Last week my garage door wouldn’t open, leaving our cars trapped on the inside. My wife needed to leave to go pick up my daughter from her activity, but the door wouldn’t budge.
After a bit of troubleshooting, we figured the only way to get out is to do it manually. The only problem is that the door is a large insulated door, weighing in at over 350 pounds (160 kg).
I was just reading something about how you have to train in advance of when it’s needed, because by the time the issue arises, training will be too late.
Thankfully my wife and I have trained in advance to lift heavy things, so working together, we were able to raise the door and get out.
If I had waited until that moment to get strong enough, it would have been too late 🤣
It was a good reminder to keep investing in training and personal development to be ready for unknown challenges in the future!
Kevin
A Quote
“
Have you ever found yourself stretched too thin? Have you ever felt both overworked and underutilized? Have you ever found yourself majoring in minor activities? Do you ever feel busy but not productive? Like you’re always in motion, but never getting anywhere? If you answered yes to any of these, the way out is the way of the Essentialist.
— Greg McKeown in "Essentialism"
Three Things
1 - 📗 Radical Delegation - I really liked this book on how to work with an assistant. It’s very focused on meeting preparations and post-meeting processing, but those are indeed very helpful. It also has suggestions on how to leverage AI if you can’t yet afford a human assistant.
2 - ⌤ Chevron Deference Overturned - The US Supreme Court just overturned a ruling referred to as Chevron Deference. It’s complex, as these things always are, but on the surface it seems like it’s moving authority away from information, which I generally don’t prefer in business. On the positive side it will force more clarity as ambiguous statutes pass through courts. It’ll be interesting to see how this impacts businesses (like the fishermen who brought the lawsuit).
3 - 💨 Fart Diffuser - I met a VC whose investments include a fart diffuser. I feel like this has to be half joke, half not? If you're questioning the merits of your own business idea, just remember the fart diffuser and get started.
Deeper Dive on Thinking Architecturally with Layers
Architecture is a term, like strategy, that’s difficult to understand at first, but important for leaders to learn. Architecture affects the speed with which you can build new things or make changes to old things.
Leaders have to be good architects because speed matters in business. Speed is cost. Having a team with fixed salaries, the longer it takes to build something, the greater the cost. A competitor with better architecture is going to compound their speed, take your customers, and leave you in the dust.
Today I’ll walk you through an architectural framework on layers. There are fast-moving layers and slow-moving layers.
You can improve your skill with architecture by thinking in layers, understanding why layers move at different speeds, and using that information to inform your solution designs.
I’ll explain the framework and show you the implications for your strategy or transformation work. Enjoy!
Example in Life
To introduce the idea of layers, let’s start by exploring one in general life. Let’s start by looking at a relatively fast-moving layer: fashion.
Fashion styles change pretty quickly. Every season new styles are released. Younger people dress differently than older people. It’s pretty dynamic! It also doesn’t cost too much to change your style; you just pick up a new shirt and your entire style is transformed! You don’t need new hangars or a new washer and dryer. Your new item of clothing is compatible with all of those things.
Below fashion could be building architecture. Building designs change, but more slowly than fashion. You could build a house with different doors, window styles, ceiling heights, etc., but that’s pretty expensive. All the downstream things are designed assuming certain standards, so they all become impossible or crazy expensive if you change building styles too much.
Even more slowly moving is language. Yes, different generations experiment on the edges and create new slang, but the core grammar of a language doesn’t change much.
Below language could be our social structures (how humans live together is pretty static). Below that could be biology (our underlying genetics moves glacially). Below that is the laws of physics, which as far as we know are permanent features of this universe.
All of these different layers evolve at different speeds, and they’re related. The deeper the layer, the more implications there are if it changes.
That outfit is still machine-washable.
Example in Business
Now that we’ve got the idea of layers, let’s transition to a business example. Since just about every business looks at data, let’s use a reporting example.
At the top layer you’ve got colors and other easily changed visual elements. Should the line be blue or green? Should it be dashed or solid? Those are like fashion; very quick to change. They’re inexpensive changes, too; an analyst could change a color scheme in minutes.
Below that are your visualizations. Line chart or bar chart? Should we render two charts next to each other, or one big chart? This is a little more expensive, but they can still be changed relatively quickly.
Below that are the metrics (revenue or whatever you care about) and their aggregations (e.g. median, average, percentile, etc.). Changing a metric formulation is still possible, but it’s starting to get costly. Metrics are used in many places, so if you change your underlying formulation, you’ve got a lot of changes to coordinate, and users to educate.
Next you might have your data architecture and database schemas. Below that you have event logging, then an application. Then an ecosystem of applications, which are built on hardware infrastructure. All of which is communicating over a networking architecture.
Just like in our life example, you can see that some layers move quickly (colors on your chart), and some move slowly (network architecture). Top layers can be changed more quickly and less costly. Deeper layers can get real expensive to change.
These colors are easy to change.
Deeper Layers Move More Slowly
Why do the deeper layers move more slowly? Every layer on top depends on the layer below it. You can’t change a deep layer by itself because it’ll impact the layers above it.
That means scope increases the deeper you go. Not only are you changing the layer you want to change, but you’ve got to understand the implications and change multiple layers to accommodate.
What we’re describing, in part, is “tech debt.” A change might be required in a deeper layer, but because of all the additional scope - and time and expense - businesses often choose to keep things the same rather than bother with improving it.
Because businesses avoid changing these layers frequently due to cost, by definition they’re slow moving layers. This is important because it has implications for how we approach our work.
We can use this layers model, and an understanding of how speed (and cost and time) are affected by these layers, to improve how we approach our strategy and transformation work. I’ll share four ideas below before we close out our exploration for the day.
