I was working on a long post for the main blog, reflecting on the attitudes and approaches my managers have taken to running their orgs, mapping them onto my own successes and failures, and came to a bit of an epiphany¹.

There are 3 classes of technical executive—

• Warriors: goal-directed individuals that lead from the front. They have maniacal focus on particular results and ends. They drive their teams to deliver, getting as “hands-on” as it takes and wearing any hat. They demand excellence and produce professional progeny that go on to do great things. Impatience is a virtue.
• Priests: group-oriented individuals who focus on configuring the environment for success, shepherding their constituents through change, and bearing the culture of the organization. They don’t disregard results so much as let the results come naturally as a consequence of the environment they build. Impatience is a disaster.
• Mages: technocratic leaders that emphasize domain expertise and creative problem solving. They lead their functions as the unequivocal brains of the operation. The innovations they bring into the world create durable competitive advantages for years to come. They run on their own internal clock, patience and impatience have nothing to do with the matter.

Each class works well in different situations.

Warriors can make big goals happen against all odds and defend your castle when competitive pressure is high. Priests can keep your money printer printing by leading large groups through important changes and much needed evolutions. Mages can invent the future and unlock the next major step change in capital productivity.

Maybe more interestingly, individuals can switch classes as they move through life.

Reflecting on my own story, I started out as a Mage, moving up the ranks as a technical encyclopedia and master distributed systems engineer. My early teams were highly specialized functions, R&D groups, and low-level infrastructure that lived or died by sheer technical foresight.

Once my problems got big enough, I needed a team-of-teams to get them done, and I respec’d as a Priest. I learned how to implement large-scale org changes, use incentives and process design to generate outcomes, and fine-tune sociotechnical systems to meet the demands of the business. The largest success of my career came as a Priest: transforming an org of ~1200 to ship with 2x higher velocity, 25% less overhead, 32pp more job satisfaction, and one more 9 of availability. It took 3 years of nonstop (but careful) effort to slowly boil the frog and change the culture.

As I departed that role, I was hungry to execute at the speed of tech (rather than the speed of people) and put all my accumulated expertise to the test. After holding roles across PM, SWE, SRE, DS, UX, BigCo and startup, what could I accomplish if I ran the whole show? What if I “eat what you kill?” This was when I took on the Warrior class.

Each class has a role to play

Your class matters but so too do the classes of your leadership team.

Warriors tend to leave a trail of dead bodies behind: angry peers, disgruntled employees, cultural debt. Even some lesser known detritus like vestigial teams working on dead-end projects and maintenance-mode tech. The Warrior has ignored these groups because they are not on the critical path and “pick your battles” is the motto of all Warriors. To succeed, Warriors need a leadership team filled with Priests who tend to the masses.

When I was part of Google Drive, this was my role. At our helm was a true Warrior that instilled obsessive focus on critical goals (e.g. swarming entire teams of 10-16 engineers on a single project at a time) and personally leading the product design, org influence, and technical solutioning for our top couple objectives. I ran behind her, translating mandates into vision, supplying context, tending to the gritty technical details obstructing delivery, and leading the less glamorous infrastructure groups needed to keep everything online.

The Priests in this scheme illuminate the strategic insight of the Warrior, making their mission heard and felt across the org and helping bring the whole team along for the wild ride. They tend to the process, organizational, and technical debt that hinders the mission and create the efficiency/resiliency/constancy required for an organization to prevail.

On the flip side, a Priest might create a well-oiled, highly-retentive organization but needs Warriors and Mages to ensure the group actually achieves something meaningful to the business.

The only thing to avoid is multiple levels of leader with the same class. Warrior clans burn out their teams and create cultures of fear. Mage clans produce over-engineered monstrosities that permanently increase the cost of doing business. Priest clans devolve into adult daycare complete with back rubs and story time.

Build the team you need to succeed

If you are an executive and find yourself hiring a new leader, choosing which class you need (and figuring out the true classes of your candidate pool) is paramount. There are lots of faulty intuitions and traps waiting if you half-ass this.

