Making Culture Tangible – a 15-year Learning Journey

Culture is an elusive concept. We confidently describe company cultures with terms like “innovative” or “operationally efficient”, but rarely can we put our fingers on the specifics to be able to leverage and improve our own culture. Culture is truly the only unique thing companies have. Most other things like strategies and products can be copied. If we want to be able to borrow learnings from other cultures or improve our own, we need to make culture tangible.

I started over 15 years ago trying to figure out how to bring an Agile culture that was prevalent in many small tech start-up companies to large enterprises. I had been exposed to Agile practices at a small start-up I led.  I was fascinated at the energy and productivity of small, self-directed squads that were loosely coupled but tightly aligned. Embarrassingly, my product development team of 40 was far more productive than my prior experiences as CIO where I had thousands of employees.  After the start-up experience, I started a CIO role in 2007 at Suncorp – a large bank / insurer in Australia.

My experience at Suncorp

I had a dream to see if I could create a culture that had the innovation of the best tech companies and the scale and opportunity of a large bank/insurer. The concept we developed at Suncorp is called “operationalizing culture”. We wanted to make an Agile culture something tangible that you could learn, practice, measure and improve. Our strategy was to break culture up into a set of practices. We could then train, practice and measure the maturity of those practices across our company and then make course corrections and improve.

Our metaphor was a great sports team. In this case, we chose Manchester United (Man U) in the EPL and its leader Sir Alex Ferguson. We did this because Man U had demonstrated sustainable excellence over a long period of time with both a great leadership system and high talent density.  

Our strategy had 4 key components:

1. Playbook of practices: 3 key sets – Leadership, Collaboration and Delivery
2. Team members that care about their craft and are willing to practice/build skills both individually and as a team
3. Coaches that help train, give advice, course correct
4. Keep score to ensure we are getting better

A key learning from reviewing Man U was just how much time Sir Alex Ferguson spent on recruiting and ensuring he had the right talent and chemistry to win in the marketplace of the EPL. He constantly evaluated his team and made adjustments throughout the season and after the final results. He also made sure the Playbook he used matched the talent he put on the pitch. The bottom line: he cared about attracting and building great players, listening and learning from others, training hard, and keeping score.

Another key learning from Man U was appreciating how much learning and chemistry is built from teams staying together. In the IT world, we form “projects” where we routinely pull people together to accomplish an initiative. We treat people like identical objects that come together with seamless chemistry and are productive right off the bat. The reality is quite different. In most cases, by the time the project team learns together and gets productive, they are broken up to form other project teams – and in many cases – at the height of their productivity! We would never do this to a world class sports team. Great teams like Man U have replenishment strategies that happen on a continual basis to course correct and improve over time – but rarely is there wholesale tearing apart of teams and reforming by scratch. That is a recipe for disaster.

We changed our strategy to create “product” teams which would form around an application set or business problem. When new initiatives would come up, instead of forming new teams to take on a project, we would feed the work to existing product teams that would manage the work from discovery through completion into production.

In addition, we created a method for designing an organization model that was counterintuitive to most HR and management consulting firms. The conventional wisdom was to start at the top of the organization and move down layer by layer until you finally reach the people that do the work. My desire was to put the priority on the people doing the work rather than the leaders managing the work. We turned this approach on its head and decided to do the design bottom up. We loosely built what people now call the Spotify model. We just called them “small cross functional teams”.  We followed the “work” – forming squads around the work. We then grouped squads into Tribes, then grouped Tribes into Domains. By doing this, we reduced layers from 13 to 5. We would have never got to that result going top down – leaders would have protected their empires. The key learning was “turkeys don’t vote for thanksgiving” – follow the work – not what the existing bureaucracy is defending.

Ultimately, with a much leaner leadership structure, there was less “noise” distracting the squads and they had clarity of purpose of what problems they needed to solve.  Velocity and throughput increased because they learned through practice and course correction how to improve. We also saw the competitive spirit drive passionate curiosity to look outside their own areas to listen and learn from others. Product teams learned new tactics like test automation, continuous integration, continuous delivery, etc. Squads as well as individual skillsets improved more in a short amount time than they ever had before. More than anything, people learned how important it was to care about their craft and accountability to their teammates.  

