Person comparing papers beside a laptop, with a train and aircraft outside the window.
Illustrative travel-planning scene, not a real trip record or inventory result.
Diagram text description

A synthetic scene shows papers, a ticket-like document and a laptop with coloured routes. The visible train and aircraft suggest several travel modes; no readable distance, emissions amount or company case is established.

A defensible business travel emissions inventory starts with the journeys employees actually made, not the number of rows in a booking export. Classify each activity, match its evidence, apply a compatible method and retain the decisions behind the total.

The recurring problem is duplication and incomplete coverage at the same time. An agency ticket may appear again in an expense claim, while a directly booked train is absent from the agency report altogether. Adding every row overcounts; keeping only one channel can omit real travel.

Build a trip-level register with separate transport legs and document references. For third-party employee business transport, GHG Protocol places the transport providers' relevant Scope 1 and Scope 2 emissions in the employer's Scope 3 Category 6 account. Controlled vehicles and ordinary commuting need different treatment. The operating workflow below connects those boundaries to method choice, duplicate resolution and monthly review. A deliberately fictional calculation demonstrates the arithmetic without supplying real emissions factors. The result should be an explainable accepted-activity set and a visible exception list, not merely a plausible-looking headline number.

Start with the trip, not the expense total

Use a parent trip record for the business purpose and a separate activity record for each transport leg or included stay. A booking reference identifies a commercial transaction; it does not necessarily identify one journey. One booking can contain several passengers and connections, while one travelled leg can appear in the agency export, corporate card feed and reimbursement system.

Decide what the counted row represents before importing everything. A useful transport row represents an identified leg on a specified date, with its mode and traveller coverage. Keep source-system references alongside that row rather than creating another emissions activity whenever a new document arrives.

  • Retain the business purpose, travel dates and a suitable traveller or group reference.
  • Identify the origin, destination and leg, including an evidence-based distance where available.
  • Record the mode and relevant service detail, such as cabin or vehicle type when the method needs it.
  • Retain booking, ticket, card and expense references, their status and their relationship to the activity.
  • Identify included hotel stays separately, with the night basis required by the selected factor.

Only collect personal information needed for the inventory and its controls. A restricted traveller reference can connect records without making passport or payment details part of the calculation file. Keep access to identifiable supporting documents proportionate to the task.

The purpose of this structure is continuity. A corrected expense description can improve the evidence for an existing leg without creating a second journey, and a replacement ticket can be distinguished from an additional flight. Preserve the earlier source reference so someone reviewing the period can understand the correction.

Which travel belongs in the boundary?

Classify the activity using both its purpose and the reporting company's relationship to the vehicle. An expense label saying “business travel” is not enough to determine its GHG account. The same finance code can contain an airline ticket, company-car fuel and an ordinary commuting reimbursement.

ActivityAccounting route to assess
Employee business transport in third-party vehiclesScope 3 Category 6
Fuel or electricity for vehicles owned or controlled by the companyScope 1 fuel or Scope 2 electricity, as applicable
Ordinary home-to-work travelScope 3 Category 7 employee commuting
Leased vehicles operated by the company and not included in Scope 1 or 2Assess Scope 3 Category 8 upstream leased assets

Table 1, shown one row at a time. Every field is retained.

Row 1

Activity
Employee business transport in third-party vehicles
Accounting route to assess
Scope 3 Category 6

Row 2

Activity
Fuel or electricity for vehicles owned or controlled by the company
Accounting route to assess
Scope 1 fuel or Scope 2 electricity, as applicable

Row 3

Activity
Ordinary home-to-work travel
Accounting route to assess
Scope 3 Category 7 employee commuting

Row 4

Activity
Leased vehicles operated by the company and not included in Scope 1 or 2
Accounting route to assess
Assess Scope 3 Category 8 upstream leased assets

The table reflects the distinctions in GHG Protocol's Category 6 guidance. Apply the organisation's consolidation and leasing treatment rather than deciding from the word “rental” alone. An employee's business journey in their own car can differ from their ordinary commute, even if both claims arrive through the same reimbursement system.

