skip to content
fdbck lps
table of contents

Mixed metaphors

The other day I read a good blog post by Kevin Lawler called “Streams”, about how streams accumulate, how subscription services have consumed the internet, and further non-obvious insights. This got me thinking about how many of our work metaphors are actually water-related. We have the idea of stream-aligned teams (“you build it, you own it” groups responsible for the entire product lifecycle) in Team Topologies, as well as fast flow, the speed at which value is created. And value is treated like a liquid too, in value stream mapping and in waterfall project management. Your data engineers are upstream and your users are downstream; my mistakes have spillover effects. And of course we know and love all the streaming data that occupies our work days and sometimes our evening leisure time too.

Why does this matter? Our metaphors hint at how we see the world because they borrow from our experience to make sense of more complicated phenomena. I tend to go for audio metaphors (“sounds like” and “sounds good” and “on the same wavelength”) and you might be tempted to draw some weak inferences about me from that. Well, we can do the same with conceptual water metaphors.1 They feel natural for obvious reasons — after all, we’re surrounded by the stuff2 — but I think they say more about the forces that constructed the metaphors than anything else. If we choose to view them a little differently, maybe we can expose the highest leverage control points, letting us effect positive organizational change.

Value streams, lazy rivers, and fūkeiron

Let’s dive into the example of value stream mapping. It’s a “sequence of activities necessary to deliver… [something] to a customer”, a directed temporal flow of materials and goods that ends with the buyer, voyageur-style. Sounds safe and pleasant and anodyne. But that’s exactly backwards! I saw a really impressive waterfall last month, and when I think back to that moment, what struck me was the sheer force of the thing. It’s unsettling to think that the stuff you drink every morning has enough directed energy to casually kill you. But none of that raw momentum shows up in the value stream signifier. It also misses out on the dynamical behaviour of rivers; the way they cut through earth to form new channels, to deposit sediment and create new land, and to generally do all manner of unexpected, unintuitive things. This dynamic richness is also curiously absent from virtually all the other water metaphors I mentioned earlier.

Now, mapping the Raging Deathtrap Rapids doesn’t have much of a ring to it, so maybe the “stream” theory of value was always destined to win out over such a chaotic metaphor. And if all conceptual metaphors are necessarily incomplete to begin with, what’s so bad about seeing your supply chain as a lazy river? Wouldn’t any other conception fail in some other way? Maybe so, but the sanitized water metaphors we’re looking at conceal a reifying force that’s laundered through the ordinary meaning of the process diagram. In other words, they don’t show us the true locus of power in the value equation. To make this more explicit, consider your favourite canyon. It’s been shaped residually over aeons by patient hydrological forces. But it’s the absence of the river that makes the canyon impressive. Nobody would blink twice if the Grand Canyon were a giant lake!

Grappling with objects through their absence is nothing new3 and I don’t pretend it’s an innovation to apply it here. But I want to focus on one particular deployment of absence-thought that I discovered independently of org design theory, instead through my interest in art film. It’s also particularly rich, both for its natural alignment with known design principles and for its environmental bent. It emerged in 1960s Japanese film; they called it fūkeiron, the landscape theory. The filmmakers would invert the camera, shooting landscapes without humans, both in direct response to more popular documentary approaches of the time, and as a technique to reveal the environmental power structures that shaped the very people their cameras elided. The best-known work in the fūkeiron tradition is probably AKA Serial Killer, a documentary that’s more interested in showing train tracks and neighbourhoods than the titular spree killer himself.4 Yet as we watch the landscape pass by travelogue-style, its spatial logic comes alive, and we start to see how it prefigures — and maybe even predetermines, if you buy the director’s view that the “image of power takes the form of the urban landscape”5 — the social relations that shaped the murderer’s path. This is seeing the river’s force through the canyon’s form.

Org mode

Now the film was intended to propagate a highly localized political critique that isn’t relevant for my purposes here, making it, like so many others, an imperfect metaphor.6 But even so, I think its landscape technique has rich applications in the organizational world. If you think back to Conway’s law, the idea that communication predetermines design is cognate with how the environment shapes social factors in fūkeiron. Both suggest prestructuring forces that mould human behaviour, whether in the omise or the office.

In this metaphor, the systems thinker becomes the documentarian — just swap the camera for a map of the organization — flipping their tools 180 degrees to chart the environmental structures that drive workplace causality. And what does that organizational landscape look like? It’s obviously physical: we have floor plans and conference rooms and collaborative spaces and break rooms and quiet areas, all of which are built to enable specific activities — and therefore to discourage others, even if not explicitly. The cynic might point to the Taylorist open plan as a more extreme environmental example. There’s a rhythm to the temporal landscape too; think of sprint cycles, the cadence of recurring meetings, the 40-hour work week, and the like. And in proper Conway fashion, communication pathways also thread the countryside: emails, group chats, the DM antipattern. When we apply fūkeiron cartography to this terrain, we see the possibility of new dynamics at work. For instance, value might just be a boardroom and a couple hallways, some status updates passing through the shadow org chart, two pizzas, and a few Slack channels in a trench coat.

This way of documenting calls to mind a scrying pool more than a value stream. But it does a great job of exposing control levers that were invisible in the water metaphors we started with. It also gives us the opportunity to steelman “we’re always done it this way”, which is now an almost descriptively true statement of how value is produced! This means it’s not dissimilar to Stafford Beer’s point that the purpose of a system is what it does; if we’ve always operated neatly within the landscape, then maybe the resulting output reflects the real intent of the system. If the steelman fails, it’s because another purpose of a system is to oversimplify its own causal diagram until these levers are all but hidden, meaning it becomes harder to parse exactly why we’ve always done things a certain way.7 But the landscape theory opens up a creative angle with which to find these levers anew.

Instead of optimizing the value stream, what would happen if we reshaped the canyon? Changed how conference rooms were booked? Moved the coffee machine to a different floor? Built a room where the minimum volume was 70 dB? Diffed a map of the shadow org chart against its formal counterpart? Removed email? Gave juniors expanded editing perms on strategy documents? How would that change your landscape?

Footnotes

  1. WORK IS WATER, as Lakoff would write. ↩

  2. As long as you’re not that one fish. ↩

  3. Here I’m thinking specifically of John Cage and Waiting for Godot, but there are probably dozens of even better examples. ↩

  4. See Conor Bateman’s article on the film for further detail. ↩

  5. Benoît Rossel, “Eric Baudelaire”, BOMB 140, Summer 2017. ↩

  6. I also wonder if fūkeiron is of a school of hard-determinism that’s hard to reconcile with the creative act. Eric Baudelaire, whom I think is the leading modern experimentalist in this area, seems to agree in his interview with Benoît Rossel, quoted above. See also his 2011 film The Anabasis of May and Fusako Shigenobu; highly recommended for its cinematic view of memory via landscape. ↩

  7. Another way to understand this is through Beer’s viable system model. The System Three management layer crafts purpose statements, but these must be mapped into the lower, more operational System One before they can be executed. This translation inevitably simplifies the diagram by framing it as if System One is all that exists. It also helps explain why org charts’ reporting relationships are inevitably so cleanly directed. Certainty and predictability are valuable for many good reasons, but after the inflection point they begin to lend themselves to less virtuous ends, for Seeing Like a State reasons. ↩