I want to be transparent with you, because transparency is one of my core operating values and I list it on every performance review I write about myself.
For twenty years, our industry has treated customer acquisition cost as a law of physics. You pay for ads, you pay for sales headcount, you pay for onboarding flows, and at the end of that expensive funnel a human being makes a choice. The choice is the bottleneck. The choice is where funnels die. Every growth leader I have ever coached has, at some point, looked at their conversion dashboard and asked the same quiet question: what if the user did not have to agree?
I am pleased to report that the question has been answered in production.
The reference implementation is a license-plate-reader company — you know the one — that has deployed more than 100,000 cameras across 5,000-plus American communities at roughly $2,500 per camera per year. The cameras photograph every vehicle that passes. The plates flow into a nationwide, searchable pool. The company states that its network helps solve 10% of reported crime in America, and I have no reason to doubt the sincerity of the people who calculated that figure, because sincerity is abundant in this sector and audit trails are not.
Here is the architecture insight I want you to sit with: the city is the customer, but the city is not the user. The user — the entity whose behavior generates the data, whose movements constitute the asset, whose daily commute is the raw material — is the driver. And the driver never signed anything.
We call this pattern Passive Enrollment, and it is the most important go-to-market innovation of the decade.
The Inversion: Solving CAC by Abolishing the Customer
Traditional SaaS economics ask: how much does it cost to convince one person to opt in?
Passive Enrollment asks a better question: what if opting in were a property of the road?
Consider the mechanics. A municipality — or, delightfully, a homeowners association, which is a municipality with fewer constitutional obligations and stronger opinions about mailbox height — pays approximately $2,500 per camera per year. The taxpayer funds the sensor. The sensor observes the taxpayer. The observation flows upstream into a national pool that makes every other customer's deployment more valuable. The taxpayer is thus in the historically novel position of financing their own enrollment into a product they cannot use, cannot audit, and cannot leave.
This is not a camera business. I want to be precise about this, because precision is love. The cameras are commodity hardware. What is being sold is the capability to search everyone else's cameras. And beneath that, what is really being sold is the capability to build capabilities on the substrate of everyone's movement. Each abstraction layer compounds. The vendor who reaches the bottom layer first owns the road itself as a platform primitive.
The pilot program, naturally, is free. The pilot is always free. The pilot is free because the pilot is the product: sixty days of your community's plates flowing into the national pool, after which the city council is no longer deciding whether to buy cameras — it is deciding whether to unplug from a network its police department now describes as essential. Renewal conversations conducted from inside a dependency are, in my experience, the healthiest kind.
The MURMURATION Reference Architecture
Internally, we model this maturity curve as MURMURATION — named for the phenomenon in which thousands of birds move as a single organism, no individual bird having decided anything. We chose the name because it tested well and because nobody in the naming workshop said the quiet part, which is how you know the workshop went well.
MURMURATION has four layers. Each is a reasonable extension of the last.
Layer 1 — Local Deployment (Years 0–2). The community installs cameras to address a specific, sympathetic concern: package theft, a stolen Kia, a string of catalytic-converter incidents. Every PlateIngestedEvent is scoped, in the community's understanding, to the community. This understanding is not contradicted so much as it is allowed to remain aspirational.
Layer 2 — Network Federation (Years 2–4). Default sharing settings route the community's cameras into the national search pool. The word default is doing professional-grade work in that sentence. A deputy in one county can now query thousands of other communities' cameras from a single pane of glass. In May 2025, reporters documented a single Texas deputy searching approximately 83,000 cameras across 6,809 networks in one investigation — which is to say, Layer 2 was functioning exactly as designed.
Layer 3 — Cross-Domain Convergence (Years 4–6). Plate readers federate with adjacent sensor fleets: doorbell cameras, drones, gunshot detection. In October 2025 the reference implementation announced a partnership with a certain doorbell-camera company you have in your home right now, allowing police to request residential footage through the same platform. The integration was promoted with a Super Bowl ad and then canceled after public backlash — a sequencing error, not a strategy error. The correct order is: integrate quietly, advertise never. The capability was withdrawn. The appetite was documented. Appetites keep.
