The open web is not a marketplace. It is a network of data nodes.
The industry’s dominant mental model is a transaction marketplace: impressions, auctions, clicks, conversions. That model describes what happens. Understanding why some participants compound in value over time while others plateau requires a different framing.
The open web is best understood as a network of independent nodes where data objects move and communicate between them. Every major participant in the ecosystem is a node.
A publisher sitting on years of behavioral and contextual signals. An SSP aggregating inventory across thousands of sites. A DSP building targeting and measurement models from impression data. An agency managing CRM lists on behalf of dozens of brands. A publisher aggregator sitting on top of independent publisher graphs at scale. Each one is a distinct data asset with its own signals, its own value, and its own strategic position in the network.
What that node holds matters enormously.
Transaction metrics are outputs, not drivers.
The transaction-first view of the open web trained an entire industry to optimize at the impression level. Floor prices, bid density, win rates, CTR. These are real metrics and they matter — but they are outputs. They describe what happens when data objects arrive at the auction, not what determines the quality of those objects before they get there.
The node-network view asks an earlier question: what happens to a signal between the moment it originates and the moment it reaches a buyer? How much fidelity does it carry? How accurately is it attributed? How cleanly does it resolve against the identity infrastructure on the other side?
Signal fidelity, match rates, attribution accuracy — these are network properties rather than auction properties. Value is created or destroyed at every point data moves or connects between nodes.
The asset is the node. The advantage is the network.
Every node type has a different version of this problem — and the node-network frame makes each one visible where the transaction view obscures them.
For publishers, the question is whether their signals are traveling as an independent, well-attributed data object or dissolving into a pool someone else controls. A publisher with a structured, owned identity graph sends a different kind of signal into the bidstream than one whose data has been aggregated, anonymized, and re-packaged by a third party. The first appreciates in value as the graph deepens. The second gets commoditized.
For SSPs, competitive differentiation is increasingly a function of signal quality on the inventory they route, not just the volume of inventory they access. The nodes they connect to determine what they can offer the demand side. Better-connected nodes produce richer bid requests, and richer bid requests attract stronger buyers.
For DSPs and agencies, model accuracy is downstream of signal fidelity. Targeting models trained on well-attributed, consistently structured first-party signals from independent publisher nodes behave differently than models trained on pooled, processed, third-party-touched data. The quality of the network you’re drawing from sets a ceiling on what you can build.
Across all of them, the common thread is the same: in a network of nodes, competitive position is a function of the quality of the signal and the connections between nodes — and who controls them.
Infrastructure isn’t neutral by default.
The industry has operated on an assumption that the infrastructure connecting these signal nodes is a passive utility. A highway. You pay a toll, your data travels, you get reach. Publicis’s announced acquisition of LiveRamp made that assumption impossible to hold.
When a company that has functioned as core identity ‘neutral’ infrastructure for the ecosystem gets absorbed by an agency holding company, every node that was routing data through it must now face the question: does the infrastructure connecting me to the rest of the network now have a stake in the outcome? Is it dispassionate? Do advantages accrue to a non-neutral entity understanding my signals and their place within other buyer strategies?
Publicis committed that LiveRamp will operate as a neutral, interoperable platform. Whether that holds or is believed is a question every participant will have to answer for themselves. But the node-network frame makes the underlying reality visible: every participant is already running on infrastructure built and owned by someone.
Adtech is the current use case. It won’t be the last.
The open web will keep changing. Which data nodes build durable, independent positions will come down to who owns the signal and who controls the connections.
Trip Foster is Co-founder and CRO of GrowthCode. GrowthCode provides identity and data infrastructure as-a-service for the digital advertising ecosystem.