  You also want to make sure that you’ve established the right communication to the company.  My advice is once the decision is made, you need to quickly communicate it to the company.  Best that the information comes from you, rather than through the rumor mill.

Some might suggest “two-in-the-box management” as an option.  I have found this structure easy to implement and something that never works.  It is a nightmare to manage and is a weak organizational structure.  Make the decision upfront, put in a simple line of management and don’t take the easy way out.

Remember also that your board can be helpful in a situation like this.  In many cases, early board members and investors will know the founding team and you may be able to use a specific board member to help with the transition and coaching required to effectively re-position the co-founder.  Never feel bad about using your board to discuss key issues.  In this specific example, your investors already highlighted the possibility of the change when they funded the company, so bringing it up with them is a relatively straightforward discussion.

As for XenSource, we went with an org structure that mapped most closely to option #1 and the rest was history…