Give misclassified rows a destination, not just a deletion instruction. A controlled-fleet entry removed from the travel total should be identifiable in the relevant fleet accounting process. This prevents a correction to Category 6 from becoming an omission elsewhere. Keep a separate unresolved classification state for cases needing further facts; neither a miscellaneous code nor an unknown answer establishes zero emissions.

Capture flights, rail, road and optional hotel nights

Collect information by mode, because each mode's usable activity record differs. A single column called “distance” becomes ambiguous when it mixes passenger journeys, vehicle journeys and hotel stays. Preserve the unit and the count basis with the value.

  • For flights, retain the actual flown segments and relevant cabin information if the selected dataset distinguishes cabin classes. A connecting itinerary is not simply its first origin and final destination.
  • For rail and bus, retain the journey and passenger coverage, route or distance basis, and any service detail needed by the factor.
  • For taxis, rental cars and employee-owned cars used on business, establish whether the evidence describes one vehicle journey or each passenger's journey. Do not turn several occupants into several vehicles.
  • For hotel stays included in the inventory, retain location, dates and the room-night or guest-night basis required by the chosen dataset. Match that basis rather than treating all “nights” as interchangeable.

Hotel inclusion is optional in the cited Category 6 guidance. Decide and disclose the boundary consistently; silently adding hotels in a later period changes the meaning of the total. If they are excluded, retain enough scope documentation to explain the exclusion without implying that stays have no impact.

Do not let good agency coverage conceal other channels. Direct bookings, local ground transport and employee claims can sit outside the main travel export. The collection plan should identify these channels and an owner for their missing activity. That is a completeness question before it is a calculation problem.

Choose fuel, distance or spend without mixing units

Choose the calculation method from evidence that describes the activity, then select a compatible factor. GHG Protocol's Category 6 guidance describes fuel-based, distance-based and spend-based methods. They can serve different parts of an inventory, provided their boundaries and overlaps are controlled.

MethodActivity and factor denominatorKey limitation to resolve
Fuel-basedFuel quantity or allocated energy with the corresponding fuel or energy factorWhole-provider consumption needs an appropriate allocation to the reporting company's travel
Distance-basedVehicle-distance or passenger-distance with a matching factorPassenger and vehicle bases are not interchangeable
Spend-basedEligible travel spend with a compatible currency and price-year factorPrice changes and booking fees can change the proxy without changing distance

Table 2, shown one row at a time. Every field is retained.

Row 1

Method
Fuel-based
Activity and factor denominator
Fuel quantity or allocated energy with the corresponding fuel or energy factor
Key limitation to resolve
Whole-provider consumption needs an appropriate allocation to the reporting company's travel

Row 2

Method
Distance-based
Activity and factor denominator
Vehicle-distance or passenger-distance with a matching factor
Key limitation to resolve
Passenger and vehicle bases are not interchangeable

Row 3

Method
Spend-based
Activity and factor denominator
Eligible travel spend with a compatible currency and price-year factor
Key limitation to resolve
Price changes and booking fees can change the proxy without changing distance

An airline fuel total is not automatically the fuel attributable to one employer. Equally, a ticket price is not a fuel quantity. Where data cannot support a more specific route, use a justified fallback and identify which rows use it. The fallback should improve coverage without pretending to have information the source does not contain.

Avoid applying distance and spend methods to the same underlying activity and summing both outputs. If a richer record replaces a spend estimate, retain the change history and remove the superseded contribution from aggregation. Different methods can coexist for genuinely different journeys; they must not become parallel estimates both counted for one journey. This makes method selection a row-level decision that can be explained, rather than an undocumented adjustment to the final total.

Turn a distance record into a calculation

For a passenger-distance method, calculate each leg's emissions as leg kilometres multiplied by the number of relevant passengers and a factor expressed per passenger-kilometre. The passenger count must describe that row; do not multiply again if the provider's activity value already includes all travellers. Add the results only after the included legs have been reconciled.

