Collecting Scope 3 energy data from suppliers and tenants

For most companies, the largest part of the carbon footprint is not in their own operations. It is in the value chain: the goods they buy, the assets they lease, and the buildings their tenants occupy. CDP notes that supply chain emissions are, on average, far larger than a company's operational emissions, which is why Scope 3 now dominates serious disclosure.

The catch is that the energy data behind Scope 3 belongs to other people. A landlord needs consumption from tenants; a tenant needs it from the landlord; a manufacturer needs it from suppliers. Collecting that data, in a usable and consistent form, is the hardest part of value-chain reporting, and it is a data logistics problem more than an accounting one.

The two Scope 3 categories that trip up building owners

The GHG Protocol Corporate Value Chain (Scope 3) Standard defines fifteen categories. Two of them turn on who controls the meter. Category 8, upstream leased assets, covers assets a company leases and operates but does not own, where the energy shows up in someone else's accounting boundary. Category 13, downstream leased assets, is the inverse: assets a company owns and leases out to tenants, whose energy use is the owner's Scope 3.

This is where landlords and tenants get stuck. A property owner reporting Category 13 needs each tenant's electricity and gas consumption. A tenant reporting Category 8 needs the landlord-controlled common-area and base-building data. Neither side usually has the other's bills, and lease terms rarely require sharing them.

Primary data beats estimates, but it is harder to get

There are two ways to fill a Scope 3 category: spend-based estimates using average emission factors, or primary data collected from the actual supplier or tenant. The GHG Protocol's guidance on collecting data favors primary activity data where it is available and material, because estimates cannot show real reductions. When a supplier cuts its energy use, a spend-based model will not detect it; only their actual consumption will.

How the leaders collect value-chain data

Large buyers do not chase suppliers one email at a time. They use structured programs. Through CDP's supply chain program, hundreds of purchasing organizations representing trillions of dollars in spend request environmental data from tens of thousands of suppliers each year. The model works because it standardizes the ask and the format, so responses are comparable.

The same principle applies to buildings. The most reliable way to get tenant or landlord energy data is to make sharing structured and low-effort: standard data formats, sub-meter reads, and automated feeds rather than annual requests for scanned bills. Where interval data or standardized utility data access exists, the collection can be automated instead of negotiated every year.

Response rate is the quiet killer of value-chain programs. Even with a structured request, many suppliers and tenants do not reply, or reply with incomplete data, in the first year. The reporters who improve over time treat engagement as an ongoing relationship rather than an annual demand: they explain why the data is needed, make the format easy, and let counterparties reuse what they already report elsewhere. Each year the share of primary data rises and the reliance on estimates falls, which is exactly the trajectory assurers and investors want to see.

Green leases and the data clause

The structural fix for landlord and tenant data is contractual. A growing number of leases include provisions that require both parties to share energy and water consumption data, often called green lease clauses. Where those clauses exist, the collection problem becomes a process problem rather than a negotiation. Where they do not, the reporting party is left asking counterparties for data they have no obligation to provide, which is why response rates for tenant and supplier requests are often low. Building the data requirement into new and renewing leases is the durable answer for a property portfolio.

Not every category needs primary data

Primary data is the goal, but chasing it everywhere wastes effort. The GHG Protocol expects reporters to prioritize by materiality: spend the collection effort on the categories and counterparties that dominate the footprint, and use estimates for the long tail. For a property owner, downstream leased assets may be the single largest category, which justifies collecting tenant meter data directly. For a manufacturer, a handful of key suppliers may account for most of the upstream emissions. Mapping materiality first tells you where primary data pays off and where a spend-based estimate is good enough for now.

A collection approach that scales

1. Map which Scope 3 categories are material, especially upstream and downstream leased assets for property portfolios.
2. Decide per category whether primary data is worth collecting or a spend-based estimate suffices for now.
3. Standardize the request: one format, clear fields, and a fixed period, whether for suppliers or tenants.
4. Prefer automated data access (interval feeds, standardized utility data) over manual bill collection where available.
5. Keep every collected figure with its source and method so assurance and year-over-year comparison hold up.

The data-quality problem underneath

Even when suppliers and tenants do share data, it arrives inconsistent: different units, formats, billing periods and commodities. A pile of PDF bills and portal exports from a hundred counterparties is not a Scope 3 inventory. It has to be extracted, normalized and reconciled before it can be summed. That normalization work is exactly what breaks Scope 3 programs built on spreadsheets and manual entry.

Data quality is now part of the disclosure itself. The GHG Protocol and the reporting frameworks increasingly expect reporters to describe how good their Scope 3 data is: whether it is primary or estimated, how complete it is, and how it was calculated. A tenant electricity figure pulled from an actual bill is a different quality of evidence than one modeled from floor area, and a credible inventory records that difference. When suppliers and tenants send raw bills, preserving the link to the source document is what lets you defend the quality rating later, and it is the first thing an assurer will look for.

The moment a program moves from a handful of counterparties to dozens or hundreds, manual collection stops working. Each supplier or tenant sends a different format on a different schedule, and a team keying those into a spreadsheet cannot keep pace or keep the quality consistent. Errors compound, the audit trail frays, and the inventory becomes something the reporter cannot fully stand behind. The programs that scale are the ones that automate extraction and normalization so that adding a counterparty is a small marginal cost, not another manual task.

How MartinAI fits

MartinAI removes the manual work of turning other people's utility bills and interval data into structured, comparable numbers. Whether the data comes from a tenant, a landlord or a supplier, and in whatever format or commodity, it runs through the same extraction and normalization, so consumption is isolated and tied to its source. That gives Scope 3 categories built on leased assets and supplier energy a primary-data foundation instead of a spend-based estimate, without a team hand-keying counterparty bills. MartinAI structures the data; it does not file or assure your inventory.