
I'm a designer, researcher, and maker løving in Copenhagen.

I'm [on Twitter](https://twitter.com/readywater) and [Insagram](https://instagram.com/lovettbarron)

I'm also on [Github](https://github.com/lovettbarron) and [Printables](https://www.printables.com/social/115124-andrewlb)


Andrew Lovett-Barron

Small Builds, Big Research


I am often “getting started.” Part of the theory of Diverge, to be honest, is being in that frequent state of beginning something new as past explorations manifest in new curiosity. Usually, these things end up being related. For much of this year, I’ve gone back and forth in 3D printing, modelling, and some rudimentary game development. That’s included a few (maybe helpful) diversions, like getting a bit too interested in flight simulators.

In each of these things there’s the necessity of having the basic things in place to get started, and transition from curious to amateur. I’m writing a bit about the nature of being amateur next week, but that space before being amateur — the state of becoming amateur — is something that I think is somewhat unnecessarily fraught thanks to the nature of the internet and its dominant business model.

Most hobbies and skills require some kind of infrastructure. It might be a book, it might be paints, it might be a machine, it might just be space. A fundamental part of navigating any new hobby then becomes navigating the infrastructure that makes that hobby possible. Take sewing. In theory, you can sew with a needle and thread. Some artisans even make a point of primarily working in that mode ([Mutsu is an example of that](https://www.youtube.com/c/MUTSUBYPROSPECTIVEFLOW)), but really you need a sewing machine to make a diverse array of things. But then which sewing machine? An industrial Juki can do a lot more than a home sewing machine. A [Sailrite walking foot machine](https://www.quiltersreview.com/sailrite-lsz-1/) (like what I have) works great for thick materials and allows me to make bags and accessories, but then working with thinner or stretchier materials becomes a hassle. Then you start getting into overlocking, coverstitches, etc. etc. etc. — it spirals quickly.

![](/assets/f62121c5864b61a52b2fe652d250ede6f7381bd3-1100x600.jpg)

When you’re first learning about something, there’s a fairly dangerous period between your desire to get enmeshed in the thing and acquiring enough knowledge to understand how different tools act together. During this period, you might be tempted to get cheaper equipment or, worse yet, find a getting started kit.

Never, ever buy a getting started kit.

Or, maybe reframed, never ever buy into the getting started kit.

Getting started kits are everywhere, and have an obvious value proposition: to the individual seeking to acquire new knowledge and skill, all the things one needs to get started are arrayed before you for some middling price. A fashion designer getting started kit might include a cheap sewing machine, pattern paper, measuring tape, pins, bobbins, etc. An electronic music starter pack (which would result in pain and horror) might include a Microkorg, an audio interface, Ableton lite, and some shutter shades. Neither of these kits really get you started.

The getting started kit is a commoditization of the journey towards having skill around a practice. It provides you with the artefacts of that practice, but not the points of reference around which those artefacts take on meaning and contribute to the practice itself. Take the electronic music example: the microkorg is objectively an awesome tool for getting started in electronic music. It’s also awful for it, depending on your need.

![](/assets/86c7796e6505dedd01837c03ac8d8e065df4ee91-1140x560.jpg)

Equipped with a vocoder, a mini sequencer/arpeggiator, 37 mini keys that are fiddly af, and a grid-based modulation section, it’s a really handy mini tool for performance and a fun tool for exploration early on. You also very, very quickly can get frustrated with it, and outgrow it — or at least on the surface.

The Microkorg’s synthesis engine is borrowed from the Korg MS2000, and they basically have the same overall capacity for audio synthesis. Except, ya know, the MS2000 is huge.

![](/assets/376d6f1768f39721816a153a958d5d3b3719ee26-512x250.jpg)

If I could repeat my synthesizer journey, I probably would have bought the MS2000 first and the Microkorg second. But in the language of getting started, the MS2000 is framed as being inaccessible, whereas the microkorg is more easily set as something commodified.

Part of my journey with synthesizers and music was getting utterly obsessed with the sound design of it. The microkorg is a great tool for performance and play — it’s small, easy, and approachable. The MS2000 has a bajillion knobs and settings and is seen as unapproachable as a result. Yet the transparency and access to those settings unlock a whole range of new possibilities with the tool.

Getting Started is a journey, not a package. Getting started in sewing for me was learning to make bags and backpacks, and then accidentally getting interested in garment making (which I’m terrible at). Getting started in synthesizers and music for me was being drawn in by bands I liked (using synths), and then falling down the Max/MSP and DSP rabbit hole. The few shows that I had the opportunity to play were done with monome, Ableton, and home-brewed VJ or audio software. What getting started pack would contain that?

The problem with the commodification of the Getting Started pack is that it locks you into a popularized and predictable model for getting into a hobby or domain — and maybe not one that reflects your needs as a creator. Had I grabbed the “getting started kit” for sewing, I never would have been able to make the cordura backpack that my wife often uses for travel, or the quilted magnetic wrap scarf I made for a friend. Picking my own machine and taking the time to orient the tools to my interests paid off in capability and low frustration. While it HAS proven frustrating from the garment making front, it’s also reframed how I think about making those garments: reinforcing stitches, thinking in terms of stress points, leveraging barracking, etc. My vision of garment making comes from bag making.

![](/assets/d08e9c1d73bb79e5af837439c441afeeef17cf0a-1556x813.webp)

Choosing your own adventure when it comes to your own, personalized Getting Started arsenal makes for the interesting set of constraints that shapes your own experience. Of course, you won’t get everything right the first time. And you definitely shouldn’t get the “best” of everything as described by wirecutter and YouTubers. But establishing your own personal constraints early on — and codifying those constraints in the tools that you choose to get started with — becomes the earliest seed of originality in what might become an incredibly fruitful hobby or path to skill development.

