Grappling With Groups

I have been grappling regularly with Dunbar’s Number of late as I attempt to rationalize and navigate a variety of group dynamics past and present. Consequently I found myself in for a treat when as I rolled out of bed yesterday I discovered that the latest episode drop for Patrick O’Shaughnessy’s “Invest Like The Best” show offered an interview of Dunbar himself. Listening to this sparked myriad thoughts across multiple epochs in my own life as well as recalled a few books I have read in recent years. Dunbar’s Number seems highly relevant in all of them though I further imagine that the artifacts of technology, process, and space play a big role.

DoD Days

Circa 2010 I found myself and a few other people handpicked from a vastly larger organization to form a team that could operate in a SWAT-like fashion and over many years gradually reimagine and reforge how a whole class of Computer Network Operations happened. The years spent in service to that quest proved some of the best and most memorable of my career and I wonder if I will ever manage to partake of such camaraderie and purpose again. I could write a book on those days (well, maybe in thirty years?) but here we will focus on the evolving group dynamics as a function of size and meatspace constraints.

The group, or arguably a precursor thereof, started with about ten people crammed into an ops room that quickly began to, as one engineer remarked, “smell of men and feet” as a consequence of too many people and not enough ventilation. There I would birth a software system, in cliched fashion upon a desk consisting of a door and two filing cabinets, that would go on to run for about thirteen years before its retirement, almost half of those years being after I left federal service for different pastures. Eventually, blessedly, we found ourselves in a new building with a new room in which to grow (and I with a desk of less body-wreckingly disastrous ergonomics).

The epoch where we actually fit into that new larger room may have been our smoothest years in terms of team dynamics. Eventually, though, as our ranks swelled from fifteen to thirty folks, we found ourselves split into two SCIFs, this triggering much hullabaloo over who would slot into which room. As it turns out the difference in required activation energy between swiveling a chair or walking sixty seconds down a hallway to have a conversation with a colleague proved so huge that in the latter case they might as well have resided in a different zip code for many matters.

As far as buildings went, though, we felt for a long time as if we had won the lottery. Working on a satellite campus, we could reliably park instantly and have maybe a 1-2 minute walk into the building instead of a nerve wracking (and often viciously competitive) extended hunt for a spot on the main campus followed by a ten to thirty minute odyssey to get from car to desk. Our in-building cafeteria, however, did truly suck, to a degree where pretty much everyone either packed a lunch that often got eaten at their desk or went out for lunch in the nearby town. This, too, exerted a subtle effect on group formation. I was part of a group of four (and sometimes five) people who went out to lunch nearly every day. Why not one or two more? I suspect that, beyond the natural affinity these people had for each other, another back-pressure existed — how many people could you comfortably fit in one sedan? Add just a few people to the lunch crew and now you need two cars. Introduce a second car and the struggle to wrangle everyone out the door simultaneously somehow explodes. And that’s before you consider that seating in the fast-casual establishments we frequented generally created friction as parties larger than four tried to dine together.

Later, as we became victims of our own success, the luxury of working on a satellite campus morphed into a burden for those of us in leadership roles that necessitated our driving over to the main campus to collaborate with sister organizations. The isolation that had allowed us to innovate quickly and effectively evolved from asset to liability as efficient coordination and inter-office trust gradually became the key currencies for encouraging broader adoption of our novel tech stack. The maddening inflexibility of office space eventually manifested frictions that found us beholden to the edict of Conway’s Law — that institutions will inevitably build software architectures that mirror organizational architectures (which in turn mirror building and campus architectures). On the whole our body of work proved a smashing success with lasting impact but some gravity wells prove too deep to escape entirely.

