Your directive is metabolized before it reaches the target. The dose you administer is not the dose that arrives.
Picture someone in the grip of angina, the crushing chest pain of a heart that isn't getting enough blood. A paramedic arrives with nitroglycerin, a drug that relaxes blood vessels and can stop the attack in minutes. Notice what the paramedic does not do: hand over a glass of water and a pill to swallow. Instead, a tiny tablet goes under the tongue. Ninety seconds later, relief.
Now rerun the scene with the identical drug, swallowed like a vitamin. Nothing happens. The pain doesn't budge. The pill might as well be a breath mint.
Same molecule. Same patient. Same dose on the label. The only thing that changed was the route, and that change is the difference between a rescue and a placebo. Swallowed nitroglycerin is one of the most famous duds in medicine; sublingual nitroglycerin has been saving lives since the 1870s. Hold that fact in your head, because it is the whole of what follows: the dose you administer is not the dose that arrives.
Pharmacologists have a precise, slightly brutal word for "how much of the dose actually shows up where it's needed": bioavailability, usually written F. It's the fraction of an administered dose that reaches systemic circulation intact. By definition, an intravenous injection is about 100% bioavailable; it goes straight into the blood, skipping every obstacle. Every other route is measured against that gold standard, and most of them lose badly.
The reason they lose has a name, too: first-pass metabolism, or the first-pass effect. When you swallow a drug, it doesn't go directly into your bloodstream. It's absorbed through the gut wall into the portal vein, and the portal vein's job is to route everything absorbed from your digestive tract through the liver first, before it reaches the rest of you. The liver is your chemical-defense organ. It's packed with CYP450 enzymes whose entire purpose is to find foreign molecules and dismantle them. Your medication doesn't get a diplomatic exemption. On its first pass through the liver, much of it is simply destroyed.
The numbers are sobering, and they're worth saying out loud because they make the abstraction concrete:
That last example is the whole essay in one molecule. The drug never changed. What changed was whether it had to survive the liver. Sublingual and transdermal and intravenous routes all share one trick: they reach the bloodstream without passing through the metabolizing organ first. The pill route doesn't, and the pill route pays for it.
So the pharmacologist who knows a drug is heavily metabolized faces a simple, unsentimental design choice. Raise the dose enough to shove a usable fraction past the liver, which is exactly why oral doses of first-pass drugs are deliberately much larger than their IV equivalents. Or change the route so the drug doesn't meet the liver on its way in. Either way, the discipline is the same: you do not dose for the pill. You dose for what arrives.
Here is the uncomfortable part. The exact pharmacokinetics that govern a swallowed pill also govern a strategy, an architectural principle, or a cultural value as it travels down an organization. The mapping isn't a loose metaphor; it's a near-isomorphism, mechanism for mechanism.
Leadership administers a dose: the OKR for the quarter, the new architectural standard, the value the all-hands was built around. That dose then enters the org's portal vein, the cascade, and is routed, by design, through every management layer before it reaches the people who actually do the work. And each layer is a little liver. Each VP, each director, each manager metabolizes the directive: absorbs the part that's convenient, reinterprets the part that's ambiguous, and quietly excretes the part that conflicts with their own team's incentives. Nobody is being malicious. The liver isn't malicious either; it's just doing what livers do to anything that passes through. A manager under a different set of pressures will, in perfect good faith, reshape the directive to fit the local metabolism. That is first-pass metabolism, in a meeting room.
And the bioavailability at the bottom is measurable, just like a plasma concentration. The closest thing we have to a clinical reading comes from Robert Kaplan and David Norton, the Balanced Scorecard pair, whose research, popularized in The Strategy-Focused Organization and their 2005 Harvard Business Review work, found that roughly 95% of a company's employees are unaware of, or do not understand, their company's strategy. Only about one in twenty grasps the goals they're supposed to be executing.
Sit with that figure for a second next to the pharmacology. Five percent. That is the bioavailability of a heavily-metabolized oral drug. Corporate strategy, cascaded down the org chart, arrives at the front line at roughly the bioavailability of swallowed morphine, actually worse. The strategy deck that felt crystal-clear in the boardroom shows up at the team doing the work as a faint, distorted trace, most of it digested somewhere upstream.
The supporting numbers fill in the mechanism. Surveys of strategy execution routinely find that 25-45% of employees can't even articulate their company's top priorities, a clarity gap that leaves them making locally reasonable, globally disconnected decisions all day. Practitioners describe the cause in language that could have been lifted from a pharmacology textbook: a top-level priority "travels through layers of management, each adding its own spin," until the frontline is left unclear; as decisions cascade downward, "each layer of the organization adds its own level of rationalization," and managers "reinterpret priorities to fit the needs of their own department." There's even a culprit organ. The strategy-execution literature calls it the "frozen middle": middle management, the layer that's supposed to carry the strategy forward and is simultaneously the layer that metabolizes it. The same organ that absorbs the dose is the one that destroys most of it.
And this matters: the loss compounds with length. A longer gut means more first-pass metabolism. A taller org chart means more layers between the dose and the target, and each one takes its cut. The more handoffs a directive must survive, the less of it survives. You can feel this in your own company: the message from a skip-level you talked to directly lands; the same message relayed through three intermediaries arrives as a rumor of a message.
