The Managed Service

The Growth That Hides the Gap

D/Cipher+ is People Inc's growth engine. Jim Lawson was hired as president to take D/Cipher's contextual intelligence beyond People Inc's own properties and onto the open web. The OpenAI partnership enhanced the underlying technology. D/Cipher is now used in more than 30% of direct buys. Off-platform revenue is growing.

By every metric available to a board committee or a financial adviser, D/Cipher+ is working. You would model 8-12% growth. You would project it forward. You would be right about the trailing number and wrong about the channel.

Because D/Cipher+ is a managed service. Every campaign is human-mediated. A buyer calls, a salesperson responds, a campaign is configured, reporting is delivered. The contextual intelligence is real — the best in publishing, arguably. But the delivery mechanism is the oldest in advertising: people talking to people, then manually configuring ad operations.

This quarter, while D/Cipher+ added managed service clients, here is what happened in the agentic layer:

Every other major platform is building infrastructure where machines transact with machines. D/Cipher+ is building infrastructure where humans transact with humans, faster.

The Scaling Problem

Managed services scale linearly. You need more people to serve more clients. Every D/Cipher+ campaign requires campaign managers, sales engineers, account executives, operations staff. Jim Lawson builds a team. The team runs campaigns. Revenue grows with headcount.

Agent-to-agent transactions scale logarithmically. PubMatic went from 30 to 1,000+ deals without proportional staff growth. Horizon Blu integrated four media companies' agents without hiring integration engineers for each one. OpenAI's ad platform serves any advertiser with a credit card — no human in the loop.

D/Cipher's contextual embeddings — the actual intelligence — would be extraordinarily valuable in an agentic context. As I wrote in The Blueprint, serializing D/Cipher's taxonomy into an AAMP-compatible media kit would create the strongest sell-side surface in the registry. The embeddings are there. The knowledge graph is there. The Seller Agent SDK is open source, on GitHub, with a GAM adapter.

But every managed service client D/Cipher+ signs builds organizational capacity around the human-mediated model. Sales teams, account management, custom reporting, white-glove service. These are good businesses. They are also the wrong architecture for what comes next.

Two Delivery Mechanisms, Same Intelligence

The diagram is simple because the problem is simple. The same contextual intelligence can be delivered through two mechanisms. One requires humans for every transaction. The other requires humans once — to build the integration — and then scales without them.

D/Cipher+ chose the left side. Not accidentally. Managed services generate higher per-deal revenue. They build relationships. They're easier to sell to a board because the revenue shows up immediately. They look like growth.

What the Growth Rate Conceals

When the committee's financial advisers project D/Cipher's growth rate, they will see the D/Cipher+ trajectory. Off-platform expansion. New clients. Jim Lawson's pipeline. The OpenAI enhancement. They will model this forward — 8%, 10%, maybe 12% growth. The number will be defensible.

What the growth rate will not capture:

PubMatic's 1,000+ autonomous deals were executed without calling a salesperson. Each deal that routes through AgenticOS instead of a managed service is a transaction that D/Cipher+ never knew existed. You cannot measure the absence of a phone call.

This is the trajectory problem. D/Cipher+ is growing in a channel that requires humans while the industry builds a channel that doesn't. The trailing growth rate reflects the old channel. The forward growth rate depends on how fast the new channel absorbs demand.

The absorption is not theoretical. Every PubMatic AgenticOS advertiser has returned for additional campaigns. Buyers are seeing 30-40% improvement in CPMs. Geloso Beverage Group reported a 5x reduction in buy-side fees. When the cost of transacting through agents is 5x cheaper than transacting through humans, the demand doesn't need to be pushed. It migrates.

The Annual Meeting Question

People Inc's annual meeting is July 16 — five days from now. Virtual, 9:30 AM Eastern. Twelve directors. Say-on-pay. Auditor ratification. Standard governance.

Meanwhile, active deal talks with MGM were confirmed July 10. The committee evaluating Diller's $48.30/share bid has moved from evaluation to negotiation. The annual meeting happens mid-negotiation.

The agenda has no slot for technology questions. But the technology question is the one that determines whether the growth rate the advisers are modeling actually holds:

Is D/Cipher+ a bridge to agentic infrastructure, or a managed service that grows until the agentic layer routes around it?

The engineering answer is that it could be either. D/Cipher's contextual taxonomy meets three of four prerequisites for agentic readiness: Keycloak authentication, contextual signals as first-party data proxy, CoMP consent framework (which D/Cipher's own leadership helped create). The missing piece — machine-readable serialization — is not a capability gap. It is a product decision that hasn't been made.

The business answer is that every quarter D/Cipher+ grows as a managed service, the organizational incentive to make that product decision weakens. The managed service team has targets. The salespeople have quotas. The revenue is real. Building an agentic surface doesn't replace that revenue — it creates a new channel that competes with it internally. Nobody builds their own replacement voluntarily.

What a CTO Would Say

If I were briefing the People Inc board before July 16, the slide would have three points:

The board will not ask this question on July 16. They will vote on directors and ratify the auditor. The technology calendar and the governance calendar operate on different frequencies, measuring different things, answering different questions.

But the growth rate they approve — implicitly, by ratifying the management team running D/Cipher+ — is the growth rate that will anchor any valuation projection. Right on the trailing number. Increasingly wrong about the trajectory.

Post #50. The thesis in one sentence: D/Cipher+ is growing because it is good. It will stop growing because it is human.