.png)
In carbon accounting, real time does not mean the dashboard refreshes every second: it means the latency of the source data. A dashboard that updates instantly with electricity data from three weeks ago is not real time. The useful frequency for an organisation is set by the slowest source it needs to make a decision, not by the software.
This guide ranks data sources by the frequency they genuinely allow, explains what cannot be tracked continuously, and shows how to build monitoring that supports management without breaking the annual inventory.
What does real time mean applied to emissions?
A carbon footprint is calculated by multiplying activity data by an emission factor. In continuous monitoring that means two separate clocks:
- The activity data clock. When you know how many kWh, litres or tonne-kilometres were consumed. It runs from seconds, on your own meter, to a year, for a supplier's footprint.
- The emission factor clock. When you have the correct factor. It can be hourly for the grid mix, or annual and revisable, like Spain's MITECO registry factors, published around May of the following year.
Two practical consequences follow. First, you almost never measure gas at a stack: you estimate from consumption. Second, the figures on a continuous dashboard are provisional and get restated when the final factors arrive. An honest system flags which figures are provisional and which are closed.
What frequency does each data source allow?
| Data source | Realistic frequency | Typical latency | How it is automated |
|---|---|---|---|
| Own sub-metering: meters, IoT sensors and building management systems | Minutes | Seconds to minutes | Direct integration by industrial protocol or API into the platform |
| Electricity at the supply point | Hourly | Several days, until the distributor publishes the curve | API from Spain's Datadis, the retailer or the energy management system |
| Emission factor of the Spanish grid mix | Hourly | Near immediate | Red Eléctrica and e·sios data APIs, which publish generation structure and associated emissions in real time |
| Natural gas | Monthly or bimonthly for most supplies, hourly with telemetering for large consumers | Weeks | Automatic invoice reading or retailer telemetering |
| Own fleet fuel | Per refuelling, daily or weekly | Hours to days | Fuel card file and vehicle telematics |
| Business travel | Weekly | Days | Corporate travel agency and company cards |
| Contracted transport | Per shipment | Days to weeks | Logistics operator portal with ISO 14083 compliant data |
| Waste | Monthly or quarterly | Weeks | Waste transfer documents and authorised manager delivery notes |
| Purchases and materials | Continuous on spend, annual on supplier data | Spend is immediate; the supplier footprint arrives a year later | ERP for spend and a data campaign for the product footprint |
| Supplier Scope 3 | Annual | Months to a year | Structured questionnaire or exchange of verified product footprints |
What cannot be tracked in real time?
It is worth saying plainly, so nobody promises the impossible:
- The Scope 3 categories that depend on third parties. Purchased goods and services, use of sold products, end of life and investments close once a year at best.
- Official emission factors. They update annually, so the current year is calculated with last year's factor and restated afterwards.
- Fugitive refrigerant emissions. You learn about them at recharge, not continuously.
- Market-based Scope 2. It depends on guarantees of origin and contract settlement, not on instantaneous consumption.
The sensible approach is to monitor continuously what can be monitored, with hourly or daily data, and keep the rest on an annual cycle with documented estimates.
How to build continuous monitoring step by step
1. Define the decision the data has to support
Spotting a consumption spike at a plant is not the same as preparing the annual report. The decision sets the latency you need, and that latency sets how much integration work is worth doing. Start with the use case, not the dashboard.
2. Classify each source by achievable latency
Use the table above to tag your sources. Anything that already arrives daily or faster goes into continuous monitoring. The rest stays on a monthly or annual cycle, with the reason documented.
3. Automate capture
Integrate invoices, meters, telematics and ERP so consumption arrives on its own. The less manual entry there is, the more reliable the system and the slower it degrades. Automatic document and invoice reading is usually what unlocks the sources that arrive on paper or as PDFs.
4. Freeze the methodology and version the factors
Record which factor was applied to which data point and on what date. Without that versioning, a mid-year factor change makes the series incomparable and complicates verification.
5. Set indicators and alert thresholds
A dashboard with no thresholds does not get looked at. The indicators that work are intensity based, not absolute: tCO₂e per unit produced, kWh per square metre, litres per hundred kilometres. The useful comparison is against the same period last year, corrected for production or degree days, with an owner assigned to every alert.
6. Close the year with a verified inventory
Continuous monitoring feeds the annual calculation aligned with the main standards, it does not replace it. At year end you apply the final factors, complete the annual categories and submit the inventory for verification where relevant.
What is continuous monitoring actually good for?
- Catching deviations while they still matter. Abnormal weekend consumption gets fixed in days, not at year end.
- Checking whether a measure works. It is the only way to validate the expected saving in a reduction plan before a year has passed.
- Shortening the annual close. If data is collected and traced through the year, the inventory stops being a three-month project.
- Answering customers and investors. Footprint data requests arrive on short notice and with tight deadlines.
It is also the basis of data and AI integration in ESG reporting: without a continuous flow there is nothing to automate.
How it fits with the inventory and certification
Standards still think in years. The GHG Protocol and ISO 14064-1:2018 require a defined reporting period, and the MITECO carbon footprint registry only accepts footprints covering twelve consecutive months calculated with official factors. Continuous monitoring is a management layer on top of that: it adds reaction time, not rigour.
For the dashboard and the inventory to reconcile at year end, the organisational boundaries, the Scope 1, 2 and 3 split and the allocation rules must be identical in both. If the dashboard uses a different definition of a site than the inventory, reconciliation eats more time than the monitoring saves.
Frequently asked questions about real-time carbon footprints
Does a real-time carbon footprint measure gases directly?
In practice, almost never. Emissions are estimated from activity data and emission factors. Direct gas measurement is reserved for specific industrial stacks with continuous emission monitoring systems.
What latency can you realistically achieve?
Minutes with your own sub-metering. A few days with the distributor's hourly curve. Weeks for gas and waste. A year for supplier data. An honest corporate dashboard shows different latencies per source instead of a single headline figure.
Does it replace the verified annual inventory?
No. It complements it. The annual inventory brings rigour, official factors and verification; continuous monitoring brings the ability to react and a shorter close.
Why do dashboard figures change at year end?
Because final emission factors are published after the close and because third-party data replaces earlier estimates. Good practice is to flag which periods are provisional and to log every restatement.
What do I need to get started?
Electricity supply points with access to the hourly curve, the fuel card file if you run a fleet, and an ERP connection for spend. Those three sources already cover much of Scope 1 and 2 at daily frequency.
Does it work for transport and logistics?
Yes, if the operator delivers data per shipment. The GLEC Framework and ISO 14083 define the common format that makes that data comparable across providers.
If you want to move from an annual calculation to continuous monitoring without losing traceability, Manglai's carbon footprint platform automates capture from invoices, ERP and telematics with artificial intelligence features, and keeps the factor versioning a verifier will ask for.
