Industry
Part of What can a company honestly claim about its sustainability ethics?
Sustainability ethics metrics: what to put on the tracking sheet
A sustainability ethics metrics tracking sheet: ten rows worth keeping, four fields to fill in for each, and the rows to leave out.
Sustainability ethics metrics begin as a sheet with a few rows, and one rule shapes every row: almost every environmental figure a company publishes is a calculation rather than a measurement.
So a row holds more than a number. It holds four fields: basis, boundary, baseline, and the absolute figure paired with intensity. If you understand what each field does to a number, you can read anyone's report, including your own.
What to take away
- Basis, boundary and baseline decide most of the number. Two honest companies with identical operations can publish very different figures by choosing differently, and both can defend their choice.
- Intensity figures can improve while absolute figures rise. Keep both on every row or the reader cannot tell which happened.
- A change of methodology and a change of comparison year in the same report, without both being shown, is the most common way a figure misleads without anyone lying.
- Precision is a claim. Three significant figures on a number built from generic factors is asserting an accuracy the inputs do not support.
The ten rows worth keeping
Ten metrics earn a row because each has a defined unit and a published method. The right-hand column gives the concrete fact that anchors the row, so the number can be checked against something outside your own spreadsheet.
Metrics worth tracking
| Metric worth tracking | What it is | Concrete fact |
|---|---|---|
| Scope 1 greenhouse gas emissions | Direct emissions from owned or controlled sources, such as natural gas boilers and company vehicles | Defined by the GHG Protocol Corporate Standard; report in metric tons of carbon dioxide equivalent (tCO2e) |
| Scope 2 location-based emissions | Purchased electricity, heat, steam or cooling, using the average grid intensity | The U.S. EPA eGRID database publishes regional emission factors |
| Scope 2 market-based emissions | Same electricity, but using contractual instruments such as renewable energy certificates (RECs) | The GHG Protocol Scope 2 Guidance requires both location-based and market-based figures when instruments are bought |
| Scope 3 Category 1 purchased goods and services | Upstream emissions from what the company buys | One of 15 categories in the GHG Protocol Scope 3 Standard; often the largest source for manufacturers and retailers |
| Energy intensity per unit of production | Kilowatt-hours per ton, per barrel, or per vehicle | Compare efficiency across a growing business; report it beside the absolute total |
| Water withdrawal in water-stressed basins | Cubic meters withdrawn, split by basin stress level | The WRI Aqueduct tool maps baseline water stress; CDP's water security questionnaire asks for this split |
| Waste diversion rate | Percentage of waste diverted from landfill and incineration | TRUE certification requires at least 90% diversion from landfill, incineration, and waste-to-energy for 12 months |
| Lost-time injury frequency rate (LTIFR) | Recordable injuries per 1,000,000 hours worked | OSHA Form 300 logs recordable incidents; pair LTIFR with days away to show severity |
| Gender pay gap | Median hourly pay difference between men and women | UK employers with 250 or more employees must publish it; the U.S. EEOC collects pay data on the EEO-1 Component 1 form |
| Supplier audit coverage | Share of suppliers or spend covered by a social or environmental audit | Sedex, SMETA, and the Responsible Business Alliance's VAP are common; a count of audits without coverage percentage tells you little |
For the location-based row, the factors come from sources such as EPA's eGRID database, a source of data on the environmental characteristics of almost all electric power generated in the United States, including emissions and emission rates. Record which edition of the factors you used.
The four fields every row needs
Fill four fields before a row goes into any report. Leave one blank and nobody can test the figure, including you in a year's time.
| Field | Question to answer | What it prevents |
|---|---|---|
| Basis | Measured, estimated, modeled, or reported by someone else? | Comparing figures that are not comparable |
| Boundary | Which sites, entities and activities sit inside the count? | A total that moves because the line moved |
| Baseline | Which comparison year, and would you use it if it made the number worse? | A reduction that required nothing |
| Pair | Is the absolute figure shown beside intensity? | Efficiency gains hiding a rising total |
Field 1: basis, where the number came from
Four kinds of basis exist, and they are not comparable. Write one of them beside every figure. It costs one word per line and changes how the entire sheet reads.
Four bases, not comparable
Measured
- Source
- Own meter
- Data quality
- Honest
- Best use
- Direct energy
- Tracking change
- Reliable
Estimated
- Source
- Activity x factor
- Data quality
- Generic factor
- Best use
- Rough totals
- Tracking change
- Cautious
Modeled
- Source
- Spend x factor
- Data quality
- Weak for change
- Best use
- Spotting big items
- Tracking change
- Poor
Reported
- Source
- Supplier
- Data quality
- Unseen
- Best use
- Supply chain
- Tracking change
- Unknown
- Measured:a meter on something you own. Rare beyond direct energy use at your own sites, and the only category where the word measurement is doing honest work.
- Estimated:activity data multiplied by a conversion factor. The activity data may be good. The factor is generic and was derived from something that is not your operation.
- Modeled:a spend figure multiplied by an intensity factor, or a physical model of a process. Useful for spotting where the big items are, weak for tracking change, because the model does not know what you changed.
- Reported by someone else:a supplier's declaration. It carries whatever quality their own basis had, and you usually cannot see it.
