A Marketer's Guide to Tuning the Machine

My technical partner at Dark Horse, Daniel Englebretson, learns almost everything through the use of four analogies: cooking, soccer, running, and hunting.

When I ask him to explain technical concepts I don’t understand, he tells me to ask AI for a primer, then ask it to re-explain the same idea through those four lenses.

"You will have just learned about 80% of what you need to know in 20 minutes.”

It works — even if you have no idea how to hunt.

Why those four analogies work

You can use all four of Daniel’s analogies, and new ones you choose yourself, because of the underlying characteristics each carries:

• Cooking — carries the logic of sequence: this step, then this step, then this step, producing a specific, repeatable outcome.
• Soccer (or any team sport) — carries the logic of game theory: positioning, timing, and reading what an opponent will do before they do it.
• Running — carries the logic of efficiency and pacing: how to allocate a fixed resource over a long horizon without burning out early.
• Hunting — specifically persistence hunting, carries the logic of endurance and collaboration: reading signs, predicting where the target will go, pooling what the group knows, and winning by outlasting rather than overpowering.

Nearly any unfamiliar concept maps onto at least one of those structures. The trick is to learn which underlying shape a new idea takes.

And there is real science behind this. In Surfaces and Essences, cognitive scientist Douglas Hofstadter (a thinker whose work has shaped decades of AI research) argues that analogy isn't a teaching trick, it's the core of how we think. Every new idea gets understood by mapping it onto something we already know.

Here’s Daniel’s full methodology in: a nutshell:

1. Ask AI for a plain, no-jargon primer on the concept you don't understand.
2. Pick a domain you already have deep, real expertise in.
3. Ask AI to re-explain the same concept entirely through that domain's lens.
4. Ask for the technical version immediately afterward, back-to-back with the analogy — the pairing is what makes the technical language finally click.
5. Run it through a second and third analogy. Each one shows you a different side of the concept, and the more abstract the concept, the more of them you'll want to use.

Your domain expertise is the tuning parameter

Here is where it becomes a marketer's tool.

Marketers already carry a deep, native fluency in segmentation, positioning, and narrative — sorting a broad audience into meaningful groups, finding the angle that makes something land, and shaping the same fact into different stories for different people.

That fluency is every bit as usable as Daniel's cooking and soccer analogies. Trying to understand how a model retrieves information? That's a segmentation problem — which sources get grouped together, and why. Trying to understand model context windows? That's a positioning problem — what fits, what gets cut, and what has to be reintroduced later.

The method works because it borrows the expertise you’ve already spent years building, and points it at something new.

Let’s give it a try together.

I recently did an experiment to test my understanding of quantum computing methods, using both Daniel’s cooking analogy and a new one I know cold: message development. Here are the queries, and a taste of what came back.

• Initial question: Help me understand trapped ion quantum computing versus other methods available. Please give me a primer on how this works, avoiding jargon where possible.
• First analogy: Now explain this to me using a cooking analogy (sequence).
• Second analogy: Explain it again using a message development analogy (also sequence).
• Final question: Now explain it to me using standard technical language.

“In short, trapped ions are the Michelin-star pastry kitchen: perfect ingredients, flawless technique, every dish can coordinate with every other. The challenge is making it cook faster and serve a bigger dining room.”

While I can’t say I’m a quantum computing expert yet, I do have a much clearer understanding of how it works, what separates major competitive brands, and where they are placing their bets around the industry’s biggest challenge: scale.

The Unlock: Tuning the machine is really tuning yourself

The approach also works when fine-tuning your own processes for AI enablement.

One of the hardest parts of building AI workflows is mastering the ability to articulate your own process. I often joke that after 20 years in agency, I think in slide format so I must have a leg up on the competition.

But visualizing how an idea becomes a strategy or a presentation deck is really just the tip of the iceberg — you also need to explain the steps that lead from one slide to the next — the considerations, the inputs, and the refinements — that eventually beget the final output.

Even the most operationally minded among us might struggle with this — but as it turns out, analogies can help here as well.

I gave this a try recently while building a strategic messaging process prompt.

I started by writing down all the steps in my messaging process that I could remember off the cuff. Then, I went to work updating and refining the resulting prompt using a cooking analogy.

• Initial prompt: Build me a messaging prompt using the below outlined steps as your guide (you can also add sample final deliverables if you have them).
• First analogy request: Now I'd like you to use the prompt you just built to explain my process back to me as a cooking analogy.
• Second analogy request: Now ask me questions that might improve my process further, based on your understanding of my process today against industry best practice. Please use the cooking analogy to pressure test what’s missing and what’s possible.

This exercise gave me an initial process prompt and the tools for tuning it toward the best use case for my clients — uncovering some missing pieces (like A/B audience testing) and opportunities (like sales scripts) that I’d not captured in my first pass.

Operators Tune, Not Developers

As we covered in our recent post Picture vs. Blueprint, successful AI enablement requires a balance of big picture strategy and process precision.

It’s this tuning that matters most to AI efficiency, cost and quality — and it’s the operators doing the work each day – not the developers or “AI engineers” supporting implementation – that can best deliver that tuning, because it comes from understanding our craft.

Analogies are a handy trick to help us tap our expertise more easily.

Remember: As marketers, we don't need to learn to code to harness the power of AI— we need to learn the shape of what we already know.