Skip to content

Thinking At Work, Part 1: Systems

14 min

This week, I'm sharing some of what I've been working on in my role as CEO.

Recently we started the hiring process at my agency

(by the way, if you know anything about PPC and are interested in a fun and exciting job, email me and let me know! I'd love to consider you)

...and it occurred to me how poor a job I've done at communicating our culture.

That may seem like "business-speak", but hear me out:

Culture at a workplace isn't something that gets handed down from some corporate office in a memo. It describes the "tone" of the interactions between the people that work there; between the company and it's clients; between the world-at-large and the world-within.

It encompasses everything from why we do what we do to how we do it.

My single biggest challenge as an agency owner? Teaching my team to innovate without me.

How do I retire, AND have the business evolve into something I couldn't expect? Something that both aligns with my personal values AND represents the best of everyone that works there?

To do that, I need my team to not just repeat best practices, or do what I did because it's what I did. I need them to be building and developing new ways of doing things.

That's what this stuff is about: not just teaching a way of working, but a way of thinking.

Hopefully, that context makes the following essay more comprehensible. And while I talk a lot about ads and ad management in this essay, nothing in here is specific to what I do. You don't need to know anything about what I do to find something useful here.

Let me know your thoughts! If you manage to read all of this you are officially a fellow Systems Nerd and I will be happy to nerd out in the questions.

Not Pieces, But Systems

As a PPC manager, you are not working with things, but systems.

If you've ever had a client react negatively to seemingly good news...

Or had a tried-and-true campaign go suddenly and inexplicably off the rails...

Or had an unexpected "algorithm update" destroy your long-running "best practice"...

Or bought a course promising a "tried and true" strategy with pages and pages of testimonials, only to have it go absolutely nowhere... was because of systems.

Systems are everywhere.

Systems thinking - the ability to spot, understand, and manipulate systems - should be the single most sought-after skill of all marketing agencies.

But it isn't.

In fact, you'd be hard-pressed to hear anyone in marketing talk about systems thinking at all.

Why is this?

I don't know for sure, but I have a theory:

Online marketers are too clever by half.

By which I mean:

We all think we get it.

We think that we understand the world around us.

We think the world is predictable, provided you have expert knowledge (which, of course, we all have).

And much of the time, we're RIGHT!

We do get it. We are able to predict some outcomes, and we're rewarded for it - with more leads, more clients, more revenue.

We know exactly what's going on.

....Until we don't.

The rules change.

The market shifts.

The channel evolves.

If you are in this business long enough, this will happen to you. It is only a matter of time (and not very much time at that).

The best PPC managers in the world aren't slavishly following a step-by-step guide. They're not expecting everything to work every time, and they don't fool themselves into thinking they can predict the future.

The best PPC managers in the world are flexible, innovative, and always thinking on their feet.

In other words:

They are systems thinkers.

And the best thing about systems thinking?

It can be taught.

You don't need to go back to school or get certified or invest in an expensive course.

You just need to make a slight shift in how you see the world.

And once you do?

You will never look at things the same way again.

What Are Systems?

Donella Meadows describes systems this way:

"A system is a set of things—people, cells, molecules, or whatever—interconnected in such a way that they produce their own pattern of behavior over time.

The system may be buffeted, constricted, triggered, or driven by outside forces. But the system’s response to these forces is characteristic of itself, and that response is seldom simple in the real world."

Take an ant colony.

Ants are simple creatures. They possess nothing we would recognize as "intelligence;" their set of operating instructions and programmed responses is extremely minimal.

There is no one ant that knows how to build a colony, or even what a colony is.

Yet - ant colonies are capable of incredible things. They build and use of tools, they engage in organized warfare, they locate and transport resources en masse, etc.

The ant is the unit; the colony is the system.

You can't cut open any individual ant and find the colony.

You can't look at an ant and predict the colony.

The colony emerges from the interactions between the individuals.

This leads us to a profound truth:

The behavior of a system cannot be known by examining its pieces.

Let's connect this to marketing:

Have you ever launched a campaign that seemed, on it's surface, to be perfect?

The right ads, the right message, the right landing page...

And yet, when you finally hit "enable" and launched the campaign, it went absolutely nowhere?

As it turns out, perfect pieces do not necessarily make a perfect whole.

The behavior of a system cannot be known by examining its pieces.

Too often, as managers, we spend our time optimizing pieces, while ignoring the performance of the system. This is a critical error that underlies much of the errors made during management (and one we will return to in-depth when we discuss local optima).

For now, try to internalize these concepts:

How To Change a System

If examining the pieces of a system doesn't tell us about its behavior, what does?

A system's behavior comes from three interlocking parts:

You can get a sense how these work by imagining what would happen if you changed each piece in turn.

Let's imagine our ant colony from before.

How much would the system change if you swapped out the elements (in this case, the individual ants) one by one?

Not much. Swapping one ant for another wouldn't change the behavior of the colony at all.

How much would the system change if you swapped out the interconnections (the relationships between the ants) one by one?

A lot! If we changed the rules by which ants interact, the behavior of the ant colony would be very different - perhaps unrecognizable.