Field 2: boundary, what is inside the count
The boundary decides more than most people expect. It is usually settled by convention rather than by anyone's deliberate choice, so write it down on the row.
Boundary honesty rules
- Write the boundary beside the number
- Hold the boundary steady across periods
- If changed, restate prior year on new boundary
- Show both old and restated figures
- An unstated boundary cannot be tested
Does the figure include leased sites? Franchised operations? The joint venture you do not control? Contractor vehicles? The product's use by customers? Its disposal?
Every one is defensible either way. Moving a boundary can change a total by more than a decade of operational improvement. Two rules keep boundaries honest: write the boundary beside the number, and hold it steady across periods.
When you must change it, restate the prior year using the new boundary and show both. A written boundary is also the first item on a corporate responsibility checklist, because an unstated boundary cannot be tested.
Field 3: baseline, what you compare against
A reduction is a comparison, and the choice of comparison year is a choice about how impressive the reduction is. Name the baseline year on the row and the reason for it.
Is this baseline honest?
Would you use this baseline if it made the number worse?
defensible comparison
marketing artifact
Baselines chosen after the data is known are a marketing artifact. Baselines that happen to be an unusually high year produce reductions that required nothing. Baselines that predate a divestment credit you with someone else's decision.
The test applies to any comparison you chose: would you have used this baseline if it had made the number worse? Write the answer down before you look.
Field 4: absolute beside intensity
Intensity figures, per unit produced or per unit of revenue, can fall steadily while the absolute total rises, if the business grows faster than the efficiency improves. Every row therefore carries both.
Absolute vs intensity
Absolute
- Unit
- Total tonnes
- Can fall while
- Output shrinks
- Growth effect
- Direct
- Reporting rule
- Publish both
Intensity
- Unit
- Per unit or revenue
- Can fall while
- Total rises
- Growth effect
- Hidden
- Reporting rule
- Publish both
Neither figure is dishonest. Reporting only the one that moved in your favor is. Where they point in different directions, say so in a sentence rather than leaving the reader to notice.
The related trap is a target set on intensity and described as if it were absolute. Those are different commitments. The difference is the entire growth rate of the business.
The offsets row
An offset is a transaction, not a reduction in your operation. Whatever its merits, netting it into the same figure as your own emissions makes the resulting number uninterpretable.
Gross, offsets, net
- Gross emissions
- Offsets purchased
- Net figure
- Reader forms own view
Give gross, offsets and net three separate cells. A reader can then form their own view about the offsets, which they cannot do if the three have been collapsed into one. It is the same discipline a corporate responsibility framework applies to every obligation you carry.
Rows to leave out
Ten common rows add little or mislead. Treat them as caution flags rather than bans, and ask what each one hides before it earns a place on the sheet.
Metrics to ignore
| Metric to ignore or treat with caution | Why |
|---|---|
| Total carbon footprint as one number | No scope breakdown; Scope 3 categories can double count |
| Carbon neutral claim from offsets only | Offsets are transactions, not reductions; Verra and Gold Standard project integrity varies |
| Unbundled renewable energy certificates as reductions | Market-based only; the report must also show location-based emissions |
| Tree planting counts | Survival rates are often low; no common monitoring standard |
| Paper saved or paperless claims | Small material impact; distracts from supply chain emissions |
| Employee volunteer hours | No standard; not comparable across companies; can be gamed |
| One-off office energy savings | Not material to a total figure |
| Supplier questionnaire scores without verification | Self-reported; not the same as an audit |
| Zero waste to landfill excluding incineration | The FTC Green Guides caution against broad claims; incineration is not diversion |
| Intensity-only targets presented as absolute | Business growth can outpace efficiency gains |
On claims, the FTC's Green Guides say marketers should not make unqualified general environmental benefit claims. Where figures become marketing claims, some markets also apply rules on environmental advertising.
Four checks before you log a change
Before concluding that a movement means anything about the world, work through four questions in order.
Four checks before reporting
- Did the factor change?Conversion factors are updated periodically. A published update can move your whole series without anything happening at your sites.
- Did the boundary change?An acquisition, a divestment, a reclassified site.
- Did the basis change?A figure that moved from modeled to measured usually moves, and the movement is about the method.
- Only then: did the operation change?
Most commentary on environmental figures starts at four. Starting at one is slower and produces conclusions that survive the next reporting cycle.
What to publish from the sheet
Short and evidenced beats comprehensive and defensible-only-with-explanation. State the basis, boundary and baseline for each figure. Show uncertainty on estimates rather than a single number carried to three decimals. Report what got worse.
What can honestly be claimed from the sheet at all is the subject of sustainability ethics.
Common questions
If most of it is estimated, is it worth measuring?
Yes, for finding the big items and for tracking things you actually control. What it does not support is fine-grained comparison between companies, or confident claims about small year-on-year movements.
Should we set a target on any of this?
Set it on something you control and would still want if it were gamed. Targets on figures dominated by factors and boundaries reward methodology work, which is cheaper than operational change and produces the same graph.
A supplier gives us a figure we cannot verify. What do we do with it?
Use it, mark it as reported by someone else, and do not present it with the same confidence as your own metered data. Where it matters a lot, ask for the basis and the boundary rather than a better number. The supplier ethics metrics reading steps apply here too.