A key element to having actual results match high aspirations is the ability to measure the results – like Man U – we needed to keep score. I had a goal where I wanted to be able to visit any squad and have them describe the maturity of their culture with tangible measures. We developed what we called the “Agile Maturity Assessment” (AMA).  It was essentially a traffic light assessment of the 20 key Agile practices we rolled out. We did this every quarter with each squad self-assessing their maturity on a scale of 1-5 with Agile coaches moderating the results to ensure consistency. We made this process transparent and the squads got serious about getting better. At the beginning, many squads were not strong in most practices (e.g. continuous delivery). But once they saw we were only concerned about them improving – not what their actual scores were – they gained skills in seeking constructive feedback and learning from others. One of our key goals was to accelerate improvement by having squads seek out advice from other squads that had more maturity in the practice they were trying to improve.

Overall, the water level of talent and skills improved at an accelerated rate because we had clarity of purpose, one culture – Agile – underpinned by 20 practices that we practiced, adjusted, measured and evolved. Through my tenure at Suncorp, my scope expanded to take on all the shared services of the company. This included Finance, Procurement, HR, Real Estate and Risk. As CEO of Suncorp Business Services, I applied the same playbook – and although not all the practices were applicable to every squad, more the half were the same. I saw the same results come through in non-IT squads because we set high aspirations to know what best looked like and sought ways to get there.  We ultimately became one of the first financial services firms to put production banking and insurance applications in a public cloud with AWS. We couldn’t tackle this until we had built the talent to match the tactics required to design, build and deploy applications in a public cloud. A highlight for our people was being able to talk about our story as part of the AWS Re-Invent keynote in 2013.

Key Lessons Learned / Mistakes I made:

• Do the organization design and restructure quickly – I spent close to 9 months doing this. It leaves too much anxiety in the system. You’ll make mistakes – but better to complete the restructure quickly and course correct as you come across problems.
• Make your culture tangible and mandatory – don’t give people an out. Initially, I made using Agile practices voluntary – thinking everyone would want more productive, leading edge practices. I was wrong – lots of people opted out. When you are doing a major transformation, clarity of purpose, grit and consistent principles are essential – don’t give people an out with ambiguity.

My experience at IBM

From this experience I took on the Global CIO role and leader of the Agile transformation effort at IBM. I followed a similar playbook to what I had done at Suncorp. I built a Playbook with the Agile practices that we wanted to operationalize as our culture. I hired Agile coaches, designed on organization structure from the bottom up with a similar Spotify model of squads, tribes and domains. We ended up reducing layers of leaders from 15 to 5. Although the playbook was essentially the same, IBM brought a few more wrinkles to this approach. One was size. My CIO team was roughly 23k and IBM had 375k people in a distributed global workforce. IBM had larger scale and complexity – but largely the problem space was the same.

The biggest issue I had to overcome at IBM was the existing culture. IBM’s culture drove the love of the rank of position over the love of craft. IBMers were conditioned to look for opportunities to become executives (of which there are multiple exec levels) because of the entitlements that came with those positions. For example, when people became an exec – they would get a plaque congratulating them on becoming an executive – not what they actually accomplished. They even had quarterly celebrations for the newly minted executives. Leaders even got different options for office sizes, carpet quality, choice of artwork, etc.  – based on the level they occupied. I had never seen anything like this. IBMers also respected hierarchy to a fault they were conditioned over time to know where to sit in a meeting, who they could speak to, etc.  Conversely, an Agile culture has no hierarchy in communication, innovation, or entitlements. The best ideas have to win.

I needed to bring purpose to my squads. We needed a big problem to solve on behalf of all IBMers. We decided to tackle creating a productive work environment. For example, from a device perspective, Mac’s weren’t really allowed because of the extra cost. In addition, the software collaboration tools were made up of IBM tools that were out of date and not productive. IBMers were dissatisfied with their environment and wanted something different and modern. This was the big hairy problem to solve. Ultimately, we co-developed with Apple a way to provision and deploy Macs zero touch that made the cost similar to PC’s and the employee experience much better. We also branded the environment as Mac@IBM. It went so well we expanded the effort to the entire collaboration environment and implemented Box@IBM, Slack@IBM, etc.  IBMers realized how important a world class collaboration environment was to improving productivity and employee engagement.

But something was still missing. It took me a while to figure this out. It had nothing to do with IBM. It had to do with me. I had spent years focusing on building high performance teams and finding ways to operationalize culture and measure results for people doing the work. Conversely, I had failed to come up with an equivalent maturity assessment for Leaders.

