How to calculate your carbon footprint with incomplete data

You can calculate a carbon footprint with incomplete data, and the standards explicitly allow for it. The conditions are that every gap is filled with a declared method, that it is written down which figures are measured and which are estimated, and that uncertainty is documented. A calculation with disclosed estimates is auditable; one that hides them is not.

What do the standards actually allow when data is missing?

Considerably more than most people assume. ISO 14064-1:2018 requires an uncertainty assessment at inventory category level, and explicitly accepts a qualitative assessment where quantitative estimation is not possible or not cost effective, provided the decision is justified. It also requires documenting the data characteristics of each source and justifying the choice of calculation models.

The GHG Protocol works the same way: it accepts secondary data and estimates, and in exchange requires the report to describe the types and sources of data used and the data quality of the emissions reported. What neither framework allows is presenting an estimate as a measurement. The detail on how the two frameworks differ is set out in the comparison of GHG Protocol versus ISO 14064-1.

In Spain, Royal Decree 214/2025 requires companies within its scope to calculate their carbon footprint annually and to prepare and publish a reduction plan. It mandates scope 1 and scope 2 and leaves scope 3 voluntary for private companies. Registration in the MITECO registry is also voluntary for private companies, and mandatory for state public sector bodies. In other words, the minimum legal obligation is covered by data that usually sits in invoices.

A phased method when the data is not complete

The most expensive mistake is doing this backwards: spending three months chasing an exact figure for a source worth 0.4% of the inventory while the line that carries a third of it stays untouched. The screening phase exists precisely to prevent that. If you have never built an inventory, start from what a carbon footprint is and how it is calculated.

What is a proxy and when is it acceptable?

A proxy is a substitute figure that stands in for real data following an explainable rule. It is not making a number up: it is applying a known relationship to a quantity you do know. These are the typical gaps and the proxies that survive a review.

Working rule: a proxy is acceptable if another person, reading your documentation, can reproduce the same number without asking you anything. If they cannot, it is not a proxy, it is a guess.

How do you document uncertainty and keep the calculation auditable?

With two tools. The first is the GHG Protocol's five data quality indicators: technological, temporal and geographical representativeness, completeness and reliability. Scored line by line or category by category, they show immediately where the weak data sits. The GHG Protocol also publishes dedicated guidance on uncertainty assessment in GHG inventories for anyone who wants to quantify it.

The second is the assumptions log, which in practice is what saves a verification. One row per estimated line, with: emission source, period, figure used, where the figure came from, method or proxy applied, emission factor with its source and year, who entered it and when, and a confidence level. ISO 14064-1 asks for exactly this: documenting the data characteristics of each source and justifying the models used.

One detail that gets forgotten: uncertainty is not declared only for the total, but per inventory category. You can have a very solid scope 1 and a mostly estimated scope 3, and that is entirely valid as long as the report says so clearly. If formal verification is the goal, this log is the core of the evidence the verification body will review, as covered in Manglai's ISO 14064 solution and in the explainer on what ISO 14064 is.

How much can you leave out?

Less than you would like and more than you fear. ISO 14064-1 requires you to define your own significance criteria, apply them through a documented process to decide which indirect emissions are included, and justify any exclusion of significant indirect emissions. It sets no percentage: it sets a process.

The GHG Protocol does not impose a coverage threshold for scope 3 today either. It is revising its whole corporate suite, the Scope 3 Standard included, so this point may move and is worth tracking; in the meantime the reference remains your own significance criteria, documented and applied consistently every year.

To decide which scope 3 categories to calculate and which to leave out with a defensible rationale, the criteria are set out in the 15 scope 3 categories of the GHG Protocol: estimated magnitude, financial significance, influence over the activity, risk exposure, stakeholder interest and sector guidance.

Which data is non-negotiable even without an internal team?

There is a core set that does not admit a proxy, because the data always exists and anyone can get hold of it:

• Purchased electricity per supply point, in kWh, from the retailer's invoices.
• Natural gas and other stationary fuels, in kWh or invoiced physical units.
• Own fleet fuel, in litres, from fuel card invoices.
• Refrigerant top-ups, in kilograms and by gas type, from maintenance records.
• Floor area, average headcount and turnover for the year, which underpin every intensity metric and every later extrapolation.

Those five blocks close scope 1 and scope 2 with real data, which is precisely what Royal Decree 214/2025 requires of the companies within its scope. Everything else can start as an estimate. Obligations and deadlines are set out in the guide to the carbon footprint registry and the Royal Decree 214/2025 obligations.

A three year data improvement plan

This plan has a second benefit: it makes years comparable. If you improve the method, recalculate the base year with the new method and say so, or your apparent reduction is just a methodology change. Once data capture is automated, the reporting frequency stops being annual and monitoring becomes continuous, a shift explained in the analysis of real-time carbon footprint data.

Mistakes that do invalidate a calculation

• Mixing years: 2025 consumption with 2022 factors, undisclosed.
• Changing method between years without recalculating the base year.
• Using monetary factors without deflating to the factor's reference year.
• Double counting the same emission, typically between scope 2 and scope 3 category 3.
• Leaving exclusions unjustified when the emission is significant under your own criteria.
• Losing track of who entered each figure and where it came from.
• Rounding estimates down so the total looks better.

Frequently asked questions

Can I submit a carbon footprint with estimates to MITECO?

The MITECO carbon footprint registry accepts calculations that follow a recognised methodology, and those methodologies contemplate the use of documented estimates. What gets reviewed is the coherence and traceability of the calculation, not whether every figure is metered. For private companies, registration remains voluntary.

What is the difference between an estimate and an invention?

Traceability. An estimate starts from a known figure, applies an explainable rule and is recorded with its source and method. An invention is a number with no origin. The first survives a verification with an observation; the second sinks it.

How much uncertainty is too much?

There is no universal threshold. ISO 14064-1 asks for uncertainty to be assessed per category and accepts a qualitative assessment where quantification is not feasible or cost effective, provided the decision is justified. In practice the problem is rarely the level of uncertainty, it is that nobody declared it.

Is it worth starting with spend data?

Yes, as a first layer. A screening with monetary factors tells you within weeks where most of your footprint sits and stops you spending the budget on the irrelevant part. The mistake is staying there for several years: a spend factor cannot tell an efficient supplier from an inefficient one, so it will not reflect your reductions either.

Do I need an internal sustainability team to do this?

Not for the first inventory. Scope 1 and 2 come from invoices finance already holds, and the scope 3 screening comes from cost accounting. What you do need is one person accountable for maintaining the assumptions log and keeping the criteria stable from year to year.

Working with incomplete data is a traceability problem, not a formula problem: knowing at all times what is measured and what is estimated. For a first figure without setting anything up, the carbon footprint calculator handles the initial screening, and Manglai's carbon footprint software keeps the origin of every figure on record for when verification comes.