All of this said, sometimes a getting started kit is totally okay. Arduino and Adafruit make really good getting started kits. [Take this one by Adafruit for building your own computer with a raspberry pi.](https://www.adafruit.com/product/4796) These kits come with everything you need to build skill in a hobby and are “consumed” in the process of acquiring knowledge. Usually, a getting started kit consists of core tools you’ll be using over an extended period of time, and consequently, your constraints are dictated by those tools. But in these cases, the kit gets you started on a path to knowledge, where you end up being better equipped to make decisions on your next project based on that knowledge.

Just to wrap this up and give a bit of a TL;DR: we are shaped by our introduction to skills in rather nefarious ways. The getting started kit is the worst way to get into something because it takes us down a path of likely mediocrity, and doesn’t encourage us to pick up the tools that fit our skills and interests as we discover what those skills and interests are. They construct an artificial set of constraints around us that maybe don’t reflect the intrinsic qualities that got us interested in this thing in the first place.

Instead, it’s perhaps better to learn through doing small and easily discarded projects that don’t lock us into one tool path. We’ll accrue tools and knowledge along the way, and those tools will better reflect our interests in the thing itself.


Avoiding Kits


We’re in a golden age (cue: ominous sound that comes with a loaded word) of prototyping. For most digital designers, it’s never been easier to express ideas in forms that can be clicked, tested, interrogated, and discarded. This should lead to a Cambrian explosion of creativity and diversity — but instead, we find ourselves in a period of incredibly homogenized experiences. Websites follow similar patterns, coffee shops and restaurants barely differ around the details, and connected devices compete for control over your networked ecosystem.

Actually, this isn’t all bad. The obvious example is that many of these experiences are put together by absolutely tiny teams, and so the homogeneity comes from these teams leveraging existing tools. The creativity — at least for our purposes now, let’s call it the novel thing applied to a problem space that hasn’t encountered that thing before — is around some other margin that we might not even recognize until much later in the process.

So, in a way, our golden age of prototyping is actually facilitating something very specific: the ability to get creative on the things that normally don’t make themselves accessible to iteration and development until business and teams are quite established. This is the great side of this work, and honestly is something that facilitated a more strategic type of design work. But there is a problem in that it does really limit the playground around that core site of inquiry. The business model innovation you’re working on can be quickly realized through templates and off-the-shelf design systems, but that very act of framing narrows what might otherwise have been possible.

So, in a way, our golden age of prototyping is actually facilitating something very specific: the ability to get creative on the things that normally don’t make themselves accessible to iteration and development until business and teams are quite established. This is the great side of this work, and honestly is something that facilitated a more strategic type of design work. But there is a problem in that it does really limit the playground around that core site of inquiry. The business model innovation you’re working on can be quickly realized through templates and off-the-shelf design systems, but that very act of framing narrows what might otherwise have been possible.

Great design is emergent — from teams, from moments, from what we immerse ourselves in, from talent, from practice. But it also emerges from the iterative nature of design itself. I create a dozen, three dozen, a hundred things — from scratch or in parts — and the form of the things we design starts to emerge, and the decisions that we make around that design are pulled by the gravity of those past iterations.

So the real problem with this is that when we leverage these toolkits for moving and prototyping quickly without recognizing them for what they are — scaffolding — we run the risk of mixing the detritus of many, many past decisions in with our otherwise novel inquiry. A disruption becomes disrupted.

##

Sprues

There’s thankfully tons of solutions for this, and one that occurred to me recently was around kit bashing. I [shared a really interesting article about this last week](https://www.denofgeek.com/movies/greebles-how-tiny-details-make-a-huge-star-wars-universe/) about ILM using Kitbashing techniques to help design the “used future” feeling in Star Wars. Kit Bashing is basically just a technique for taking a lot of different scale model kits — a tank, a giant robot, a train, a car — and combining the individual cast parts into some greater whole.

That way, you can go from something like this (called a sprue card):

![](/assets/82c5cc6b4f899f34a0c9681f379d0ac336217f44-1431x820.jpg)

And, after combining it with lots and lots of other such contraptions, end up with something like this — a just-believable-enough take on another world.

![](/assets/1ca5ec1f4a396ae5ee8cc6ef7ee8893cb40920fc-1024x768.jpg)

Now, in theory, what I just lambasted as a potentially poisonous design approach falls into this category: after all, isn’t a toolkit approach the same as using spurs? What is a tool like [Uber’s Base Web](https://eng.uber.com/introducing-base-web/), or [Twitter’s Bootstrap](https://getbootstrap.com), or the [US Government’s Web Design Standards](https://designsystem.digital.gov) but a collection of UI sprues?

Kitbashing is different though. You can think of a lot of design templates as almost like sticker sheets: you get an aesthetic, a form, a prescribed function that gets pasted over different types of functionality — and those stickers end up being recognizable cues for the functionality underneath. Dropdowns select options, buttons trigger interactions, a large or small X demarcates exiting an experience, etc.

Kitbashing, on the other hand, takes many different pieces of potential functionality — an engine shape, a wheel, a barrel, an antenna — and mixes them up into different forms that eventually emerge something all-together different.

One of the classic kit bashing examples is from designer Kow Yokoyama, who created a [model making brand in the 80s called SF3D](https://en.wikipedia.org/wiki/Maschinen_Krieger_ZbV_3000). Working quickly with two friends in creating a narrative and visual aesthetic for the world, he took a large collection of existing scale model kits — planes, tanks, motorcycles, etc — and starting drilling, gluing, and sculpting connections between the disparate parts of those kits to create something entirely new.

