The $25 Million Reason Georgia-Pacific Rebuilt Its Data Foundation

How Matt Robuck is turning America’s paper giant into a decision-making powerhouse — one trusted dataset at a time.

In a paper mill, every hour of downtime carries a steep price tag, with tens of thousands of dollars lost the moment a machine goes offline. To protect against that risk, site managers do exactly what sound operational logic dictates: they maintain robust inventories of spare parts. It's a responsible decision made at the local level, repeated across numerous manufacturing sites nationwide.

But when that same behavior plays out simultaneously across a company the size and complexity of Georgia-Pacific, the cumulative effect creates a paradox. Inventory accumulates faster than it can be tracked or reconciled. Parts that are plentiful at one facility get purchased new at another, because without reliable, unified data, there's no practical way to know what already exists (or where). Over time, that gap between what the data says and what's actually on the shelf compounds into a big business problem.

For Matt Robuck, Vice President of Data at Georgia-Pacific (GP), this wasn't a failure of operations. It was a trust problem. And solving it required something more fundamental than better tooling. It required rebuilding confidence in data itself.

Engineering a data team business impact

Three and a half years ago, Georgia-Pacific had no centralized data function. Data engineers were embedded across individual business units, each operating with their own tools and priorities, with little shared infrastructure, no common language, and no unified view across the company's vast operational footprint. The raw talent was there, but it was fragmented, and its potential to drive company-wide impact was largely untapped.

That's the environment Robuck stepped into, and the opportunity he was hired to change. "Our central data team has been around for three years," he notes. "Before that, we had data engineers scattered throughout the organization. Building something like our Data Intelligence Platform, named GP Data Pulse (Alation), wasn't even really a possibility at Georgia-Pacific, given how decentralized we were until three years ago."

GP is not a small operation. With 32,000 employees, facilities in over 150 locations, and three major business divisions — Consumer Products (Angel Soft, Brawny, Dixie), Building Products, and Packaging and Cellulose — the company spans an enormous operational and data landscape. Add international teams in Mexico, Canada, and India, and that complexity becomes exponential.

Robuck's mandate was clear: build a data capability from the ground up that could actually move the business forward by driving bottom-line impact. To do this, his mission statement was deliberately simple: "Enable more informed decisions faster."

But the journey from that mission statement to organizational reality would require a fundamental philosophical shift.

Rethinking the modern manufacturing firm

The insight at the center of Robuck's strategy is deceptively straightforward… and yet it reframes everything.

"We manufacture products that are in 65% of US households, but we also manufacture decision-making systems."

That sentence carries enormous weight. Georgia-Pacific has been perfecting the physical act of manufacturing for nearly a century. Automation has transformed the factory floor. But the decisions that govern those factories (when to buy, when to transfer, when to invest, when to stop) were still being made on incomplete, untrusted, and often contradictory data.

Robuck set out to change that. His team built the centralized data intelligence platform (internally branded "GP Data Pulse") that serves as the data connective tissue between GP's three businesses, its delivery teams, and its executive stakeholders. The goal wasn't just to catalog data. It was to create conditions in which the entire organization could act on data it actually trusted.

The $25 million discovery in a spare parts bin

The spare parts problem crystallized everything Robuck had been building toward.

Paper mills run 24 hours a day. When a machine goes down, every hour of downtime costs tens of thousands of dollars. So site managers do what any rational person would do: they stockpile spare parts. The trouble is, with over 200 sites doing the same thing independently, Georgia-Pacific had accumulated enormous quantities of duplicate inventory — parts sitting idle at one site that were desperately needed (and being purchased new) at another.

The data to solve this problem existed. Thirty different data sources, spread across the organization, held the answers. But when GP's teams tried to use that data, they ran into issues.

"When we get into your data solution, I look at the numbers one day, and then the next day they double, it's all over the place — I can't trust it," Robuck recalls users telling his team. The problem wasn't just technical. Business process issues, pipeline issues, and inconsistent data entry had eroded confidence to the point where people had simply stopped looking.

Robuck's team, along with business teams, took action.  Among a number of solutions was shoring up the data foundation and data quality, which included efforts like connecting data quality tools to the platform and eventually connecting those tools directly to the GP Data Pulse platform — in context, alongside the documentation, metadata, and lineage users needed to make sense of what they were seeing. The goal was to make data quality visible and trustworthy at the moment of decision, not as a separate technical exercise happening somewhere downstream.

As the foundation strengthened, data quality and business processes improved, and now, instead of automatically ordering a replacement motor from an external supplier, a procurement manager at one site can stop and ask: Does another GP facility already have this? The answer, increasingly, was yes.

"Last year we did about $25 million worth of intercompany transfers," Robuck says. “The GP Data Platform, which includes Alation as a critical component, has become the ecosystem that transformed how our teams discover and trust data, enabling initiatives like this that significantly impact our bottom line and position us to solve the next set of high-value opportunities even faster.”

The operating model behind the outcomes

Results like that don't happen by accident. They require an organizational model designed to connect data investments to business value, explicitly and continuously.

Robuck built his team around what he calls "integrated planning and business engagement." A dedicated solutions group within his organization is responsible for bridging the gap between data capabilities and business outcomes. They write what GP calls "venture summaries" — documents that project the P&L impact of specific data initiatives before a single line of code is written, tracked via net present value.

"If I talk to my executive team about data platforms, it can be a tough discussion," Robuck says. "But if I talk to them about how we’re moving the P&L with data, they listen."

That discipline has made the difference when renewal conversations come up at the executive level. When executives ask what a platform investment is actually delivering, Robuck doesn't cite catalog coverage metrics or governance scores. He points to venture summaries. He points to $25 million.

"These NPVs, these values, and these outcomes are enabled in part by the Alation platform,” he shares. “You take that out, and it's going to be much more difficult."

Building for the AI future

The spare parts story is a proof point, not a destination. Robuck is focused on what comes next.

Last year, GP deployed eleven AI agents. This year, the team is building many more. For example, the team is working on agents that analyze marketing campaign performance across platforms — automatically surfacing recommendations on where to shift spend, informed by years of historical data and institutional knowledge encoded into a governed knowledge layer.

"It’s likely that in a short time, we will have agents not only recommending and building digital marketing campaigns, but the technology will allow those agents to take actions and buy the media," Robuck notes. The stakes of getting the underlying data governance right are no longer theoretical.

This is where the role of a platform component like GP Data Pulse becomes something larger than a catalog. It becomes the governance layer that controls what an agent can know, what a playbook says, and who has the authority to change it. The ecosystem that makes AI trustworthy at enterprise scale.

The leadership lesson

Robuck is candid about the realities of leading data transformation in a complex enterprise. CDOs average 18 to 24 months in the seat. The business moves fast. Executive patience is finite.

His advice to data leaders navigating the same terrain is hard-won: connect to business outcomes early, visibly, and continuously. Don't wait for a perfect platform before starting the conversation about value. Find the footholds — the governance champions embedded in individual business units, the executives who are already data-curious — and build momentum through them.

Because at the end of the day, the goal was never to build a data catalog or a data platform. It was to help a 100-year-old company make better business decisions faster.

The motors are moving. The agents are running. And Georgia-Pacific is just getting started.

Curious to see how Alation can help you transform your business? Book a demo with us today.