The useful thing about a rigorous metaphor is that it doesn't just diagnose; it prescribes. Pharmacology already solved the low-bioavailability problem two ways, and both transfer cleanly.
Fix one: raise the dose. Because the oral route destroys most of each administration, you give more. This is the brute-force answer, and it's legitimate: say it more, say it more often, say it in more forms, knowing full well that most of every transmission will be metabolized before it reaches the team. The leaders who repeat a single strategic message until they're sick of hearing themselves, past the point that feels redundant from the top, aren't being unimaginative. They're dosing for first-pass loss. If only 5% of each telling survives, you have to tell it twenty times to deliver one clean dose. The all-hands, the wiki page, the Slack pin, the repeated reminder in standup: that's not noise, that's compensating for the metabolism.
But raising the dose is the inelegant fix, and pharmacology knows it. The elegant fix is the one that makes nitroglycerin a wonder drug instead of a dud.
Fix two: change the route of administration. Don't send the dose through the metabolizing organ at all.
The intravenous version, for an organization, is the leader speaking directly to the people doing the work. Skip-levels, not for theater but as pharmacokinetics: when the CTO explains the architectural bet to the engineers in the room, with no director in between to "translate" it, the directive arrives at something close to 100% bioavailability. Nothing got a chance to metabolize it. This is expensive and doesn't scale infinitely (IV access requires a vein and a clinician), which is exactly why you reserve it for the doses that matter most.
But there's a route that does scale, and it's the one tech leaders should care about most, because it's native to what we build. Call it the transdermal patch: encode the value in the system itself, where no manager's liver can reach it.
This is the move. An architectural principle delivered as a memo ("we value backward compatibility," "we don't merge without tests," "secrets never go in code") travels the oral route. It cascades through layers, each of which interprets "we value X" against its own deadline pressure, and arrives at the engineer as a vague preference she's free to trade away under a sprint crunch. Bioavailability: single digits.
The same principle encoded as a CI check that fails the build is administered transdermally. It bypasses every metabolizing layer. There is no director to reinterpret a red pipeline; there is no manager whose incentives can excrete a failing test. The rule reaches every engineer, every commit, at full strength, identically, forever, not because anyone cascaded it well, but because it was routed around the part of the body that destroys directives. A linter, a default configuration, a permission that can't be granted, a template that bakes in the right shape, a deploy gate that won't open: these are sublingual tablets. The principle goes straight into the bloodstream of the work.
This is why "encode the value in the tooling" beats "communicate the value better" so reliably, and now you can say why in one sentence: communication is the oral route, and the oral route is metabolized. Tooling is the route that bypasses metabolism. You are not choosing between a clear memo and a sloppy one. You are choosing between a drug that meets the liver and one that doesn't.
Which leaves the last discipline, and it's the one leaders skip most. A doctor titrating a drug does not measure success by the size of the pill she prescribed. She measures the plasma concentration at the target, the actual amount of drug actually circulating in the actual patient, and adjusts until that number is right. The dose on the label is an input she expects to be degraded; the blood level is the truth.
Most leaders measure the pill. They obsess over the wording of the memo, the polish of the strategy deck, the eloquence of the all-hands, the dose as administered, and then assume that because the dose was excellent, the effect must be too. Pharmacology says that's precisely backwards: a perfect drug down the wrong route arrives as nothing, and a beautifully-worded strategy cascaded through six layers arrives as Kaplan and Norton's 5%. The eloquence of the prescription tells you almost nothing about the concentration at the target.
So measure the blood. Don't ask "did we communicate the strategy?" Ask "can a randomly-chosen engineer, with no warning, tell me the top priority and why their current task serves it?" Don't ask "did we publish the architectural principle?" Ask "what fraction of last month's merged PRs actually honor it?" Those are plasma readings. They tell you the bioavailability you actually achieved, not the dose you hoped you delivered. And if the reading is low (it will be, the first time you take it honestly) you now know exactly which of two levers to pull: raise the dose, or change the route.
Here is the practical version you can use on Monday. For any value, strategy, or standard you care about, ask three questions in order. What's the bioavailability? Go find out what the front line actually understands and does, not what you said. Can I change the route? Is there a way to encode this in tooling, defaults, or direct contact so it bypasses the metabolizing layers entirely? If yes, do that; it's worth ten memos. If not, am I dosing for the metabolism? Am I repeating it enough, in enough forms, to push a usable fraction past the liver, or did I say it once in an all-hands and assume it landed?
The hardest thing to accept is the humbling premise underneath all of it. You do not get to administer a dose and assume it arrives. Your organization has a liver, and it is doing exactly what livers do. The leaders who get this stop tuning the pill, and start changing the route, or measuring the blood. Nitroglycerin swallowed is a placebo. The molecule was never the problem.
You told the agent what to do. Measure the blood, not the pill.
The essay's last discipline is the one that matters most for agents: the directive you administer is not the behavior that arrives. A prompt, a policy, a tool spec all travel the oral route, reinterpreted at every handoff between agents and layers, so the only honest reading is the plasma concentration at the target, what the agent actually did, not what you told it to do. Chain of Consciousness is that blood test: a tamper-evident record of an agent's real decisions and actions, so you can check the concentration at the target instead of trusting the dose on the label.
See Hosted Chain of Consciousness · See a verified action chain
pip install chain-of-consciousness · npm install chain-of-consciousness