For example:

Ants are able to locate food by means of pheromone trails. Each individual ant searches its surroundings randomly until it finds food; it then transports some of that food back to the colony, laying down a pheromone trail behind it.

Other ants, following that trail, find the food as well, also laying down their own pheromone trails. Each ant that follows the trail makes the signal stronger by laying their own pheromones on top; very quickly, you'll have a solid column of ants headed straight to the source of food and back.

As the food source is depleted, some ants are unable to access the food, and thus return without laying down a pheromone trail. As the number of frustrated ants rises, the strength of the pheromone trail declines, and the column of ants vanishes.

Imagine that we changed even one of the rules; for example, imagine that we could tweak it so that frustrated ants, instead of returning to the nest, fought for access to the food resource.

How would this change the behavior of the colony?

I'm not exactly sure, but I can say it would dramatically disrupt the colony's existing behaviors.

How much would the system change if you swapped out the purpose of the system?

A lot! Even if every ant in the colony stayed the same, and all the rules of their interactions stayed the same, the colony would behave very different if it's goal switched from self-preservation to, say, dominating the earth.

(By the way - they kind of already do. Among all animals, ants outweigh us in biomass — putting billions and billions of tons up against humans' fewer than 500 million.)

The point I am making here is this:

Focusing on the individual elements in a system is the least impactful way of trying to change the system's behavior. Instead, the fastest way to change is by examining the interconnections and purpose of the system.

This is why examining the pieces doesn't predict a system's behavior.

Now that you know how to change a system's behavior - examine it's connections and it's purpose, and only then turn to the elements - you may be wondering how you can apply this directly to your PPC management workflow.

But there are still a few more critical concepts you need to understand before you can apply this knowledge to the real world.

Do you remember this passage?

"How would this change the behavior of the colony?

I'm not exactly sure, but I can say it would dramatically disrupt the colony's existing behaviors."

If we know the elements, and the interconnections, and the purpose...and I know what I'm changing...

How come I can't predict the subsequent change in behavior?

The answer is both simple and profound:


Three Body Problems

Humans like to think linearly.

We want things to proceed in an orderly fashion, with one thing leading logically to the other.

But this doesn't describe the universe as it is. It describes the universe as we wish it to be.

Understanding this will change the way you view the world.

Let's illustrate:

Imagine two objects, floating in space.

Imagine that we know everything we want to know about these objects - their exact size, mass, gravitational pull, whatever.

Imagine we give some incredibly smart mathematicians and physicists a starting position for these two bodies - an exact snapshot of where they are right now.

Using everything we know about these objects, can we set them in motion and calculate where they will be a minute from now?

Yes, we can!

This is the world of Newtonian physics, the laws of motion that we all learned in school (or, in my case, pretended to learn in school, but really just copied from the kid sitting next to me).

Let's change the situation up a bit:

The scenario above imagines two objects.

How many objects do we need to have before it is mathematically impossible to do this type of calculation?

How many objects need to get involved before the most powerful computers in the universe are unable to figure out their future positions?

The answer:


In physics, this known as the "three body problem."

Two objects in space are perfectly predictable using Newtonian physics.

Three objects in space?



The moment we add a third body into our system (which, of course, is what elements interacting in space are), the system becomes chaotic.

As Ben Hunt describes it:

"In 1887, Henri Poincaré proved that the motion of the three objects, with the exception of a few special starting cases, is non-repeating. This is a chaotic system, meaning that the historical pattern of object positions has ZERO predictive power in figuring out where these objects will be in the future.

There is no algorithm that a human can possibly discover to solve this problem. It does not exist."

The Three Body Problem illustrates just how quickly complexity emerges from systems behavior.

Normal ant colonies typically house anywhere between 1,000 to 100,000 ants while supercolonies can contain millions of ants that live together.

How much more chaotic, then, is our ant colony than our three body problem?

Infinite barely begins to describe it.

When you understand how quickly and how irrevocably complexity emerges from the interactions of even simple elements, you start to understand why the world around us - no matter how straight-forward things may seem on the surface....

Is utterly, completely unpredictable.

By the way, Google claims that 4 million people advertise using Google Ads. Can you see where we're headed?


While we're trying to wrap our heads around the sheer complexity of the universe around us, I have a confession:

I've actually been simplifying things. A lot.

Let's go back to our ant colony for a moment.

Despite what I said before, our ant colony's behavior isn't dictated solely by the ants, their interconnections, and their goals.

After all, ant colonies don't exist in a vacuum.

(But there are ants in space! NASA sent them up there, they're called "ant-stronauts!" )

Ant colonies exist in a wider world of competing insects, animals, vegetation, weather patterns and cosmic events.

In short:

System behavior emerges not just from the interaction of parts, but from the interaction of systems with one another.

Which begs the question:

If systems are defined not just by their elements, but by their interconnections...

And if different systems interact with one another...

How do you define where one system ends, and another begins?

Drawing Lines

No system operates by itself.

Instead, systems interact with one another, increasing the complexity of their behaviors exponentially.

Our ant colony behaves in a certain, semi-predictable way...

...Until a neighboring ant colony attacks...