1️⃣ Think Deeper
If you want to make a change to a top layer, think through what needs to be true in the bottom layers.
As an example, let’s say you want a new metric for your business. It’s pretty close to the top, so it should be pretty easy to change.
That’s true, but only if the information you need is available.
Maybe the events your data team need are in a raw data table, not in your production schemas, so they’ll have a lot of work.
Maybe you don’t even have the events or information you need; you don’t have a process or tool that covers the thing you’re thinking about! It would be very expensive to get you the metric you asked for, because the scope is many layers deep.
As a more extreme example, if the change you want to make requires a change to the laws of physics, it’s pretty much a non-starter 😁
If your solution requires people to walk on the ceiling, you'll have a hard road ahead.
2️⃣ Design in Tomorrow’s Flexibility Today
Have you ever written code, or maybe an Excel file, where you hard-coded a value in multiple places? For example, you may have a retirement calculator, so you type out “0.08” in a few places, because that’s the historical average rate of return (8%).
Later, when you want to run different scenarios, now you’ve got to hunt for all those places where you typed in “0.08” and replace it with a new value. What a headache! And you’ll probably miss it somewhere!
Conversely, if at the start you realized you wanted to run scenarios, then you could have created a variable for rate of return, and set it at 0.08. Then, when you wanted to change it, you could change it in one place.
This is designing in tomorrow’s flexibility today. You think of what you might want, or need to, adjust in the future, and design that into the solution you’re building today.
This doesn’t happen today for a lot of reasons, but often it’s just because no one thought to contemplate how the solution would be used in the future. You need to travel to a time when your solution is live, and think about what changes might need to occur there.
You might have one product today, but will that always be true?
You might have two offices today, but will that always be true?
Design a solution that makes it easy to add that new product, new office - or new anything - later when it’s needed.
Doing this isn’t free of course; someone has to think about this and design in the flexibility. But for 5-10% extra cost today, you can save your business 5-10x the cost later, and that’s often a worthwhile tradeoff.
This person doesn't realize yet that their entire company will run on this spreadsheet next year.
3️⃣ Don’t Argue About the Top Until You’ve Solved the Bottom
Yes, all the layers need to be designed holistically, but don’t worry too much about the top layers when you start executing because they’re easy to change later.
For example, if you’re in a room arguing about whether a median or an average is the right summary metric, but you haven’t solved the deeper architectural layers, then stop arguing. Median or average is a simple change later and doesn’t need to be solved now.
This is important because sometimes you can’t change the deeper layer like you want. To think of a tangible example, let’s say that you’ve got a system that records a measurement every thirty minutes. If you’re looking to make “real-time” decisions and want to build a dashboard with live updates every second, that’s just not possible.
Make sure you know what’s going on deeper in the architecture before you waste time arguing options at the top layers - some of those decisions might be moot!
4️⃣ Top Layers Get Layered
A layer that sits at the top today can become three layers deep next year.
This is what progress looks like, and something we all do intentionally. You build something new, then build more things that take advantage of those new capabilities. This creates new layers on top of old layers.
This is only a problem when the layer you build today is more of a “quick and dirty” MVP instead of robust, well-architected, solution.
Have you ever created a proof of concept in a Google Sheet, only to have it become what your business runs on a year later?
You might think that you’ll have time later to make your MVP more robust, but that’s not always true. What often happens is that by creating the MVP you put out that fire, which makes another fire the higher priority. So you go put out that fire, and so on. You turn around later and the MVP was leveraged by another team, and now it’s too expensive to go back and fix it.
MVPs have their place, but be aware that they often get layered, and you won’t always go back to fix it like you imagined. It may not cost you much to make your MVP a little more robust today, and it’ll save you headache later.
Bringing it all Together
Leaders have to be good architects because speed matters in business. Using the architectural framework of layers can improve this skill.
Things in your business, and life, can be thought of in layers. Some layers move quickly (fashion, report colors), some layers move slowly (network architecture, laws of physics).
Top layers can be changed more quickly and less costly. Deeper layers can get real expensive to change.
Deeper layers are expensive to change because every layer on top depends on the layer below it. You can’t change a deep layer by itself because it’ll impact the layers above it.
Knowing this framework, there are four things you can do to improve strategy and transformation work:
1️⃣ Think Deeper - If you want to make a change to a top layer, think through what needs to be true in the bottom layers.
2️⃣ Design in Tomorrow’s Flexibility Today - Think of what you might want, or need to, adjust in the future, and design that into the solution you’re building today.
3️⃣ Don’t Argue About the Top Until You’ve Solved the Bottom - Make sure you know what’s going on deeper in the architecture before you waste time arguing options at the top layers - some of those decisions might be moot!
4️⃣ Top Layers Get Layered - A layer that sits at the top today can become three layers deep next year; make sure it’s ready for that.
Call to Action
Use your new “architectural layers” lenses to look at your business anew. Can you see the layers? Is it clear when you’re working, which layer you’re working on? Is that the right layer, or should you shift the conversation to another layer?
Do you see an architectural issue that you’ll have in the future if nothing changes? Are the solutions you’re designing robust to natural changes in your business? Make adjustments if needed.
You’re working on this because getting better at architecture improves the speed with which you can build new things or make changes to old things - something you’re doing all the time!
I’d love to hear your thoughts on this! Email me at heykev@kevinnoble.xyz and let me know what you think. Did this resonate? Anything you’d like to add?
Kevin 😀
Thanks for reading! If you loved it, please tell your friends and colleagues to subscribe here: https://kevinnoble.ck.page/