Bridgewater Bumps

T’would be understatement to remark that my 2.5 years at Bridgewater Associates circa 2016-2019 failed to follow the initial visualization. I hadn’t prior to my arrival adequately appreciated the degree to which the firm was experiencing growing pains, struggling to navigate a succession plan, and wrestling to figure out its relationship with technology, one outcome of which included the “Tech Strategy And Incubation” organization I joined dissolving and and my hiring manager departing within months of my arrival followed by my being re-settled into Security (again with the deep gravity wells). Nonetheless that period proved among the most valuable in my career for personal growth as a result of the culture that Ray Dalio envisioned for his company (though not always exactly the way he might have hoped). When I sent a parting note to the “BW Public Goodbyes” email list Ray offered a personal reply that I will cherish in which he expressed appreciation for how thoroughly I had understood the value prop on offer and leveraged it to evolve into a better version of myself.

The Dot Collector, a real-time fully-transparent micro-feedback tool, served as a central pillar in the whole experience. At any moment you could submit a batch of “dots” with an attached comment to rate someone one-to-ten on a collection of attributes, anyone in the entire company could see it instantly, and these ratings would feed into weekly “interesting dots” reports as well as quarterly performance reviews. When wielded by calm, thoughtful, ethical, well-meaning people and received by open-minded, growth-oriented, non-emotional people the process yielded incredibly valuable and timely insights. Instead of people harboring secret frustrations with your engagement style or work product and perhaps covertly working around you they could opt to let you know immediately, you could request clarification and suggestions, and then ideally you would iterate quickly to improve with a newfound shared understanding.

I embraced the system wholeheartedly, both the giving and receiving, and felt I received enormous value from the process. The CSO, upon my re-org into his realm, remarked “I thought you must have been a ten-year employee here given the quality and quantity of feedback you are providing people”. On the receiving end my “negative dots” — Synthesis, Practicality, Prioritization, Escalation, Dealing With Ambiguity — created the opportunity to take a hard look at the personality traits and skills that were holding me back professionally. I took the mantra of “Pain + Reflection = Progress” to heart and put in the work.

When it was good it was great — the Bridgewater long-timer who regularly dinged me with negative Synthesis dots (once, particularly memorably, with the single-word attached comment “prolix”, after which he shouted down the hallway at me “GIBBS — LESS IS MORE!”) was the same guy who repeatedly organized an in-house multi-day writing course which at his encouragement I attended twice on company time. When it was bad, though, it could be kind of terrible and I suspect that Dunbar’s Number played at least a substantial role. The complete transparency of Dots, alas, created a performative element wherein the Dotter might not just be attempting to offer helpful feedback but also be aiming to burnish their own reputation or even weaponizing the system to further political goals. Worse still the real-time feature could engender a pile-on effect wherein cynical people just looking to make their weekly negative dots quota would see someone already getting hammered as a chance to attach some Me Too dots without really providing additional useful information (though I do have a delightful memory of a specific Principles Coach, bless his heart, singling out someone who gave “a soft five”, asking them to justify it, and when they couldn’t explain their reasoning proceeded to publicly crucify them).

The really valuable feedback seemed most likely to come from regular collaborators who knew you well (and thus were invested in your growth) and offered feedback that, while theoretically always a matter of public record (you were expected to record all meetings at Bridgewater), was not showcased in so splashy a forum as the Dot Collector (and so you knew was offered more earnestly than performatively). The bad behavior, inversely, seemed far more predictable when a Dotter just wasn’t materially invested in the wellbeing of the Dottee, an eventuality all the more likely as the company grew from the hundred-ish people when Ray forged his principles to the 1500 employees and equally many contractors populating the place when I joined. This stemmed, perhaps, from three causes — firstly, the more people inhabit a shared ecosystem, the less cognitive bandwidth and emotional investment one has per capita for their colleagues, and people will inevitably apportion these finite goods unequally; second, the faster an organization grows, the less likely it is to develop and maintain a coherent culture, this owing to one generation of hires enculturating the next generation of hires before before having been themselves adequately enculturated; thirdly, the more sprawling and thus complex the emergent social structures, the more inherently political any given far-reaching decision or event will prove, and we all know how politics are.

As I reflect on the experience I find myself yearning for a culture and toolchain that embodies the spirit of what the Bridgewater of my day was attempting (and in some measure provided me, albeit imperfectly) but with a better practical understanding and acceptance of what Dunbar’s Number suggests about cognitive limits and power dynamics. I wonder how, for instance, the Dot Collector might have manifested differently if it had baked in a public visibility latency long enough to disincentivize performative behavior. I wonder about a lot of things.

