Why Vendors Say “Keep Paying” and “Act Now” — And How to Tell Which One Is True

Table of contents

• Without a Model, Both Claims Are Just Assertions
• With a Model, Both Claims Become Checkable
• Why Urgency Is Structural, Not Manufactured
• What a Vendor Without a Model Can Still Offer
• The Model’s Honest Limits
• What a Model Actually Buys a Buyer

Every vendor in this category eventually says some version of the same two sentences: you need to keep paying after the initial work is done, and you should act now rather than later. Both claims are true, in the specific sense this series cares about. The question worth sitting with is what makes them true, because the answer changes entirely depending on whether a governing model exists underneath the sentence or not.

Without a Model, Both Claims Are Just Assertions

“Act now” fares no better. Urgency asserted without a mechanism is indistinguishable from urgency manufactured to close a deal faster. A buyer has no principled way to tell which one they’re looking at.

With a Model, Both Claims Become Checkable

Take the retainer question first. Byrum’s Dominance Inequality treats an entity’s citation strength as a rate — a quantity that rises or falls over time depending on whether the rate of new signal construction exceeds the rate at which unreinforced signals decay. That decay rate isn’t zero for any architecture the model has examined.

The published estimate places the required ongoing signal-construction rate at roughly 5 to 25 percent above a one-time build, depending on the specific decay coefficient measured for a given category. That’s not a sales argument for a retainer. It’s a stated consequence of treating citation strength as something that decays without reinforcement, the same way a stated consequence of treating a battery as something that self-discharges is that it eventually needs recharging regardless of how well it was built. The number can be checked, argued with, recalculated with different assumptions — everything a sales assertion cannot be.

Why Urgency Is Structural, Not Manufactured

If that’s true, delay isn’t merely costly in the ordinary sense that waiting to invest is usually costly. It’s costly at a compounding rate that a company entering a year later can’t simply buy back by spending more once it decides to start. That is a specific, structural claim about why delay behaves differently here than in most business decisions — not a countdown clock bolted onto a sales page.

What a Vendor Without a Model Can Still Offer

What it cannot offer is a checkable account of why the retainer number is what it is, or why the urgency is structural rather than manufactured. Because without a model, those questions don’t have derivable answers, only experienced ones.

The Model’s Honest Limits

Here is the honesty this series owes its own claim, stated plainly rather than buried: Byrum’s Dominance Inequality is a phenomenological model — it describes a regularity in observable citation behavior without claiming to know the internal mechanism that produces it, the same way early thermodynamics described how heat behaves before anyone understood why at the level of moving particles. Its central numerical estimates come from calibration against related literature, not yet from the model’s own dedicated validation studies, which are designed but not yet executed.

Internal consistency has been demonstrated — five thousand simulated trajectories behaved the way the model predicted they would, at the parameter ranges tested. That is not the same claim as empirical confirmation against real companies over real time, and this series will not blur the two. If a reader takes nothing else from this article, it should be that a vendor willing to state exactly what would prove its own model wrong is doing something almost no vendor in this category does — and that the model publishing its own falsification criteria is, on its own, a form of evidence worth weighing, independent of whether the model turns out to be right.

What a Model Actually Buys a Buyer

A retainer justified by assertion and a retainer justified by a decay rate can produce the identical invoice. Only one of them can be checked, argued with, or proven wrong. The same is true of urgency — manufactured urgency and structural urgency read identically in a sales conversation and behave completely differently under scrutiny.

What a governing model actually buys a buyer isn’t certainty; the model itself is unproven at the empirical level. What it buys is the ability to ask a falsifiable question instead of having to simply trust an answer.

For how this model’s core measurement — citation probability — is actually calculated, see Your AI Visibility Score: What It Is, How It’s Measured. The next article in this series removes the delivery system instead, and shows what happens to this same model when nothing builds the values it depends on.