In symbols, E_i = d_i × p_i × f_i. Here d_i is kilometres for leg i, p_i is the included passenger count, and f_i is kg CO2e per passenger-kilometre. Kilometres times passengers produce passenger-kilometres; multiplying by the compatible factor leaves kg CO2e. The inventory total sums E_i across accepted activities, not across every source document.

Activity recordDimensional chainCount rule
Distance for one traveller on a legkm × included passengers = passenger-km; passenger-km × kg CO2e/passenger-km = kg CO2eInclude each relevant passenger once.
Provider total already in passenger-kmpassenger-km × kg CO2e/passenger-km = kg CO2eDo not multiply by passenger count again.
One shared vehicle journey wholly attributable to the employervehicle-km × kg CO2e/vehicle-km = kg CO2eCount the vehicle leg once, not once per employee.

Table 3, shown one row at a time. Every field is retained.

Row 1

Activity record
Distance for one traveller on a leg
Dimensional chain
km × included passengers = passenger-km; passenger-km × kg CO2e/passenger-km = kg CO2e
Count rule
Include each relevant passenger once.

Row 2

Activity record
Provider total already in passenger-km
Dimensional chain
passenger-km × kg CO2e/passenger-km = kg CO2e
Count rule
Do not multiply by passenger count again.

Row 3

Activity record
One shared vehicle journey wholly attributable to the employer
Dimensional chain
vehicle-km × kg CO2e/vehicle-km = kg CO2e
Count rule
Count the vehicle leg once, not once per employee.

A vehicle-distance factor requires a different activity basis: vehicle-kilometres multiplied by kg CO2e per vehicle-kilometre. When several employees share one vehicle on a business journey wholly attributable to the employer, count the vehicle journey once. If allocation between reporting parties is needed, document the basis rather than multiplying the entire vehicle result by every occupant. An assumed average occupancy embedded in a passenger factor is not a second instruction to divide the output again.

Represent outward and return travel explicitly. A return ticket may already contain both legs, while a mileage claim may describe a round trip in one distance field. Confirm that convention before doubling it. Include separately evidenced transfers where they are within scope; do not invent a taxi leg merely because one seems plausible.

Hotel activity, if included, needs its own compatible night-based calculation. Add its resulting emissions to transport emissions only with the boundary made clear. A trip total is then a sum of compatible outputs from distinct activities, not a sum of kilometres, fuel quantities, ticket costs and nights.

Check flight routes, cabin and non-CO2 conventions

Use the itinerary actually supported by the records. A flight with a connection has separate flown segments even when finance stores only the overall city pair. The travel record should preserve those segments or identify the route estimate and its basis. Do not silently substitute a direct flight because its distance is easier to obtain.

Match cabin information to the selected dataset where it differentiates classes. If cabin is missing, state how the default was chosen and flag the affected activity. A default is an assumption, not recovered evidence. If a booking was changed, establish which itinerary was travelled before comparing its class or distance with the original reservation.

Also distinguish greenhouse-gas reporting from any additional aviation climate-effect convention. Some methods apply a non-CO2 or radiative-forcing adjustment; that is not a universal constant supplied by this guide. Retain the chosen dataset's treatment and disclose the adjustment if used. Check whether it is already incorporated so it is not applied twice.

These controls matter for an employer inventory even when nobody is choosing between rail and air. Two providers can return different estimates for the same ticket because their route, class or climate-effect conventions differ. Preserve the reason for the selected treatment instead of averaging the outputs as though they were independent measurements of the same quantity.

Use the right factor year, geography and unit

A factor needs an identity as well as a number. Save enough information to reproduce why it was matched to the row, including the publisher's instructions. A web page called “latest factors” is not a durable version reference because the download can change later.

  • Record the publisher, dataset release and the exact table or factor identity.
  • Record the relevant geography, mode and vehicle or service classification.
  • Preserve the denominator: passenger-kilometre, vehicle-kilometre, passenger-mile, fuel unit, spend unit or the specified hotel-night unit.
  • Identify the output basis and included gases or climate-effect conventions.
  • Preserve the activity conversions and any approved proxy assumptions alongside the original source value.

