There is a reason arthropods cannot get large. An exoskeleton is a superb piece of engineering right up until the animal inside it needs to grow, at which point it becomes the single tightest constraint on the organism's future. The shell does not grow with the animal. It cannot. It was not built by the animal's own developmental process - it was secreted once, into a fixed shape, and now the only way to become something bigger than the shell allows is to shed the shell entirely and stand there, soft and exposed, until a new one hardens around whatever you've managed to become in the meantime. This is not a design flaw the lobster forgot to fix. It is the price of a body plan organized around an external, pre-specified form rather than an internal one.
Contrast this with what an embryo does, or a river delta, or a snowflake, or - the developmental biologist Conrad Waddington's image, and it's the right one - a marble finding its way down a landscape of ridges and valleys that were shaped by the terrain's own history, arriving at one of several possible stable outcomes not because anything forced it there but because that is where the landscape, given this marble, at this speed, from this starting point, was always going to send it.1 Nobody pours a snowflake into a snowflake-shaped mold. The form is not imposed on the growing thing from outside. It is found by the growing thing, from inside, as a consequence of the rules already present in its own material and its own history. Endoskeletons, for what it's worth, get this partly right too - bone remodels continuously in response to the loads actually placed on it, which is why a tennis player's racket arm has measurably denser bone than the other one. The skeleton is listening to the organism's actual life and updating accordingly. An exoskeleton cannot listen. It can only be outgrown, or shed.
So is "log every touchpoint," and "stage-gate the pipeline," and the twenty-four-box qualification matrix laminated to the wall of every regional sales office in America. These are not neutral containers waiting patiently to hold whatever shape a deal happens to take. They are fixed forms, specified in advance, applied uniformly, built by people who were not inside the deal and could not have been, because the deal did not exist yet when the form was designed. And a real deal - a live negotiation between two organizations each of which is itself a tangle of internal politics, incentives, fears, and half-formed intentions - is precisely the kind of growing thing that does not arrive pre-shaped. It has a landscape, not a mold. It has ridges and valleys carved by who trusts whom, what happened in a similar deal three years ago, which stakeholder is quietly protecting their own reputation, what the buyer's fiscal calendar looks like, what the rep noticed in a pause before someone answered a question. None of that was specified by MEDDIC's designers, because none of it could have been. It only exists once the deal exists.
This is why forcing a deal into MEDDIC's shape does not describe the deal. It amputates whatever part of the deal doesn't fit the six boxes, and calls what's left "documented." The rep who is good at the sale is, whether she has language for it or not, doing something closer to what the marble does on Waddington's landscape - finding the actual shape available to this deal, given its history and its terrain - and then, afterward, if she is a diligent citizen of the CRM, translating that discovered shape into the pre-cut vocabulary of Metrics and Economic Buyer, the way you might describe a river in the language of a filing cabinet. The translation is not wrong, exactly. It's just not the thing. It's the shed exoskeleton, admired by an auditor who has never seen the animal that grew past it.
David Krakauer's distinction gives this the biological register it deserves. Krakauer - an evolutionary biologist and president of the Santa Fe Institute - has argued, with his co-authors John Krakauer and Melanie Mitchell, that real coarse-graining - the kind that shows up when physicists replace ten to the twenty-third particle trajectories with two numbers, temperature and pressure, and lose nothing that matters for prediction - is discovered by minimizing error against the actual system.2 It earns its compression by continuing to work. Krakauer's shorthand for this is worth borrowing directly: emergence is "more is different"; intelligence, built on top of emergence, is "less is more" - doing more with less by finding the representation that was actually there to be found, not the one that was easiest to impose.3 Nobody watched a thousand deals, asked which handful of variables actually carried the predictive weight, and arrived at Metrics-Economic-Buyer-Decision-Criteria-Decision-Process-Identify-Pain-Champion. It was assembled, instead, from what a sales manager could plausibly ask a rep to report on a Friday. It has the shape of a compressed representation - six words instead of ten thousand data points - without doing compression's job, which is to preserve the signal that lets you act. It is compression's costume. A library entry, in Krakauer's phrase, not intelligence.
And this is where the physicist Philip Anderson's "more is different" stops being a slogan borrowed from physics and starts doing actual work in this argument. Anderson - who won the Nobel Prize in Physics in 1977, three years after this argument, for a different but related body of work in condensed matter - made the claim in a short, now-canonical 1972 paper in Science, and it was never simply that complex systems surprise us.4 It was sharper than that: at each level of organization, genuinely new law appears, and no exhaustive inventory of the level below will hand it to you for free. You cannot get "wet" by cataloging water molecules with arbitrary thoroughness. Wetness is not a fact hiding inside the molecules, waiting for enough of them to be listed. It is a property of the aggregate, under certain conditions, that the molecule-level description was never going to contain.
A perfectly logged sale - every email timestamp, every call duration, every slide-view, every CRM stage transition recorded without error - is a perfect inventory of the molecules. It is not, and cannot become, an account of the wetness: the moment, somewhere inside all that activity, when a buyer stopped evaluating your solution as one option among several and started treating it as the answer to a problem they had only just let themselves fully see. That reclassification is the phase change. It is the thing that actually determined whether the deal closed. And it does not have a field in the CRM, not because someone forgot to build one, but because it isn't the kind of thing that exists at the CRM's resolution. You could add more fields forever - more granularity, more mandatory dropdowns, more "reason for stage change" text boxes - and you would never arrive at it, for the same reason more molecules never arrive at wet.
None of this is an argument that structure is worthless, or that the rep who happens to remember which stage a deal is in has learned nothing real. It is an argument that there are two different kinds of structure, and enterprise sales has been quietly mistaking one for the other for several decades. One kind is imposed from outside, fixed in advance, indifferent to the deal's own history - the exoskeleton, secreted once and worn regardless of fit. The other is discovered from inside, shaped by the deal's actual terrain, updated as the deal's own history accumulates - Waddington's landscape, or bone remodeling under real load. The first kind can be audited. The second kind is the only one that was ever going to close anything. Mistaking the first for the second does not merely fail to help. It cannot help, for the same reason a shell cannot grow: it was never built by, or for, the thing that's actually alive inside it.