S06E19: Pens

This is my penultimate week, and I seem to be working in a state of penumbra. Not as much as others, for sure, by virtue of my current role, but still not exactly enlightened.

This week I’ve been struggling with communication. I know in theory how to communicate; that is is hard; and that in a large organisation you have to say the same thing until you’re sick to death of it before it breaks through. The issue I have at the moment is mostly born from the enforced living at work we’re all going through. Stuff that would normally be communicated through overheard conversations now has to be considered and shared, and sometimes the velocity we need to work at means we don’t have time to do it. I can hear my dad telling me that more haste means less speed – a phrase I found out recently that he nicked off Plato – but slowing down enough to do this stuff would mean things would break in the here and now. In theory we all agree that – for example – breaking a poorly-set bone is better than letting it heal crooked, but when it comes to it, it takes immense bravery to volunteer for the procedure.

I wonder if our perceptions of future rewards and future punishments are the same? I wonder how it influences us day to day when it comes to decisions or conversations we know will be unpleasant?

Anyway. This week I’ve had the really exciting opportunity to review all of the digital things that the entire department wants to do in the next 5 years. There are incredible, exciting plans there that will improve outcomes for citizens and colleagues alike. I’ve also been preparing sessions with different areas of the department, identifying areas where my immediate group could support colleagues in ways they hadn’t considered. It’s communication again – even within a department as small as ours, it sometimes feels like the only way we find out about projects is through serendipity. Disseminating context at scale in a way that’s not overwhelming seems like a fascinating challenge – for example, at GDS most teams write weeknotes. They’re great, and highly informative, but there are now 12-14 different blog posts I need to read a week just to stay on top of what we’re doing right now. And GDS is technically only a business unit – a sub-section of a government department. Imagine if I got weeknotes from the entire department – 14,000 people in who-knows-how-many teams.

I suspect that this is partly the driver behind dashboards, because they seem to solve that problem: context, in real time, but in a nice obvious format everyone can understand. And I think broadly speaking that the theory is good in certain contexts (though this guy reckons all dashboards are bad, so as always take what I say with a pinch of salt).

A dashboard ought to give you information that you can act on. On your car’s dashboard is a speedometer. If you press the accelerator, the car will speed up, and the speedometer will record that. Now, if you see your speedometer noting that you are going too quickly, you can apply the brakes. There is a neat causality between data on the dashboard and actions you can take that will feed back to the dashboard.

I suspect that a dashboard showing – for example – the number of infected cases is a fundamentally useless thing because we are operating in a complex system. And in a complex system, causality is a total mess. It’s as if you pressed on the accelerator and the speedometer sometimes did nothing, sometimes went up, and sometimes went down. There is no dashboard in the world that can tell you whether you’re ready to explore an unexplored space.

(Even as I write that I resist. There is something deep within me that revolts at the unknown; that imagines that if only I knew what the factors were I could have a dashboard that told me when to lean in for a kiss or what words were needed to help out a friend. The world, my soul seems to insist, is comprehensible, and given enough data I could have a dashboard for anything. Letting go of that idea is so very difficult.)

Because dashboards by their nature are necessarily simple, and therefore they can’t give you the context. With context you can be agile, because all agile means – really means – is “able to respond to new information quickly”. If your organisation is going to respond quickly to new information, then the whole organisation needs access to the global context and the new information. And as we’ve already noted, at scale, there can’t be a shared context because people would spend their lives reading what everyone else was doing. There’s no movement happening there.

This is why agility requires being small, and why my soul shrinks within me when certain large organisations insist they’re going to be agile. Agility is wasteful. Odds are, if you’ve got a number of agile teams, they’re all building the same thing. They’re probably billing the corporate credit card for the same thing ten times over, and they’re paying for it monthly. The finance director is tearing her hair out because consolidating it to a single contract for three years would cut that bill in half.

Can you countenance that much outlay, for the sake of agility? If you can, you’ve got to absolutely stick with it, and accept that other teams are going to ask why they can’t be ‘agile’ too, where agile to them means ‘buying whatever I want without regard for proper procurement’. And office space probably shouldn’t be bought on a whim (although, as you look around, perhaps you’re reflecting on how much real estate you actually need these days…).

I don’t know what the answer to this is, by the way. Any context that’s shared by thousands of people is (in my experience) only the vaguest, most meaningless guff. (cf. “corporate values”). Small, totally autonomous teams operate effectively but can’t effect meaningful change without joining up.

I feel like the core here is how to actually communicate context to everyone else, and if there’s a maximum size above which it just doesn’t scale any more.


I started playing Factorio again, after Terence’s glowing review, and found to my surprise I didn’t hate it as much as I remember hating it when I first played it. Clearly this is a sign of my advancing age, but the prospect of neatly organising a series of robots to do tasks effectively is much more attractive to me now than years ago.

This might have something to do with my ongoing obsessions with how to predict, and thus automate, when I need things like cat litter or coffee delivered.

Thus reminds me that I am currently campaigning for the English language to revitalise thence, whence, and hence, as well as their opposites thither, whither, and hither. Russian uses them, and they’re really great for cutting down character count. For example: “whither go you” is much easier to understand than “where are you going”, and even if it’s not it’s shorter and therefore superior.


Finally: I wrote a quick, well-tested bit of code that worked really well and has saved me several hours of boring clicking, and I am reminded again of what a joy coding is and how excited I am by the possibilities of digital systems that automate the boring stuff.


Finally finally, the Post Modern Jukebox concert I was hoping to go to has been cancelled so I’ll be listening to their entire back catalogue today. Please join me in my super-niche music tastes.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s