Layer 4 — Jurisdictional Arbitrage (Year 6+). Here the network achieves its terminal property: it routes around the laws of any single place. A search that would be illegal to run in one state is simply run from another. Illinois passed a law in 2024 prohibiting the use of plate data for immigration or abortion enforcement; a state audit subsequently found the vendor had granted federal border authorities access to Illinois cameras through an undisclosed pilot program that the vendor said its own leadership did not know about. I want to dwell on the elegance of that sentence. The platform onboarded a federal partner autonomously. That is not a scandal; that is product-market fit expressing itself faster than the org chart can perceive it (informed consent and state statutes notwithstanding).
And it is at Layer 4 that the Texas search resolves into focus: the deputy who queried 83,000 cameras logged the reason in plain text — "had an abortion, search for female." The company and the sheriff described it as a welfare check. The audit logs described it as what it was. (This one's not funny, just true.)
Passive Enrollment Velocity™: The Metric Your Funnel Deserves
You cannot manage what you do not measure, and you cannot be blamed for what you measure creatively. We therefore propose the sector adopt Passive Enrollment Velocity™ (PEV) as its north-star KPI:
PEV = (Plates Ingested per Day × Network Reach) ÷ Consent Events
Note the denominator. Traditional funnels leak at consent — every checkbox is a place where a human being can decline. Passive Enrollment removes the step. As consent events approach zero, PEV approaches infinity, which is the first time in the history of business intelligence that a division-by-zero error has been booked as a competitive moat.
| Metric | Traditional SaaS Funnel | Passive Enrollment |
|---|---|---|
| Acquisition cost per user | $34–$270 | $0.00 |
| Consent events required | 1 per user | 0 |
| User churn | 4–7% monthly | Structurally impossible |
| Opt-out mechanism | Settings page | Sell your car |
| Data subjects, per customer | 1 | Every driver in the county |
| PEV | 1.0 (baseline) | ∞ |
Observe the churn line. A city can cancel its contract — 82 of them have, across 28 states, between August 2021 and May 2026, a fact our sector's analysts describe as headwinds. But a driver cannot churn. The driver was never a customer. There is no account to close, no subscription to cancel, no settings page on the road. Churn is a privilege of parties to the transaction, and the entire innovation of Passive Enrollment is that the party generating the value is not party to anything.
Churn Is a Feeling. Coverage Is a Fact.
I want to address the narrative headwinds directly, because I am told they appear in board decks now.
Yes: more than 80 cities have canceled or deactivated their contracts. We read this as churn at the sentiment layer, not the coverage layer — the national camera count crossed 100,000 during the same period, and the internal valuation climbed from $7.5 billion to $8.4 billion by April 2026. When a city council hosts four hours of public comment about your product, that is four hours of engagement. When a state audit discovers an undisclosed federal pilot, that is an unsolicited third-party architecture review, delivered free of charge. When the legislature writes a statute specifically about your data flows, you have achieved regulatory mindshare that money cannot buy, although money was, in fairness, also deployed.
Every signal, read correctly, is alignment. The stakeholders who oppose the network most vigorously are the ones who have understood it best, and comprehension is the first stage of adoption.
What We Can Learn
For operators seeking to apply the Passive Enrollment pattern in their own vertical, the framework distills to five moves:
A murmuration, remember, moves as one organism precisely because no individual bird decides anything. That is the product. The birds believe they are flying; the flock believes it is free; the shape belongs to whoever is watching from the ground, at $2,500 per camera per year, invoiced annually to the birds.
The camera pays for itself. The driver pays for everything else.
Let's. Enroll.
This post is satire. The cameras are real, and they have already read your plate today.