Transparent system boundaries
State what the calculation includes, excludes, and cannot observe.
Research foundation
Digital emissions need better measurement—and better decisions.
Data Offset is building a transparent, research-informed foundation for activity-level digital sustainability without exposing private content or overstating what an estimate can prove.
The scientific problem
Digital activity is observable. Its climate consequence is usually hidden.
Infrastructure providers report pieces of the system. Applications expose different telemetry. Electricity mixes vary by location and time. A useful estimate must preserve those boundaries, document its assumptions, and resist false precision.
Research principles
What makes an estimate worthy of action?
Source-aware inputs
Separate provider-reported activity, composite assumptions, projections, and model-derived values.
Uncertainty, not false precision
Carry ranges and fallback conditions into the result instead of burying them.
Behavior in context
Offer a lighter alternative only when it respects the purpose and quality of the activity.
Net climate benefit
Evaluate whether a recommendation’s avoided impact exceeds the cost of delivering it.
Reproducible evidence
Keep the source, factor version, and calculation path attached to every result.
Public research boundary
Credibility does not require exposing experimental design.
This public prototype explains the scientific problem, measurement chain, source classes, uncertainty, privacy boundary, and net-climate-benefit principle.
It intentionally does not publish hypotheses, cohorts, interventions, statistical plans, participant protocols, model portfolios, or proprietary calibration methods.
The standard