The Challenge
Creating a defensible decarbonization plan across a large portfolio of buildings
Amazon came to Carbon Signal with challenges that many large-portfolio owners might relate to. As an organization that owns and leases large numbers of buildings around the world, there were key issues that made it difficult to manage these properties and develop a portfolio-wide approach to decarbonization.
The portfolio itself was a black box. Utility data is hard to find and difficult to verify. Documentation of building systems isn’t always digitally available or consistently formatted. The portfolio is always shifting as leases are renegotiated and buildings are bought and sold, so the time spent setting up connections to building management systems and developing other data pipelines often goes to waste.
The scale of the portfolio made it difficult to apply traditional techniques for energy management and strategy development. In-person audits are expensive, time-consuming, and only represent a single snapshot in time. Trying to develop bespoke plans for individual buildings using these techniques, while keeping the plans up to date, can therefore be time and cost prohibitive.
Everyone across the organization was working from a different set of information toward different goals. Facilities managers are responsible for optimizing day-to-day operations. Sustainability strategists are focused on meeting long-term energy or carbon goals. And executives are responsible for aligning capital plans with overarching business objectives. There isn’t a single source of truth or an easy way for these different stakeholders to coordinate.
For Amazon, like many large portfolio owners, these challenges made it difficult to build a defensible decarbonization plan for their global corporate office portfolio. Their sustainability team knew what was important, but they didn’t have the data to back it up, and the shifting nature of their portfolio made it difficult to make progress collecting information. Without a clear baseline or quantifiable outcomes, stakeholders weren’t aligned on priorities.
The Solution
Quantifying uncertainty and building confidence
Carbon Signal was engaged to fill in the data gaps, drive alignment, and grow decision confidence. Rather than sink more resources into physical audits, Carbon Signal’s algorithmic inference engine created a full picture of their portfolio. Using only the building’s location, size, type, and - when available - monthly utility data, Carbon Signal generated a complete baseline energy model. Using AI, it inferred baseline building characteristics from available data, with a quantified level of confidence (if there’s very little data available, Carbon Signal still generates a complete model, albeit one with a wider range of uncertainty). As more information was collected, the model dynamically adjusted, narrowing the uncertainty bands and producing higher fidelity results. Carbon Signal’s AI system is constrained by physics-based relationships between building systems, so all the model inputs and outputs were transparent and coherent in relation to one another.
The Impact
Defensible plans. Team alignment.
Now, Carbon Signal provides a shared basis of understanding for all stakeholders to start making decisions. Teams no longer need a detailed building audit before they can underwrite a decarbonization plan. They immediately understand where they have confidence to act, where they don’t, and what information is needed to bridge the gap. Even though each stakeholder might still have different priorities, everyone is working from the same set of assumptions, so they can understand the tradeoffs involved with meeting different goals.
With just a small strategy team, Amazon can quickly answer important questions: What is the realistic performance target for each building? What are the biggest opportunities across the portfolio? How much will it all cost?
Rather than waiting for audits or relying on simple rules of thumb, they now have the data to back up their decisions. At every layer of decision making, they understand the risks and opportunities with quantified levels of confidence.
Carbon Signal gave us the ability to prove our hypotheses at a portfolio scale and compare interventions side by side. This changed the entire nature of the conversation.
-Amazon
About Amazon
The Global Real Estate and Facilities (GREF) team at Amazon manages and optimizes Amazon’s physical workplaces, including corporate offices, data centers, and fulfillment sites worldwide. GREF professionals handle real estate transactions, facility operations, workplace safety, and sustainability initiatives to support Amazon’s business needs. Their work ensures that Amazon employees have safe, efficient, and innovative environments to work in.