Something like this:

![](/assets/b90741d3544b6cbc9c24a8c355c04391fa403ccc-793x570.png)

## **Kitbashing Experiences**

This mechanism for design is exciting partially because it empowers us to mix and match functionality outside of its usual context. Combined and remixed, they have the potential to create something new.

One example might be [Nathalie Miebach’s](https://nathaliemiebach.com) weather data visualization sculptures. Crafted using traditional basketweaving techniques, the technique is applied not to imply the form of the basket, but rather borrow a structured and quantitatively receptive process from which the forms she displays emerge. Combined with other disciplines like sculpture, the focusing backdrop of a gallery or museum, and the polish applied in colour selection and proportion — and something entirely new emerges.

![](/assets/091f71061e71c6fabc216a039392a9bd593a9eef-1440x1742.jpg)

Another example might be the game (which I’ve yet to try, but sounds incredible) [Carrier Command 2](https://store.steampowered.com/app/1489630/Carrier_Command_2/). You are apparently put in charge of an aircraft carrier, and command a drone fleet to capture territory around an archipelago. But it does this by combining a few different styles of play, and idioms of interaction: first person movement, flight simulation and “menu surfing”, top down strategy, and tower defense (which has become incredibly popular in the smartphone gaming era). Jammed together, the design of this game seems to rock back-and-forth between the different modes of play, using that frame to keep the player off balance and wholly engaged. Something I wonder is what came first: the idea of merging these different mechanics, or the idea of situating such a game on an aircraft carrier from which those mechanics otherwise emerged?

<iframe width="560" height="315" src="https://www.youtube.com/embed/w22l8tzaA2c" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

Finally on the web side, I might point to a weekend project I did two years ago called [Fit Faj](https://fitfaj.com). The idea was to try and build a mini business in just a few days (based off of my own experience as a new father), and I did it through form of kit bashing.

![](/assets/25829a2c2c5a2bf18c3413679f455dffa8b57528-2172x1602.png)

Taking a chatbot library meant for support purposes, I used that to structure the overall signup flow and write a small script that was friendly and let me tailor plans to specific needs and timeframes. This got mashed into a payment portal (Stripe) which I was already using for [Knowsi](https://knowsi.com), and which also connected into Google Sheets to give me a quick and easy sense of who my customers were.

Then, again using google sheets, I created a quick script that customized a plan I’d commissioned a certified personal trainer to develop, and scaled it to the needs of the customer. Then, the plan itself is tied to a basic transactional email system — sending the customer that week’s plan every Sunday so they can plan ahead without getting overwhelmed, as well as potentially update their plan depending on their adherence.

Addressing the challenge expressed earlier — homogenized experiences — this project came together as an experiment around designing a product and business that considered a very specific need, but tried to take a non-traditional approach to the experience of purchasing and delivery. Using chatbots and mailing lists, the different touchpoint for the experience came about not necessarily because I thought they were the best thing available, but rather because I had some spare parts lying about my desk (or hard drive, rather). Even design wise, it’s a kitbash between **custom components** (the images, the content), **templates** (the site structure is literally just a component from Knowsi), **libraries** (the chatbot), **services** (google, stripe), and **interaction modes** (web, mail, conversation, print-outs).

The neat thing about this approach is that it can be scaled and iterated upon with relative ease. If an experiment or part of the experience doesn’t work, discard it and try something different. If two parts feel disjointed, explore a smoother experience by adding something new to the mix, or rolling a small, custom interface.

Instead of building blocks — structured pieces that fit together — designing an experience through kitbashing is about turning that block around to catch the light differently, sanding off the connectors so it fits with a different brand, and gluing the whole mess together so that the component parts are only recognizable if you squint.


Kitbashing Experience


This is a continuation of the previous post — [thinking through research automation](https://andrewlb.com/thinking-through-automation/) — which deals more with a framework on how to approach automation tasks. This time, we’re going to dig into a specific example: how to automate a recruiting process.

Let’s say I’m recruiting from an existing and engaged user base, and I want to leverage some of the relationships that my customer support colleagues have in order to increase the likelihood of response. Instead of engaging in a back and forth with customers, I can instead provide the customer support team with some tools that automate intake — meaning that I can focus on preparing the moderator guides and answering any questions that might come up.

To do that, I want to break down the different facets of that recruitment process into steps that might be automated.

It might look something like:

- **Contact**: A CS rep sends a potential participant a mostly form email requesting participation in a user study (including timing, expectation setting, and gratuity amount)
- **Screen**: The potential participant fills out a screener form
- **Schedule and Consent**: If that participant isn’t screened out, then send them a scheduling link and a consent form
- **Remind and Confirm**: If they schedule successfully, then remind then send them a reminder the day before the interview. Also remind them if the consent form hasn’t been filled.
- **Interview Followup**: If they complete the interview, send them a gratuity and a thank you email.

Each of these steps can be automated to a greater or lesser extent, but what actually constitutes a step?

A step might look something like this:

![](/assets/86f749885fb392412fb30ea3a231cdbcc930a5dd-842x595.png)

Basically a point where an input is enters (data, interaction), is processed, and is then moved to another step (exported in some way. That individual process might have different features to it as well:

![](/assets/ff1fa37af2dd2611e56fe2adbec879203297d836-842x595.png)

If we were to break the above list down into a more linear process flow, we might end up with something like this:

![](/assets/7ec3520301ad343faafb4485829f84c7bb361aed-842x595.png)

This flow helps us interrogate a broader system, and figure out what individual steps in the overall task we might seek to automate. But we need the right tools to facilitate that automation/

## **The tools**

There’s many, many different tools out there that range in complexity to setting up small web services (maybe using a scheduler and python) to connecting automation functionality on our desktops (eg. OSX’s Automator or IOS’ Shortcuts), to orchestrating various web services to get the job done. For our purposes, we’re going to start at this “easy” side of web service orchestration to implement this example.