Key lessons learned / Mistakes I made

• I still spent too much time on the organization redesign. It’s better to make the tough decisions, deal with the pain all at once and install a new culture as quickly as possible so the focus can shift to “looking out the windshield and not the rear-view mirror”.
• I should have included more leaders in the design process. I relied on a small, trusted set of strong performers, but it would have been better to include more leaders – even if that had created more friction. By being more inclusive, I would have had a wider diversity of thought. In addition, even if leaders were impacted by the leaner organizational model they participated in designing – they would have had greater appreciation of why it happened.
• Although we developed an effective AMA for squads, I didn’t have a commensurate measurement system for Leaders. This wasn’t fair – Leaders are paid more money for the responsibilities they have – they should be measured with the same discipline as the squads they manage.

My experience at World Fuel Services

I took all these learnings to my current role as COO of World Fuel Services. Our company manages $40 B dollars of energy on behalf of global customers in the Aviation, Marine and Land based markets. As COO, I have global talent in both IT and non-IT roles.

Once again, I took the things that worked well and leveraged those experiences. We created a playbook around Agile practices, hired coaches, organized our squads from the bottom up and created an AMA for our squads. But this time, I did 2 things fundamentally different. First, I included all 55 global executives in my global leadership team in the design of our Agile operating model.

To ensure broad, inclusive and diverse participation, I invited all 55 executives to work with me in Miami in intense design sessions held over a week. In following a similar Spotify model, we followed the work and designed squads, grouped squads into Tribes and finally Tribes into Domains. Every day we broke up into different design teams and shared retrospectives and showcases. At the end of the week, we collectively agreed on our overall design. The new design only required 27 executives and reduced the number of levels from 10 to 4.  We also calculated our BMI (Bureaucracy Mass Index) – which is a ratio of leaders to doers. We made sure our ratio was between 10-14%. Since all executives participated in a transparent process, there was huge buy-in – even though everyone realized this simplified design was going to impact roughly 50% of the leaders.

Following the design, we needed to complete an assessment and selection process to assign leaders to the new roles. Since everyone knew this would negatively impact 50% of the leaders, it was critical to complete the assessment and selection of the leaders quickly since anxiety would be high.  I ran an expedited interview process and reviewed the assessments and design with my peers for complete buy-in and feedback. We completed this process in about 3 weeks and coordinated all announcements for those who landed roles and those who didn’t. Overall, the process front to back from the initial design process to assessment, selection, placement and implementation was 1 month. This was the most inclusive and fastest organization restructure that I had ever been involved in and a huge improvement over my prior experiences.

The second major change I added to my playbook was a “Leadership Maturity Assessment” (LMA). Similar to the AMA that I used with squads, I wanted a similar disciplined approach so leaders could learn, practice, measure and improve their leadership practices.

We designed a LMA that had 7 dimensions. 3 dimensions were outcome based, 3 were behavior based and the final dimension was a Net Promoter Score (NPS).

Leadership Outcome Dimensions:

1. Forming Productive Squads – This is probably the most critical dimension. Ensuring the squads have the right skills, chemistry and diversity of thought, gender and race is critical. In addition, continual replenishment of talent is critical to forming and adjusting squads.
2. Distributing Work – This sounds simple – but this might be the dimension most leaders struggle with the most. Distributing the right work at the right pace to optimize the productivity of squads is essential. I see over and over again teams with too much Work in Progress (WIP) – which is the biggest detractor to productivity. Feeding too much work to squads and changing priorities within a sprint slows down velocity and throughput.  It also demoralizes squads because they want to get things done. Multi-tasking doesn’t work well – the switching costs are high. In addition, John Doerr who wrote “Measure What Matters” says that one of the biggest issues companies have is that a relatively small percentage of people work on things directly related to what is strategically important. Managing priorities, backlog grooming and sprint planning is essential to “doing the right work” and “doing the work the right way”.
3. Measure What Matters – It is critical to identify the fewest, measurable outcomes that give signals to the squads on how they are doing and how to improve. A key tenant is that these measures should be valuable to the squads – not just the leader. Examples of delivery measures could be velocity, cycle time in backlog, cycle time in WIP, etc.