Startup Life

Upon leaving Bridgewater early 2019 I spent about a year as an executive at an early stage start-up whereupon my joining the staff numbered fifteen. In the subsequent four years I have served as an independent consultant at an assortment of scaling-stage VC-backed start-ups. Common to all these companies is a story about how small size, market volatility, and evolving purpose engender wild fluctuations in relative workforce sizes during small increments of time which yields huge consequences for group dynamics and social cohesion. I’ve seen tiny companies double in size in a few months then halve their workforces in a day. I’ve seen larger companies triple in size over a few years then lay off a quarter of their staff in one swing of the cleaver. Careful and continual consideration of workforce composition and organizational architecture prove crucial in such high-churn ecosystems.

As Ben Horowitz exhorts us in The Hard Thing About Hard Things, there is no perfect person for a given job role, only the best person at a point in time. The VP Engineering perfect for scaling you from 15 to 150 employees may very well prove the wrong one to get you from 150 to 500 and vice-versa. The flat structure that works ideally when you’ve got just a single product and a staff that fits in a small room will begin to creak under the strain of a burgeoning workforce and a product suite spanning multiple interlocking systems. As your company grows you will want to put increasing thought into Team Topologies to get the most from your people while also asking hard questions that focus as much on whether a role has outgrown a person as a person has outgrown a role. And yet as you strive to control the chaos you should do so with a gentle hand that appreciates how the heterogeneity and randomness of being a little bit Messy sparks novel combinations, serendipitous discovery, and brilliant innovation, something that for all the ways I appreciate remote-work I worry a widespread pandemic-triggered Zoom-mediated present-day way of working may be undermining.

Paragliding Pitfalls

Coincident with my consulting epoch I took up paragliding. I would forgive you for thinking from your perusal of publicly available information that mostly it’s just really rad and once in a blue moon something spectacularly scary and/or heartbreaking happens. In reality, however, much like all of aviation and so many other disciplines, a continuum of bad outcomes exists with distributions of frequency and severity like you might expect. In a perfect world we would have open, honest, humble, and frank yet empathetic discussions about about all incidents at every tier of severity in the hopes that we could short-circuit more severe and more regular bad outcomes through education.

Probably the more a given form of aviation involves capital intensive infrastructure and a large swath of the general public the safer it will be. If a commercial airliner goes down then insurance gets involved, the media picks it up, and one or more federal agencies will investigate. Even something as mundane as a Cessna bending a wing reliably garners some kind of formal process and transparency. With paragliding, however, much of the mishaps involve “cheap lessons” that often get swept under the rug, perhaps because it is more likely that bones were broken than aircraft. This creates a situation where the major drivers of share-or-bury decisions about accidents center on local culture, psychological safety, and reciprocity. The fewer people who share their mishaps, the more skewed the perception of accident frequency and severity becomes, and the less likely the next person will be to share their story which creates a continual downward pressure on honest accounting and sober analysis.

So far I haven’t figured out a better way to get someone to tell their authentic story than to make a home cooked meal, crack open a bottle of wine, strive to be a good listener, and treat the divulged information and emotion as if stamped with the Originator Controlled handling caveat. The next best thing I have figured out is to own up to my personal mistakes on public forums, be open-minded to the earnest feedback, and to accept that some people can’t quite act like grownups reliably (probably because they are busy equating “hasn’t happened to me” with “won’t ever happen to me”). I haven’t yet figured out a better middle ground that both accounts for most people being less outspoken than me and yet also gets a fuller accounting of our communities’ mishaps out to the public to help in preventative education efforts. Most people aren’t comfortable sharing on a public forum consisting mostly of one thousand strangers yet probably those same people have their fave-five (or fave-fifteen) where they can share earnestly. It’s just a shame that so much of that analysis and learning stays trapped inside those small social bubbles. Surely there is a better way and with so much at stake I hope we will eventually find it. The problem is that things have to look worse before they can be better and the people willing to share their screw-ups publicly are few and far between.