[**Airtable**](https://airtable.com/)

A simple relational spreadsheet tool, Airtable has a useful data structure for facilitating different human and automated “fusion” tasks.

By using its concept of filtered views, you can create different dashboards to react to and respond to data, as well as automate those views. In fact, [this newsletter](https://divergeweekly.com) is substantially automated through Airtable — combining data intake and specific views that give me a “todo list” of curation and creation that I can power through each week.

Additionally, one of the Airtable views is form input, giving us a simple survey tool that we can automate.

(If you want to give Airtable a try, [I would appreciate your using this referral link!](https://airtable.com/invite/r/7cLYgvsO))

[**Calendly**](https://calendly.com)

Calendly is a simple scheduling tool. I use it for my [office hours](https://calendly.com/alb/officehours) as well has research recruiting. It can be set up to send out reminders and combine with other systems. An example of this is that when you sign up for office hours, for example, I receive a new todo item in my Omnifocus todo list.

[**Knowsi**](https://knowsi.com)

Knowsi is a consent management platform for research that I created. To collect consent, you can include your project’s form link directly in the email, and the researcher will be notified when that form has been filled. While knowsi currently it doesn’t integrate with other platforms, but we’ll be adding Zapier and IFTTT integration this year.

[**Tremendous**](https://www.tremendous.com/)

A simple gratuity management platform. You can create projects and budgets, and then feed a list of names and emails in to dispense gratuities. It will normally send a gift card as well, so there is a greater degree of flexibility in terms of who can accept the gratuity.

So far, I just automate this by feeding a CSV of spreadsheets and names into the tool at the conclusion of the study.

[**Zapier**](https://zapier.com)

Zapier is a glue between different tools. It’s an easy way to, for example, connect an RSS feed to Airtable, or automatically check that data exists and update a list, or delay receiving a notification until you’re ready for it. Zapier makes a lot possible that would otherwise require complicated and custom integrations.

## **Bringing it together**

Looking at each of these tools and the role they play in input, storage, processing, and connecting, we can start to create a picture of an orchestrated and automated flow.

![](/assets/6675224de919cc18eef65195007b4f806f6df6f5-842x595.png)

### **Contact**

This is a situation that we can’t really automate because it requires a degree of human connection and problem solving. But we can create information that helps start the process.

Generally, you’d want to create a description of the study, how long the interview will take, timing, gratuity, etc. And then a link to the screener. [I have a good primer on writing these here.](https://andrewlb.com/the-anatomy-of-a-research-recruitment-email/)

For our purposes, we’d list our potential participants in Zapier first, and then automate the introduction email either through Zapier or Airtable. If there’s no response, we’ll generally want to automatically remove them as well using a scheduled task from either of those tools.

### **Screen**

When creating the screener, this can be done easily in airtable. Here, I would create an intake spreadsheet with a form field, and then create two views: a “form” view and an additional grid view that has a filter. Edit the form in the visual editor to fit what you like, and then fill it out.

![](/assets/25b4968e992b5bf1b270b6bb4124b46166e02f91-1684x740.png)

In the second grid view, add a filter (like above) that will only show those respondents that fit the criteria you’re looking for.

From here, we want to create a new spreadsheet which will be filled in as our screened and ready participants filter in. We do this by creating an empty table, then reflecting the data that we want in that new table through an automation. It’ll look something like this.

![](/assets/f8227c912deebd4557a2b9c7eb0a1b4105786303-1684x1190.png)

Finally, from where, we can set up a new email to send out when records are “Created” in this participant table. This allows us to divide the records between consenting participants in the study, and our list of potential participants.

### **Schedule**

Scheduling can be both incredibly easy and incredibly hard. I’d suggest using an automated scheduler like Calendly whenever possible — but provide an option to reach out with different times if none of the Calendly times work.

![](/assets/e0e99173acf97a189bc70ccefc4fb9a96b6ee88e-842x595.png)

You can then integrate Calendly with Airtable through Zapier, and update the participant records directly with the time and place of the interview, and a flag saying they’ve moved onto the next step.

### **Consent**

Consent forms, like scheduling, should be automated wherever possible using tools like [Knowsi](https://knowsi.com) and Zapier. Privacy regimes like the GDPR in the EU and PIPEDA in Canada mandate progressive informed consent practices.

![](/assets/3854b1dafe86f742480b54d278036c5b4d5f6c42-842x595.png)

This step can be pretty simple — just send the consent form via the initial email or when the participant has completed scheduling, and the researcher will get a notification on when the participant has signed the form.

### **Remind**

Content-specific reminders can be sent out to participants who have stalled anywhere in the process.

![](/assets/a858dfae43d9dd4e54e595c55073dd8975670eda-842x595.png)

In zapier, you can set up zaps to do multiple checks on things — though I personally prefer to set up unique views in Airtable so that I can both visualize and automate that reminder process.

One example might be passing a screener but failing to schedule. Zapier and Airtable can be used to send a reminder once or twice before removing that person from the list of screened participants.

### **Followup**

Finally, upon completion of the study, the gratuity can be (mostly) automated as well by creating an automated export of participants and their emails for upload into a gratuity management platform like Tremendous.

![](/assets/7b8de2f67abb1266ad23c21668edfa6c90d41501-842x595.png)

Tremendous doesn’t have great automation tools, but you can bulk upload a CSV to their platform. So you can set an automation up to send you a weekly CSV of participants who have completed the study, and simply drop that CSV into the Tremendous platform. This is still a little bit of a manual process, but it turns what might be a half hour or hours worth of work into a three minute task.