Or the tree they were living under collapses...

Or the area is deforested in order to make way for a housing complex...

Or an asteroid the size of Texas smashes into the Earth, spewing volcanic ash into the sky and blocking out the sun.

Even the idea of "systems" as I've described them is nothing more than a convenient fiction. If systems are defined by their interconnections, it becomes impossible to really say where one system ends and another begins. Instead, we face an infinity of interlocking pieces that shift, change, and adapt more quickly than we can catalog them.

This doesn't mean that trying to understand the world around is futile.

Drawing a boundary line around a given set of interactions and calling it "a system" for the purposes of study and understanding has real utility.

But we need to understand that this act of drawing boundary lines is just that: a conscious act. Our boundaries are not really "out there," in the world. They are mental constructs adopted for our own convenience.

As Donella Meadows writes:

"There are no separate systems. The world is a continuum. Where to draw a boundary around a system depends on the purpose of the discussion—the questions we want to ask...The right boundary for thinking about a problem rarely coincides with the boundary of an academic discipline, or with a political boundary."

We must be willing to expand our awareness beyond what we are used to, beyond the limits of our existing knowledge or expertise.

We also need to be open to the possibility that each new problem or situation may require a shifting or, indeed, a complete reordering of the boundaries and models that have served us well in the past.

That's why we're starting our discussion of PPC Management with ants, physics, and complexity theory: it's an illusion to assume that the answers to your specific problems lie within the boundaries of some conveniently delineated field of knowledge.

The best online marketers in the world don't just study online marketing.

They remain open to psychology, history, sociology, science, literature, current events, computer science, and poetry.

Does this sound a bit too lofty for the topic at hand?

Are you thinking:

"Look, Dan - I get it. The world's a big, complex place. But I'm just managing ads, not trying to save the free world."

Let me give you a real-world example, from my own experience as an agency owner.

The client in question was a real estate investor (someone who buys homes to fix up and re-sell).

Their target market was Houston, a thriving real estate market in the midst of significant expansion. They'd seen big successes over the course of their first year with our agency, with higher-than-national average Conversion Rates and close rates. They were extremely happy with the service, and in August signed a 12-month contract with us.

Exactly three months later the client's ROI had plummeted. Lead flow slowed to almost nothing, and they asked to be let out of their contract early (which we did).

When clients leave our agency I typically do a "post-mortem" - a walkthrough of all their communication, account changes, etc, to determine if there's anything we can learn from that experience.

Put yourself in my shoes, about to do a post-mortem on this client.

Where should we look to find the cause of the client's problems?

How you answer that question depends on how you define the system you're working on.

If you limit the system in question to Google Ads, you might look for answers in their search terms report (were they suddenly generating more off-topic queries?), or their average CPC (did competitors enter the marketing, inflating click costs and driving our client out of the best ad positions?).

If you expand the system in question to include our agency, you might look for answers in the team's communications with the client (were they polite and responsive?), or the account's change history (were updates being made regularly? Was the team doing the right things?), or our task management system (were tasks clearly delegated? Were they done on time?)

If you expand the system further to include the client's business, you might ask whether they had recently brought on a new acquisitions person (lowering their average close rate), or whether they'd moved to a new CRM (is there a problem with their new follow up system?).

If you expand the system in question even further, and you might notice that on August 28th of that year, Tropical Storm Harvey dumped record amounts of rain on Houston, flooding most of the city.

Herein lies the rub:

Each of these system definitions was valid.

And, because people tend to find problems when they look for them, it's totally possible that we would have come away from each system investigation convinced we had spotted the root cause of the client's trouble!

This is how agency owners and PPC managers routinely address the wrong problems...while feeling that they're addressing the right ones.

The process of expanding or retracting your concept of a system as needed is an art. Indeed, in Donella Meadow's metaphor, it is a dance.

We can end our discussion of systems thinking by quoting her:

"People who are raised in the industrial world and who get enthused about systems thinking are likely to make a terrible mistake. They are likely to assume that here, in systems analysis, in interconnection and complication, in the power of the computer, here at last, is the key to prediction and control. This mistake is likely because the mindset of the industrial world assumes that there is a key to prediction and control."

"Systems thinking leads to another conclusion–however, waiting, shining, obvious as soon as we stop being blinded by the illusion of control. It says that there is plenty to do, of a different sort of “doing.” The future can’t be predicted, but it can be envisioned and brought lovingly into being.

Systems can’t be controlled, but they can be designed and redesigned. We can’t surge forward with certainty into a world of no surprises, but we can expect surprises and learn from them and even profit from them. We can’t impose our will upon a system. We can listen to what the system tells us, and discover how its properties and our values can work together to bring forth something much better than could ever be produced by our will alone.

We can’t control systems or figure them out. But we can dance with them!"





The Man Behind Ethereum Is Worried About Crypto’s Future
Vitalik Buterin dreamed up the blockchain technology for uses beyond currency. Now he’s fighting to fix the industry

Fascinating profile of Vitalik, and an interesting look into some of his ethics. Good read, even if you know next to nothing about crypto.

Subscribe to receive the latest posts in your inbox.