The Chunking Model — How Founders Lose Millions Wandering the Wrong Level of Abstraction

Dylan was barely 28 when his financial advisory business passed seven figures. Team of nearly 80. Revenue machine. And he could not sleep.

I had known him since his national service days. He attended a two-day training program I ran, and he remembered me a decade later. When we reconnected, he was already successful by every external metric. The problem was not revenue. He was gifted at getting leads. What he could not do was keep his people together. Sleepless nights, worry upon worry about whether anyone would leave. An eighty-person organization held together by charisma and sheer willpower.

He thought the fix was becoming a better speaker. He had convinced himself charisma was leadership. I told him the external polish did not guarantee the internal substance. A charismatic speaker can be compelling on stage and still fail at the one thing that actually grows a company, which is bringing culture together so people want to reach toward the vision.

That sentence deserves a pause, because it names the exact failure most founders make. Dylan was operating at the wrong level of abstraction. He was brilliant at the specific, the concrete, the measurable. Leads, calls, revenue, headcount. Those are the levels you can touch. The thing he could not reach, the level directly above all of it, was the frame that gave the work meaning to his people: culture, vision, identity. He could not see it, so he could not influence it.

Neuro-Linguistic Programming has a name for this. It is the Chunking Model.

What the Chunking Model is

The Chunking Model comes out of the observation that the human mind does not store experience as a flat list. It stores it in a hierarchy of abstraction. At the bottom are the concrete specifics: the numbers, the words, the movements, the exact sequence of a task. Higher up sit the categories that those specifics belong to. Higher still sit the values and meanings that give the categories their weight. Higher again sits the identity that holds the whole thing together.

NLP calls moving up the abstraction hierarchy chunking up. Moving down it is chunking down. Moving sideways to a related category at the same level is chunking laterally.

Every one of those moves is a cognitive operation. And here is the uncomfortable truth for founders: most people get stuck at one level and never learn to move on command. The operator who cannot see the vision drowns in detail. The visionary who never touches the detail floats in fantasy. The leader who does both on cue, and knows which one the moment demands, is the one whose company actually scales.

The intellectual foundation for this is not mystical. It traces to Gregory Bateson's work on levels of abstraction and logical typing, which Richard Bandler and John Grinder drew on when they built the Meta Model in The Structure of Magic in 1975. Bateson noticed that a great deal of human confusion comes from treating a higher-level category as if it were a concrete thing, or a concrete thing as if it were a category. Bandler and Grinder turned that insight into a set of linguistic tools. Robert Dilts later formalized the chunking operations in his modeling work, most thoroughly in Modeling With NLP and in the Neurological Levels framework, where he shows that the same chunking logic runs from environment up through behavior, capability, belief, and identity.

The word chunk itself was not NLP's invention. It entered cognitive psychology through George Miller, who in 1956 published The Magical Number Seven, Plus or Minus Two, showing that working memory holds roughly seven chunks of information at once. NLP's contribution was to notice that a chunk is not fixed. You decide what counts as one chunk by where you draw the boundary. A founder who reads a monthly profit statement as one chunk sees a different reality than the founder who reads every line item as a separate chunk. Same document. Different minds. The technique is the act of redrawing the boundary on purpose.

Let me be direct about what was happening with Dylan, because it is the pattern I see in client after client.

Dylan was trapped at the level of behaviors and specific events. He could tell you exactly how to get a lead. He could tell you the script, the call, the close. That is chunking down, and he had done it to a fine art. What he could not do was chunk up. When I asked him what his people were actually reaching toward, he had no answer that moved anyone. He had a list of targets. He did not have a rallying point.

That gap is not a personality flaw. It is a competence gap in a specific cognitive skill, and the good news is that the skill is trainable. You can learn to move between levels the way you learn any other deliberate movement. The bad news is that almost no one trains it, because business school rewards the ability to go deep on one detail, not the ability to relocate your own attention to a higher frame and come back down.

Here is what the escape from that trap looks like mechanically.

Chunk up. You ask a question that moves attention to a higher level of abstraction. What is this an example of? What is this in service of? What does achieving this give me? The purpose of chunking up is to reconnect an action to its meaning. When a founder cannot motivate a team, it is almost always because the team is looking at the same action without the meaning attached. Chunking up supplies the meaning.

Chunk down. You ask a question that moves attention to a more specific, concrete level. What does that look like exactly? What specifically do you do first? Give me an example. The purpose of chunking down is to convert a fuzzy ambition into an executable next step. When a founder has a vision but the team never starts, it is because the vision was never broken into a behavioral first action. Chunking down supplies the action.

Chunk lateral. You ask a question that moves sideways to a related category. What else is like this? What is the opposite? How is this different from that? The purpose of chunking lateral is to break a fixed frame by showing that the situation belongs to a larger set of possibilities. When a founder is stuck on one interpretation, chunking lateral introduces an alternative that was always available but invisible.

Every one of these is a question. That is the whole mechanism. You are not adding information. You are relocating the attention of the person you are talking to, or your own attention, to a different level of the hierarchy and letting the new frame reorganize what the details mean.

The Chunking Model is filed under Influence and Communication because it is the engine of both persuasion and clarity.

Consider the most common argument between two smart people. One says the company is failing because the numbers are down. The other says the company is failing because the values are weak. They argue for an hour. Here is the thing: they are both right, and they will never agree, because one is chunked down at the metric and the other is chunked up at the meaning. The argument is not a disagreement about facts. It is a mismatch of abstraction levels. The fix is not to pick a winner. The fix is to chunk both people to the same level and let them see that they are describing the same organization from two altitudes.

Once you can see this happening, you cannot unsee it. You watch a board meeting and you realize the reason nothing resolves is that no two people in the room are arguing at the same altitude. One is in the data. One is in the strategy. One is in the identity. They talk past each other not because they are rude but because they are on different floors of the same building.

The same mechanism runs in the other direction. If you want to persuade someone, you choose the level that moves them. Data-driven people respond to chunking down: the specific numbers, the exact case, the concrete proof. Vision-driven people respond to chunking up: the mission, the meaning, the identity of who they become. A leader who can detect a person's preferred level and meet them there is the leader who gets buy-in. This is not manipulation. It is the difference between speaking a language the other person can hear and speaking one they cannot.

Let me make this concrete with a decision every scaling founder faces: whether to hire a certain senior person.

Level one, the behavior. What exactly does this person do that we need? What is the specific skill, and what is the measurable output in the first ninety days?

Level two, the capability and belief. What does this person believe about how the work gets done, and does that belief match how we actually win? A brilliant operator who believes the company wins through individual heroics will wreck a team that wins through systems.

Level three, the identity and vision. Who does this person become inside the organization, and does that identity strengthen the culture we are trying to build? This is the level most hiring decisions skip, and it is the level where Dylan was blind. He could assess a producer's numbers. He could not assess whether that person made the culture stronger or weaker, because he was not operating on the culture level at all.

The Chunking Model gives you a protocol: run the decision through all three levels, in order, and only commit when the candidate passes at every altitude. Most founders commit at level one and pay for it at level three.