## **Many Small Experiments**

Ultimately, working with these tools takes time — both in learning how to use the tools in question and in setting up the automation. While I’ve used a couple of examples here that I’m familiar with, I’d encourage you to set these up with whatever format you feel most comfortable.

Generally though, get started with what you know, and try to do small but structured automation experiments to build up to a larger automation workflow. **Automate one step first**, and then look before and after that step to see what you might do next.

In the end, Automation frees up time and energy better research and can turn the frustrating time-sinks into interesting problems to solve

Also, Automation — when it fails — can reveal the real points of complexity and value in our process that can become things we double down on.

So. This is a concrete example of something that you can automate and how you might go about doing it. Last week we looked at [some tools or ways of thinking about automation as a concept](https://andrewlb.com/thinking-through-automation/). Next, we'll take a look at judging what should and shouldn't be automated.


Automating Research Recruiting


There are professional Cassandras out there and their curse isn't coming from Apollo this time — but from the institutions they serve under.

Warning intelligence is a function of the intelligence community that manifests in different ways, but at its best tends to come about as a small and senior team filled with experienced professionals with the ears of decision makers. The problem is, that while this structure emerges with some frequency (the CIA's National Intelligence Officer for Warning, or NIO/W is an example of this), these offices tend to pretty quickly get a reputation for crying wolf.

9/11 tends to be the most widely known failure in warning intelligence that many think of — a coordinated strike on American domestic targets, both civilian and military, by a well established terrorist network, whose rhetoric and actions forwarding a religiously-framed war against the United States were known to the American IC. But actually, you could argue this wasn't a failure in warning intelligence — whose purview fits the six month to several year timeline according to one of its great practitioners and thinkers, Cynthia Grabo — but rather a failure in tactical intelligence and interagency coordination (which is one of the arguments put forward in the [9/11 commission report)](https://www.9-11commission.gov/report/911Report.pdf). Warning intelligence had actually done its job. In the chapter "The System was Blinking Red," the commission outlines how many of the warning functions were operating well — sharing indications that major attacks were in the works — but the different security functions of state were not. In many ways, Cassandra was being ignored.

![](/assets/6e4ca21ef97fa44cdd5ecda5d3626fe9f8ca8525-667x673.png)

Most of us operate less in the present, and more as a kind of smear between past and future — with our present moments coalescing like a Gerhard Richter painting. Strategic thinking at its best orders this natural chaos: providing the structure and tools to look ahead, look behind, and actually make decisions in the moment that support futures. Ladder that up to decision making for not just oneself but others' time, resources, etc. and you get some pretty clear examples of strategic decision making. Taken up even further — with thousands and millions (the American Department of Defense has about 1.4 million employees when considering both civilian and uniformed personnel) people relying on different tiers of strategic decision making — and the impulse is to embed those sensory and analytical functions within specific teams and career paths. This is what makes it possible for an elected official with a four to eight year mandate to move into the chief executive role and actually do anything coherently.

Warning intelligence is one such analytical organ within the outsourced and institutionalized cognition of a chief executive. But like the weird pain or feeling of apprehension that we might easily ignore, warning intelligence isn't the thing driving the ship. Rather, it might try to influence at best — but often gets ignored.

During his presidency, Bill Clinton didn't have a great relationship with the security establishment. This was during a period when the somewhat bloated American military was being reduced following the close of the Cold War, and American engagement in overseas adventurism was manifesting in some weird ways. Amongst these was a loss of the "great enemy" of the Cold War, around which national security interests in America had swirled — which in turn shifted how the American public saw foreign wars. America's engagement in the UN mission in Somalia during the Somalian civil war, for example, was initially seen as a natural extension of American hegemony, but the Battle of Mogadishu (Black Hawk Down) in 1993 put a pretty stark end to that with the resulting death of 19 servicemen.

These events led to a kind of souring on intelligence and foreign engagements by the Clinton administration, who felt like they were being pulled into these engagements, less because America should be involved, but rather because it had the strongest capabilities — part of what drove the multilateralism that was dominant during his presidency. But this view also served to block out the warnings that the Clinton admin was receiving about other regional conflicts. Following the Battle of Mogadishu, Clinton was said to have an explicitly pessimistic view on further military engagements in Africa. He didn't want to know. But the point of warning intelligence is that things happen ANYWAY, irrespective of whether or not you want it to or not. So despite warning intelligence indicating that stockpiling and ethnic tension were occurring in Rwanda, the Clinton government [didn't do a thing](https://qz.com/384228/the-clinton-foundation-is-atoning-for-bills-failure-on-rwanda/) as a million Rwandans of Tutsi ethnicity were massacred by fellow Rwandans of the majority Hutu ethnic group.

![](/assets/92dab1148723517d9db36cbd335a95e08c582386-1534x979.png)

Warning intelligence doesn't make the decision, but rather serves as a function in intelligence with two primary roles: **to track uncertainty and threats, and communicate effectively to decision makers when those threats manifest into risk**. When driving, our eyes move constantly: flicking across our vehicle's status indicators, the road (and its external status indicators, like dividing lines and shoulder), other cars and their trajectory and the mirror, to alert us about things from behind. If we see something that's a threat (like a big truck pulling up alongside, or one that we're trying to pass), we're a lot less anxious if we see it coming. The risk is small and we have contingencies (maybe pulling onto the shoulder, maybe deciding not to drive between two trucks, etc). But if we're reeling from a near miss moments earlier, or are caught unaware by its speed, then the potential risk we're taking in responding to this new threat is substantially greater.

Later, the Clinton admin had acknowledged that many, many Rwandans could had been saved with a relatively small American deployment of just a single division (around 10,000 soldiers). But the proximity to the Battle of Mogadishu meant that even though warning intelligence was serving its function with its analytic output, it had been systematically locked out from its communication function. Apollo's curse manifested in cancelled meetings, edited documents, and shifted priorities.

