Publicis grew at roughly 6% last year. The broader agency market grew at under 2%. Publicis will tell you how: data products. These are owned assets that allow them to generate margin without depending on additional headcount, billable hours, or creative retainers.
The traditional business model for independent agencies (service- and time-based billing) is in the cross hairs of changing economic and technology-related forces. AI compresses the time-to-delivery and flattens out the differentiation that agencies used to charge for, making it harder to keep pace and prove value.
Clients and procurement teams recognize this shift; so do the holding companies that have spent billions to get ahead of these changes and now have a structural advantage that pure-service independents cannot easily overcome.
Independent agencies still have a strong foundation of expertise, relationships, and unique first-party data that provides an edge when it comes to quality of service. But without a move to product-based business models, those institutional advantages may not be enough.
Whether you’re actively building the data infrastructure to support this transition or just beginning to hear the warning bells, your agency needs a clear understanding of the opportunity, a view of the potential challenges and competitive landscape, and the right intelligence to take the next step.
What are data products for agencies?
A data product is a packaged, repeatable asset built from proprietary data and intelligence that an agency owns, operates, and sells or deploys independently of any single client engagement. Unlike a service, which is delivered through people and time, a data product generates value from the asset itself. In this way, it lightens the lift to improved margins.
It’s important to distinguish between true data products and services that are delivered using data. True data products are:
- Owned by your agency
- Unique differentiators that can’t be replicated by competitors
- Repeatable and available to a broader client base for recurring revenue
- Based on your underlying capability, not reports (e.g. proprietary audience platforms; subscription-based intelligence; licensed measurement frameworks)
What does a product-based business model look like in practice?
Agencies that are embracing this approach have built an ecosystem around delivering this model called the consumer data layer. This is a unified environment where client first-party data, identity infrastructure, and licensed intelligence combine into a continuously updated view of the customer across devices and channels. The outputs of that data layer can be monetized in products because they cannot be produced by your competitors. They also allow for consistent measurement across accounts, so it’s easier for you to demonstrate your value with outcomes.
AI is a big part of this story, too. It’s the engine that allows agencies to turn the consumer data layer from a static repository into a predictive asset. It continuously processes consumer signals at scale, for actionable, individual-level insights of who audiences are now, in real time. This means that the products produced on top of this foundation are current and defensible. In other words, worth paying for.
Holding companies have invested years and billions of dollars to build an infrastructure to support data products, and independent agencies have been trying to replicate this without the same raw materials. The technology infrastructure (often modeled at data lakes, standing up in Snowflake, connecting identity resolution, even some of the AI capability) is within reach, but the intelligence that powers the consumer data layer has proven harder to replicate.
There are some common data challenges that agencies have to tackle before they have the proper intelligence to compete against the holding companies.
Why data products require more than demographics and transactions
Most population-level data available to agencies suffers from limitations when it’s not enriched. Population-level data is demographic and transactional, modeled at the household level. It tells you where someone lives and approximates what their household bought last quarter. As a predictor of future behavior, that is a weak foundation.
Holding companies solved this by building proprietary data collection and modeling infrastructure. They generate the most recent individual-level predictions, revealing not just who a consumer is but why they act and what they are likely to do next.
This understanding is possible when you know the values of your customers, the motivations that lead to a purchase, the context of their lived experiences that just don’t show up in transaction data. This level of understanding is called predictive consumer intelligence, and it’s what gives data products differentiated value.
It is also what has been largely inaccessible to agencies that cannot acquire a data company to get it.
There is also a structural problem with how most enrichment products work. Standard data enrichment requires a client to submit a first-party file before anything can happen. You send a list of customer IDs or hashed emails, the data gets matched and returned. That works for targeted enrichment, but it creates a bottleneck for agencies trying to build a foundational data layer.
What building a product-based model actually looks like
Agencies that have successfully made the shift to data products share a few things in common. They treat consumer intelligence as an owned infrastructure investment, not a campaign line item. They build for their entire book of business, not a single client. And they choose data partners that give them access to predictive consumer intelligence without restricting how, where, or with whom they use it.
That last part is especially important. Many data providers sell population-level data with licensing structures that create real friction: per-client terms, usage restrictions, requirements to route activation through the provider’s own systems. For an agency trying to build a proprietary product, those are huge restrictions. The data is only as useful as the operational freedom to use it.
A real data foundation installs directly into the agency’s own environment, refreshes continuously, and carries no downstream restrictions. One license that covers the full population. One source of truth that works across every client in the portfolio without renegotiating terms each time a new account comes on.
Moving from the consumer data layer to data products
Once that foundation is in place, the product development follows. Client first-party data layers in on top, giving the agency and its clients a shared view of the right audiences to target and why. Measurement becomes consistent across accounts because the underlying consumer view is the same. Audience products built on individual-level predictive data can be packaged, sold, and scaled. This is where margin comes from.
The agencies that make this shift now will have a durable advantage over those that do not. The ones that wait are going to find their clients asking harder questions about what, exactly, the agency provides that they could not get from a platform or a holding company’s proprietary stack.
Predictive Consumer Intelligence Install: The Cheat Code to Level the Data-Product Playing Field
Resonate built Predictive Consumer Intelligence Install to address exactly this problem. Rather than requiring a client file to initiate enrichment, this offering delivers Resonate’s full catalog of individual-level predictive attributes for use in an agency’s own environment: their data lake, their Snowflake instance, their analytics spine.
It covers 12,000+ attributes across psychographics, personal values, in-market intent signals across 30-, 90-, and 365-day windows, media consumption, and consumer preferences.
One license covers the full US adult population, across all your known ID types. Agencies can apply it across every client without per-client friction or usage restrictions. The data is built by Resonate directly, so there are no co-licensing complications and no requirements to route activity through Resonate’s systems (unless you choose to). The intelligence belongs to the agency to use how and where they need it. And when client first-party data comes in, it layers directly on top of that foundation, producing a more complete picture of the customer than either source provides alone.
Predictive Consumer Intelligence Install is available as a standalone Data-as-a-Service offering and alongside Resonate Ignition, the agency operating system built for insight, activation, and measurement.
The holding companies spent years getting to this starting point. Agencies that want to compete with them on product do not have to spend the same years getting there. To learn more about leveling the playing field and taking the next step in your agency’s evolution, schedule some time to discuss Predictive Consumer Intelligence Install with us today.