Behavior Dimensions

1. Attract, Develop and Retain Talent – Ensuring your leaders are “magnets” for talent and not “repellents” is critical to having high talent density. Key evidence to look for can include: retention of best resources, how large the attraction canvas is (how far away are people attracted to a leader – other groups, other companies, etc.) and length of time in a squad (many people stagnate by staying too long in a role in a squad).
2. Listen, Learn and Leverage – This is the most important dimension to me. Finding people that have passionate curiosity to look outside their own frame of reference is a rare trait. Someone that learns from others in your own company is good, but someone that reaches outside to find what “best looks like” and strives to meet that is infectious and critical to true innovation and productivity.
3. Influence and Impact – Finding leaders whose impact/influence is wider than the actual scope of people they manage is essential. Building Leaders whose impact is larger than their remit builds resilience and adaptability in your leadership system. It also enables strong succession bench strength if your leaders already have insights / knowledge outside their own responsibility.

The final dimension is NPS. We ask 2 questions of all employees for NPS – “would you recommend your leader to an outside colleague?” and “would you recommend your squad to an outside colleague?”. These questions are direct, transparent and useful in determining whether leaders are detractors or attractors.

Operationalizing a Leadership Maturity Assessment is critical to giving meaningful, timely feedback to your leaders. We do this every quarter for every leader. We use a rating of 1-5 for all 6 outcome and behavior dimensions and we add in the NPS score to the LMA.

Leader Maturity Assessment:

Another benefit of this system is it is applicable to all leaders at all levels in any discipline. It doesn’t matter whether you manage 5000 people in 500 squads or 5 people in 1 squad. The assessment is the same. This means you can learn it, practice it, measure it and get better. Many HR executives like to invent multiple leadership programs such as first line leader training, middle manager training, executive training, etc. I don’t believe in this – the essence of Leadership is the same at all levels. The only difference is scale and complexity – and having one leadership program is simple and equitable. By having one LMA, you create clarity of purpose and enable people to build their leadership skills over time as they take on new assignments and additional responsibilities. This also ensures people take on work for the love of their craft and not the love of their rank. In too many instances, Leaders equate new leadership programs by rank as a way of getting ahead and falsely perceiving career advancement.

Key takeaways from 15 years of making mistakes and learning:

1. Clarity of Purpose – Simple, clear, objectives are key to getting your whole organization behind the change you are leading. Instead of having 10 concurrent strategic objectives – pick 3 – get those done, then go on to the next 3, etc. An example Strategic OKR (objective and key result) that we have at WFS: Objective: Move all applications to the public cloud: Key result: all data centers closed in 2 years. Defining OKR’s that everyone can understand regardless of role or level is critical – shared understanding matters – and the objectives should be aspirational to drive transformational thinking.
2. Grit, perseverance, tenacity – there will be many, many things that will go wrong and lots of people around to remind you of your mistakes. Having a key mindset of “course correction over perfection” and the determination to get up off the mat when most people would quit is essential to overcoming the inevitable obstacles that come with anything that’s difficult, but worth doing.
3. Talent Density over Process Density – the number 1 critical success factor will be the talent you are surrounded by. This is equally true for both your internal resources and external partners. For our external partners, I have a principle that we pick those whom “we aspire to be like, not necessarily like us”. In addition, we prioritize our partner’s Leadership and Talent over their product features. We believe our partner’s culture is forward looking and their product features are backward looking. Lastly, both sides of the partnership should benefit outside the commercial relationship. Our key partners at WFS have assisted us in a variety of ways such as college recruiting, Leadership and Technical practices, etc.

Overall, I’ve made many mistakes, learned many things and course corrected over time. I believe by listening, learning and course correcting, I have come up with a decent recipe for operationalizing culture. I continue to adapt and seek out “what best looks like”. I know I’ll never get there – but if I can get close enough that it forces me to reset my aspirations, I know I’ll continue to improve.

Finally, my core belief is that life is short enough – and your career is even shorter – it might as well be memorable!  Doing something of substance, of significance is incredibly rewarding. Also – your roles may change as you move through your career – but your friendships and experiences you take with you the rest of your life. They might as well be great! Overall, aspiring to be valuable works out better than aspiring to be successful.

I would also like to say a special thank you to some wonderful people who I have learned from and who have had a huge influence on me…

Tom Mendoza.       Former President & Vice Chairman at NetApp

Aaron Levie.             CEO at Box

Eric Yuan.                 CEO at Zoom

Phil Abernathy.       Founder & Managing Director at Purple Candor

Patrick Snowball.    Former Group CEO at Suncorp

Andy Jassy.            CEO at AWS

Mark Lelliott.           Managing Partner at NGS Global

Stuart Butterfield. CEO at Slack