EPA's factors hub provides current and archived releases, illustrating the need to retain a particular release rather than a mutable page label. The appropriate dataset for a journey still depends on the organisation's method and reporting circumstances. A factor published in a particular year may describe underlying data from another period; keep those dates distinguishable.

Convert activity units deliberately when needed. Converting miles to kilometres does not convert passenger-distance into vehicle-distance, nor does converting a currency establish a valid spend-factor price year. If a provider supplies an already calculated emissions result, obtain its method and coverage before accepting it in place of a new calculation. Do not apply a factor to that result again.

Prevent duplicate agency, booking and expense records

Reconcile evidence around the underlying activity, not around amounts that happen to match. An agency ticket, card settlement and employee expense attachment can all describe the same travel leg. Keep those references, but give only the accepted activity one contribution to the inventory total.

Use an explicit decision trail:

Source situationReconciliation decisionWhat remains visible
Agency and expense rows match the same travelled legCount the leg onceBoth references and the matched identity
Replacement booking supersedes a cancelled bookingDetermine the travelled itinerary before countingCancellation and replacement relationship
Expense has no matching agency recordInvestigate as possible additional travelUnmatched status, owner and evidence needed

Table 4, shown one row at a time. Every field is retained.

Row 1

Source situation
Agency and expense rows match the same travelled leg
Reconciliation decision
Count the leg once
What remains visible
Both references and the matched identity

Row 2

Source situation
Replacement booking supersedes a cancelled booking
Reconciliation decision
Determine the travelled itinerary before counting
What remains visible
Cancellation and replacement relationship

Row 3

Source situation
Expense has no matching agency record
Reconciliation decision
Investigate as possible additional travel
What remains visible
Unmatched status, owner and evidence needed

Ticket or segment identifiers provide a useful starting point where available. For records without a shared key, compare traveller or group reference, service date, route, mode and the relevant document details. A matching date and payment amount alone do not prove identity, especially when two employees make similar journeys. An uncertain match should remain an exception rather than being automatically discarded.

Distinguish charges from physical travel. A refund can correct the financial record without representing negative kilometres, and a cancellation fee does not prove a journey occurred. Preserve financial reconciliation separately from the accepted travel activity and its calculation. Where a spend-based method is used, apply that method's defined eligible-spend treatment rather than quietly borrowing a distance-method assumption.

The consolidation figure applies only after the records have been matched. It does not say that every agency, card and expense row belongs to the same leg. Retaining both the counted activity and the excluded duplicate references lets a reviewer explain why several financial rows produced one emissions contribution without losing evidence of genuinely unbooked-channel travel.

Matched agency, card and expense records converge on one counted travel leg.
Only records confirmed to describe the same activity consolidate into one counted leg.
Diagram text description

Three source references point to one activity identity. Their arrows mean supporting evidence, not additional passenger journeys. An unmatched or disputed record remains outside this consolidation until investigated; the picture does not imply that all travel documents describe the same leg.

A fictional two-leg calculation

This example is entirely fictional. Two employees take one invented rail trip with an outward leg of 180 km each and a return leg of 240 km each. The routes are not real locations. The coefficient 0.04 kg CO2e per passenger-kilometre is invented solely to demonstrate arithmetic; it is not a published rail factor and must not be used in real reporting.

Assume both employees travel both legs. The distances describe each person's leg, not a provider total already multiplied by passenger count. No hotels, transfers or additional climate-effect adjustments are included. Agency and expense records have been matched to the same two group-leg activity rows, so they do not create additional journeys.

Fictional group-leg rowInvented passenger activityResult using invented coefficient
T-EX-73-L1 outward180 km × 2 passengers = 360 passenger-km360 × 0.04 = 14.4 kg CO2e
T-EX-73-L2 return240 km × 2 passengers = 480 passenger-km480 × 0.04 = 19.2 kg CO2e
Trip total for both employees840 passenger-km14.4 + 19.2 = 33.6 kg CO2e

Table 5, shown one row at a time. Every field is retained.