Consider some common scenarios that might lead you to hire a new leader.

Execution Swamp
Money goes in, nothing comes out. The lead seems too soft, too checked out, too ineffective. Your instinct is to hire a Warrior who will come in and show ‘em how it’s done.

Resist this urge with all your might. It will not work. The thing is, no one comes to work wanting to suck at their job. If an organization is failing to execute collectively, the problem is going to be a complex, nuanced combination of systemic factors, individual performance problems, and technical issues that have been deferred far too long.

Warriors will quickly become discouraged by the magnitude of the change required to achieve anything, impatient at the slow rate organizational-change-at-scale unfolds, and will create secondary disasters by applying too much brute force to a delicate operation. Instead, grab a Priest who will carefully orchestrate an organizational pivot with their trademark combo of personal and professional care.

Engineering Hairball
The systems are chronically over engineered. The brain trust has successfully turned your simple web app into a distributed replicated cache replete with CAP theorem problems. Or the inverse: your majestic monolith ain’t cutting it anymore. Incidents are piling up and nothing is getting better. Either way, the issue seems to be a lead who doesn’t know any better and lets their team get away with this shit…maybe even sponsoring some of these gaffes.

Your instinct is to hire a Mage who will up-level everyone’s technical acumen. This too, will not work. The Mage does not actually know how to scale their expertise. They will become a bottleneck at best, or frustrate and scare off all your talent at worst, and won’t even realize it. The Mage lives by the credo “technically correct is the best kind of correct.”

In a truly nightmare inducing scenario, the Mage will fall victim to full-blown Second System Syndrome and make everything worse. Their inherent genius will enable them to sell you on the idea, none-the-wiser. You’ll wake up 4 years later in the middle of an unfinished migration from hell unsure what just happened.

What you really need is a Warrior. Their obsessive focus on results and slightly aggro approach to the job will put an end to any unnecessary science experiments. The complexity perverts (they call themselves “architects”) will become frustrated and quit of their own volition. Meanwhile, a lean core of doers will emerge and rise to the top, basking in the warm glow of the Warrior’s sun.

Urban Sprawl
You have 3 different workflow engines, 5 different release tools, and 9 little fiefdoms each with their own petty overlords. The lead seems unwilling or unable to make the group behave like a single team, bring duplication into check, and focus their efforts into a cohesive strategy.

Your instinct is to hire a Priest: they will unify the group, bring about process and incentive change, and simplify the sprawl into an optimized value chain with fulfilled and energized workers. Will this work? lol, no. Your Priest will be overwhelmed by the complexity and won’t know where to start. Their model of cause-and-effect will be wrong and resulting organizational interventions will be mismatched to the actual problem. Completely stuck, they will lean on their strengths and turn the problem into a group brainstorm with the same predictable results as design-by-committee.

In these scenarios, you need a Mage to end around the kingdom building and incentive gaming. Their outsized capacity for complex and complicated scenarios (word up Cynefin) will allow them to conceive of technical solutions that can outcompete the fiefdoms’ bloatware. Free market forces will take it from there.

Picking your next role

This framework cuts both ways. Just as you’d use it to staff your leadership team, you can use it to evaluate the role you’re being asked to play.

Having just come up with this while puttering around ORD waiting for my connection, I’ve never actually put this to the test, but I suspect this could be pretty helpful for picking your next job.

During interviews, get some grounding on the situation you’re walking into, map it to the appropriate class, and ask yourself if you want to do that for the next couple years.

You can also use it as a bit of a shit test for your new boss: do they know enough to look for the right class? Before accepting any offer, ask your would-be-boss, “Why me? What do you think you’re going to get?” Their answer says everything and you need to read a lot into what they say. Every single word carries meaning. Even if they are someone you’ve worked with before and you know they are good, do not skip this step. They might have known you 2 or 3 class changes ago and it is extremely difficult to override/update the impression left by your past self.

1. This one might be off the deep end so I’m keeping it over here on my personal site 🤣 ↩︎