These days, there's more of a structure that John Gentry calls "Every Analyst a Warning Analyst" in his book [Strategic Warning Intelligence](https://bookshop.org/a/19778/9781626166554), which basically tries to diffuse this warning function across many different analysts. The problem is that perceiving threats from the miasma of risks and uncertainties out there requires a particular disposition and is frankly a job. Thinking about something really, really hard — worrying about it until something emerges and then validating that what emerged isn't just in your head — that's a full time thing.

In a lot of orgs, designers and researchers can play this kind of role by serving in a kind of foresight and prototyping function. While not attuned to decision making per se, the designer and researcher surfaces the right context for product decision makers to move effectively. In others, product owners do this type of work, looking ahead to where something needs to be and gaming out the risks to the product and the business surrounding it. These risks are surfaced incrementally in roadmaps and are responded to and acted upon by decisions makers.

I'm personally curious to learn a lot more about these kinds of institutionalized foresight functions within organizations (though I'll say that I think warning is different from foresight in specific ways — both in lens and communication), so expect some more writing like this.

However, if you're interested in warning specifically, take a look at some of the links spread throughout, as well as some of these resources below. Cynthia Grabo's book is free and was recently declassified after decades of service as a key resource in strategic warning training. I also wrote about Gentry's Strategic Warning Intelligence book, which is fascinating and very much worth your time.

## Some further reading

[National Intelligence Warning: The Alert Memorandum (U)](https://www.nsa.gov/Portals/70/documents/news-features/declassified-documents/cryptologic-spectrum/national_intell_warning.pdf)

[Cynthia Grabo's Anticipating Surprise](https://www.ni-u.edu/ni_press/pdf/Anticipating_Surprise_Analysis.pdf)

[Strategic Warning Intelligence](https://andrewlb.com/reading/strategic-warning-intelligence/)


Professional Warning People


_(Originally written March 23, 2020)_

Why is it that we can see something coming and still not prepare? I’ve been asking myself this question over the past 10 days that we’ve been on lock-down here in Denmark. While we set ourselves up pretty well (how is besides the point here), we set ourselves up in a very general way. We have some medical supplies and training, we have food, we can keep ourselves and our son protected… we think. But what if we could be more specific?

It was this question that spurred my wife [Ayla](http://www.aylanewhousedesign.com/) (a design director at Fjord) and I (design all-arounder and co-design consultant) to do what we have consistently done in our relationship: turn to goddamn post-it notes.

We realized that we needed a different tool to deal with this complexity, and so we created a game. It’s not — strictly speaking — a particularly fun game. It’s not dealing with a particularly fun topic. But it follows that structure and it was (and is) useful to us. **This is a long post: if you want to jump to the rules and the tools we created, just head to the end.**

## Anxiety, Collected

Our first task was to enumerate our anxieties. That was fun.

![](/assets/6572a232ac59347bbf3be5253919fce03de6e803-2048x1365.jpg)

We started off just listing everything we could think of

Sickness and death in our family (all on another continent), our son Emil getting sick, losing our job(s), overwhelmed public health systems, closed borders, and economic bailouts failing were the obvious near-term concerns. Greater extremes came quickly after: governments focusing on their citizens above foreigners (we’re foreign workers in Denmark), seized assets. Even broader: what of a European Union unable to fulfill its “union” mandate? What of existing fragility and its fallout, such as armed conflict between Turkey and Russia?

These anxieties are overwhelming when taken together, and can quickly spiral. I’d argue they are effectively the same as reaching out frantically in the dark from one's current position. What we needed is a way to tackle these anxieties in sequence.

In the past, we’ve often used a tool like [backcasting](https://en.wikipedia.org/wiki/Backcasting): looking at a future or point in time that we WANTED to get to, and figuring out the steps required to get there. But we aren’t in that kind of situation. **We aren’t defining a strategy so much as developing a posture from which to engage whatever the world is throwing at us in a time of extreme uncertainty.**

We want to build our resilience. We know (insomuch as one can) where we are right now: so what does it mean to scout forwards in time?

## Anxiety, Organized

Our next step was to simply cluster the anxieties where they seem to make sense. A few obvious ones emerged:

![](/assets/30b87b85e4267912061d5e400b1fdcab42753aec-3024x4032.jpg)

Some very initial clustering

### Environment

These were things like “One of us gets sick”, “Emil gets sick”, or “One of us loses our job.” Events that we are living with, in a time of change or decision making. Something that is probably traumatic, but also something that frankly most of us will be experiencing in some way over the coming months.

### Resources

These started as inflection points, things like “Hospitals get filled to capacity”, “No testing means no idea of actual risk”, and “Internet rationing or curfews.” But these clusters could be simplified down to “Public Health”, “Risk Appetite”, and “Community and Connection” respectively. Either private or public resources that shape outcomes. There’s a cluster of other ones as well, including Cash, Consumables and Food, Sanity, etc. Ultimately, anything that is a resource in balance.

### Inflection Point

Inflection Points are specifically in tension with the Environment category and have a lot of overlap. These are mostly framed as a CHANGE that we will have to make a decision around. An example we listed is that “Denmark starts emphasizing Danish citizens over foreigners for medical care” (unlikely but possible), “Indefinite lockdown is announced”, and “No one is hiring.” In each of these cases, it requires a non-obvious response given the reality we find ourselves in.

### Catastrophe

Finally, we have a “Catastrophe” category, which is something that throws an already vulnerable situation into complete chaos. This might be the death of a parent or partner, assets being frozen (for whatever reason), a national or international crisis, or — most likely — a full recession. Significantly, these are things that might actually be quite probable: but their likelihood doesn’t dilute the extreme emotional, financial, physical, and psychological impact of the event.

