Thanks for this interesting question! Many companies are tackling this challenge right now.

When dealing with uncertainties in emission data, especially Scope 3, the goal is not to eliminate all ambiguity (which is nearly impossible!!), but rather to:

1. Identify, quantify, and transparently communicate the uncertainty. This is necessary under GHG, ISO and CSRD reporting.

2. Improve data quality over time through better processes and supplier engagement. Step by step is a good approach and do not be too overambitious. thsi will take time.

3. Use conservative estimates to avoid underreporting. Communication of correction is way more difficult if you have underreported!!

Let’s break it down for most relevant Scope 3 categories:

Scope 3.1 Purchased Goods and services: 

Uncertainty sources:

• Emission factors used (e.g., average industry vs. supplier-specific)

• Incomplete or inaccurate data from suppliers

Mitigation strategies:

• Use supplier-specific data where possible (e.g., request LCA or EPD reports)

• Engage key suppliers on their carbon accounting practices

• Segment suppliers by spend or risk to prioritize engagement

• When using averages, apply conservative emission factors (Over reporting is from my point of view better than underreporting. Communication of correction is way more difficult if you have underreported!!)

Scope 3.11 Use of Sold Products:

Uncertainty sources:

• Customer usage patterns (duration, frequency, energy mix) –> primary data is mostly never available

• Assumptions about product lifespan

Mitigation strategies:

• Base estimates on standardized use cases or agreed upon use cases from industry networks

• Gather data from user studies or product telemetry

• Apply sensitivity analysis for different usage scenarios

If you found this helpful, we can dive into other Scope 3 categories as well.

Thank you for this question João! From my experience with companies, uncertainties in emission data are very common – what matters most is transparency about where estimates are used and how the methodology is chosen. But I am curious how others here approach this in practice.