## Anxieties, Iterated

![](/assets/69f1711ead24889aa6c20fe083b60f5fd43e3971-3024x4032.jpg)

This is the part where things started to come together. Ayla and I were trying to figure out how to actually approach the forecasting part. We landed on a grid that looked a bit like this:

The initial canvas

![](/assets/aca63a7cf2e27d78aaf4b9150907299a2185cdf3-1440x1024.png)

First, one axis became time extending from now, +3mo, +6mo, +12mo, + 18mo, and +36 months. Then the other axis became the “order” of decision: environment to resources, resources to inflection point, inflection to catastrophe, with the two assessments beside it.

### Setting the Scene

First, **Environment** cards were chosen based on what we were facing right now. (We started with one of us being sick, though hopefully, we aren’t). Then looked at the **resources** that we had access to and what we were lacking. In the above scenario, we’re well set up for supplies and consumables, but honestly, don’t feel well covered by the public health system. This is an entirely subjective judgement: we might actually be well covered, but it’s just the feel there.

Next, we move on to the **Inflection Point**, which in this case is the made-up “Denmark explicitly emphasises Danish citizen wellbeing over foreign residents.” This being purely speculative, we wanted to explore what would happen.

![](/assets/c15ca7dc85e6a737f710947ea2cd14ec2b3fcbfc-2048x1365.jpg)

Looking at the first scenario

### The First Assessment

![](/assets/68d3bd5aabfaf43d6eeee82ed84a8da40bb72078-1440x244.png)

The first assessment canvas

After we laid out these four cards, along the “Now” time, we asked ourselves: “Okay, what would we do?” And at this first assessment, there actually wasn’t much. Even in a scenario where we were sick and our health as foreign workers was being devalued, we’re healthy, have resources, and have a community. So we decided that we’d simply stay home, and try to weather this sickness as best we can.

### The Second Assessment

![](/assets/31517ea3b13e68abca2a2df0855db9bab4b93531-1440x244.png)

The second assessment canvas

Then we rolled a dice. We had each of our catastrophy cards labeled with a number, since for this initial exercise we only had a few. And it rolled a 4: “A Death In the Family.” **This was the moment where the game hit home and we both said “Fuck” in unison.**

In a scenario where our own health was at risk and we had made the decision to shelter in place, suddenly our sibling or our parent calls us with crushing news. We’re between an 8 and 12-hour flight back — itself a multiple of risks to ourselves, our loved ones, and those whose paths are simply coincident with our own.

This is where the second assessment took place. We listed what would likely have to happen: hunker in place. Look after our physical health. Look after our mental health. Find a grief councillor who can work with us remotely. Redouble our efforts to stay connected with family. Take on the responsibility of planning and estate considerations, or support the family member who does.

### The Next Turn

Exhausted by this, we drew the next card: 3 months from now, in June. Suddenly, one of us is unemployed. We select our resources: we have good savings, but in such a situation our risk tolerance is going to drop (meaning we might not be able to jump on opportunities like we had in the past). Then a corresponding inflection point: “No one is hiring.”

We laid out our first assessment, which actually did not look too different from what we would do outside of a pandemic and looming recession. But the catastrophe card threw this into stark relief: the borders were fully locked down. Suddenly, even though we had not been planning a move, the scope of potential forward actions dramatically shrunk. Maybe the embassy? What do we do?

## Anxieties, Acknowledged

We played several rounds of this, seeing another family member pass, seeing cabin fever set in, and finally seeing both of us face unemployment. Each of these turns as they marched forward in time prompted stark decision making and discussions on what might come.

Ultimately, we looked at our decisions and took stock — what comes next? If these things might happen, what can we do now to offset the risk? If joblessness is a likelihood (or at least a distinct possibility), can we plan our capital and budget accordingly now as though we anticipate that? If the death of a family member is possible in the near term instead of the indeterminate future, what needs to happen?

**This wasn’t a fun game.** It was upsetting, painful, and anxiety-inducing. But though crudely simulating what might come, it helped map out many of the blindspots and fragilities we’ve had all along. And maybe, if we run through a few more times, it’ll mean we’re in a better posture when the time to act comes.

![](/assets/a3ed73356f37bc0044786aac36a91d8dba9b4271-2048x1365.jpg)

Dealing with the first game required a beer and serious decompression

## Rules of Play

Okay. The above was a long diatribe about why we made this. Here’s the TL;DR on the rules and how to play.

### Setup the Board

1. List out everything you’re thinking about on post-its or small squares of paper.
2.

Organize into:

- Environment (something ongoing, like sickness or unemployment)
- Resources (cash, community, etc)
- Inflection (an event that forces a decision)
- Catastrophe (something big and unexpected)
- Wildcard (Something possible but very unlikely, like a shooting war between Russia and Turkey or the EU collapsing)

1.

Create two axis:

- one of time starting from now,
- one with environment, positive resource, negative resource, inflection, catastrophe, wildcard. Then two more columns: Assessment 1, Assessment 2

[You can download the canvas here](undefined). We didn’t play with the wildcards initially, so didn’t include that on the canvas.

### First Round

1. Select an “Environment,” either what you’re most worried about now or what is already the case. Put that on “Now” and “Environment”. Then select two resource cards for what you have access to (positive) and what you are missing (negative). Then select a random inflection card.
2. Think about what you would do in this scenario and discuss. What do you hope you’d have? What kind of decision would you make? Write it down on a post-it and put it in assessment #1.
3. Then roll a dice or select a catastrophe card at random. Put it in the catastrophe slot.
4. Reflect on and discuss the new scenario given your previous decision. What changes? What can you do? How bad are things? What opportunities might exist? Write down your reflection and put it in assessment 2.
5. Normally the round is over, but if you want to add something, add an additional wildcard and an assessment #3.

### Subsequent Rounds

Each round builds on the last, so if your next time slot is three months from now, then it is following the decisions you made in the “Now” timeslot. Simply follow the same steps as the first round and see where it takes you.

### Reflection

Play until you feel like you have an idea of what might come. Before wrapping up, look at your assessments. What can you learn in terms of preparing for some of these worst cases now? Is there anything you need? Decisions you need to make? Processes to put in place?

And then if you like, play another round with a different starting point. Who knows what different decisions you might make, and what you might discover you’ll need on the other side.

### Resources

- [List of topics to help get started.](https://docs.google.com/spreadsheets/d/1kxSbkwnDXJZfGT_-6vx8Yu8Xa4zb8taRV3CGYYlHy5E/edit?usp=sharing)
- [Template canvas to print out](https://andrewlb.com/f87d372c75f53c5a605f786953391152/canvas.pdf)


Covid Anxiety: A Foresight Game


I love Rick Owens’ runway shows. They’re a consistent (and he is remarkably consistent) source of visual inspiration that I find bleeds into a few different sources of cultural and creative output. And as a designer consuming design work from a different material context, it’s almost inevitable that I get drawn into investigating process.

![](/assets/637d3398841e314a3ab3d013dcb18e0f9a19fca7-1600x2400.jpg)

Across a few interviews that I’ve been reading, one of the things that stood out was the way Owens engages with the trajectory of his work.

_“I’m always looking at the arc of my work and thinking about the next logical step. The fabric comes first: I usually start with 50 fabrics and edit them down to a dozen. Then I’ll work with black-and-white copies of archive looks, cutting up and manipulating them, or redrawing lines with a marker. I can plan a whole collection that way. I’ll ask my factory in Bologna to make toiles and, over one weekend, when there’s nobody else around, I’ll be in there working on them.” ([Interview w/ Rick Owens](https://www.matchesfashion.com/intl/mens/the-style-report/2017/01/the-paris-fashion-week-issue/my-desk-rick-owens-designer-interview))_

A toile is a mockup of a garment made in cheap cotton fabrics, called muslin or calico. It’s a way to test the construction, drape, and fit of a garment in a fabric similar to the weight of the end garment, without taking on the costs of finer fabrics — which can get incredibly expensive. In the earliest stages of designing a garment, draping muslin on a dress form means cutting, pinning, and manipulating that fabric into the forms that map onto the vision that the designer is bringing to their work. He’s taking his previous refined works as prototypes to interrogate and remake in a new collection.

![](/assets/44f9dfd5f1c61659602eb723cfd555795075edbd-2360x1428.png)

By working in this way (and Rick isn’t unique here), he is actively weaving the trajectory of his previous work into the demand for something new, with an upcoming collection. A silhouette might become accentuated or diminished through paneling; or a shoulder made more extreme or muted when unpicked on a toile — all incrementally moving towards the output that is ultimately communicated through polished forms and the communication medium of runways and lookbooks.

At its simplest, design is a perfusion of inputs into a process that results in an output. The work of design is the discipline of structuring and metabolizing those inputs into a form that — broadly — reflects those inputs. Perhaps more than most, designers are what they eat. But neither are designers distinct in this.

![](/assets/12c704a82eff6f81d5040c9986de11b7ba181351-850x461.png)

The OODA loop — articulated by the F16 designer and military strategist John Boyd — is a good example of how these actions are actually fairly banal. The OODA loop represents an observation, orientation, decision, and action loop that everyone is going through all the time, as we consume new information and decide on what it is that we’re going to engage with. Boyd’s idea was, that by shortening a pilot’s OODA loop, it would make them more competitive in a combat situation. For most of the rest of us, the OODA loop serves to frame how many of our actions are defined by rote response (jumping from orientation to action without engaging with it critically), analysis (jumping from orientation, to decision, back to observation without acting), instead of the full loop — conscious engagement with our influences as we move from decision to action.

Boyd’s model does an interesting job of balancing between the introspective and critical lens when engaging with decision making, as well as the more animalistic and reflexive lens — the action that bypasses critical consideration. In Boyd’s model, this bypass lends speed to action, by short-circuiting the decision step with instinct and directly leaping into action. What we maybe don’t discuss enough, is how that same tension exists in the quotidian grind of design work.

![](/assets/1a2384c4fb04c3f8ab1fc05f28187c36351e43bc-1600x2400.jpg)

Rick Owens' work is incredibly emotional and physical. It’s intuitively and perhaps natively evocative of fantasy, sensuality, violence, and a kind of fragile calm that initially belies the intellectual underpinning of his approach. But that only makes sense as a tension, if we believe that the reflexive and intuitive is in contrast to a more intellectual stance, instead of complementary to it.

One of the things that Rick Owens’ work has taught me so far, is that the critical and the intuitive flow from the same source. But frequently, we’re so wrapped up in suppressing one or elevating the other that we don’t make space for the two to intermingle effectively. My sense is that Rick’s process of catabolizing his own work is at the core of this creative process — and a key to focusing on his instincts. His process critically engages with the more intuitive and emotional aspects of his identity and process — using that short circuit as a tool of predictable creativity that others might dismiss as irrational or undesirable.

![](/assets/1c6a9101363d618fb84f9cb192c68e0cbe1a2e66-1800x1800.jpg)

The metaphor of metabolism is really at the centre of a reconstructive design process — old material is broken down for energy through catabolism, and new work emerges from its predecessors through an anabolic process. Repeated over and over, and the form that emerges is new, yet not. Structured, yet emergent. Consistent, but novel. A reconstructive design process emerges over a lifetime and, properly fuelled, creates a legacy.