Row 1

Fictional group-leg row
T-EX-73-L1 outward
Invented passenger activity
180 km × 2 passengers = 360 passenger-km
Result using invented coefficient
360 × 0.04 = 14.4 kg CO2e

Row 2

Fictional group-leg row
T-EX-73-L2 return
Invented passenger activity
240 km × 2 passengers = 480 passenger-km
Result using invented coefficient
480 × 0.04 = 19.2 kg CO2e

Row 3

Fictional group-leg row
Trip total for both employees
Invented passenger activity
840 passenger-km
Result using invented coefficient
14.4 + 19.2 = 33.6 kg CO2e

The same fictional answer follows from 840 passenger-km × 0.04 kg CO2e/passenger-km. It is a total for both employees together, not a per-person value to multiply by two again. Unequal outward and return distances are explicit scenario assumptions, not evidence that a real route takes those distances.

If the expense export repeats the outward ticket, link that document to T-EX-73-L1 rather than creating another 14.4 kg contribution. If an unmatched receipt might describe an additional leg, investigate it; do not force it into this example's total. Real use requires actual supported activity and a suitable sourced factor. Correct arithmetic can confirm the calculation mechanics, but cannot validate an invented coefficient or an unresolved travel record.

Explain uncertainty before comparing periods

Explain the limitations of the register before interpreting changes in its total. Coverage, estimation and factor choices can change independently of how much employees travel. A precise output with missing journeys is not necessarily more informative than an explicitly qualified estimate.

For example, an employer might capture agency flights in one year and add direct-booked rail and taxis in the next. A higher total can partly reflect better channel coverage. Compare the consistently covered population where appropriate, while also showing the expanded inventory. Do not remove the newly visible journeys simply to preserve a favourable trend.

Identify which activity uses estimated routes, unknown cabin defaults or spend proxies, and which channels remain incomplete. Explain whether optional hotels have changed scope. Where practical, describe the affected activity or data share; do not manufacture a percentage when the denominator is unknown. Changes in prices can alter a spend-based estimate without demonstrating changed travel activity.

Keep physical activity changes distinguishable from methodological changes. A different factor release, corrected duplicate or improved allocation can change the result even when an itinerary is unchanged. Follow the organisation's established policy for any historical recalculation and retain the earlier version with its reason for change. The conclusion may be that periods are only partly comparable. That is a useful accounting finding, not a reason to disguise uncertainty with a single unexplained reduction claim.

Review the inventory as a repeatable monthly process

A monthly review is a practical operating choice, not a universal GHG Protocol deadline. Choose a cadence that allows travel, finance and the inventory owner to resolve exceptions before period-end pressure turns assumptions into permanent defaults. The same sequence can be used at another appropriate interval.

  1. Preserve the received agency, direct-booking, card and expense extracts with their source periods and file versions. Check whether a feed covers bookings, settlements or travel actually undertaken.
  2. Create or update trip and leg identities. Classify fleet, commuting, leases and optional stays using the chosen boundary rather than a finance label alone.
  3. Match overlapping evidence and retain counted, duplicate, cancelled, replaced and unresolved states. Assign ambiguous matches to an owner rather than forcing the import to balance.
  4. Resolve method inputs and factor selections, keeping missing routes, unknown classes and spend fallbacks visible. Confirm that a richer record replaces rather than supplements an older estimate for the same activity.
  5. Calculate accepted rows, check units and recompute a sample including shared journeys, return legs and a corrected booking. Reconcile the aggregation to the accepted activity set.
  6. Have the responsible inventory owner review coverage and unresolved exceptions, then preserve the input, calculation and factor versions supporting the period result.

Keep a correction route for late claims and changed itineraries. A new source file should trigger a review of affected rows, not an unexplained overwrite of the signed-off total. Mark whether the correction changes identity, activity, classification, method or only supporting evidence. That distinction makes later comparisons and investigations much faster than reopening every booking.

The monthly output should therefore include the accepted activity inventory and a separate exception view. A total without either cannot show whether missing travel was investigated or whether duplicates were removed for a defensible reason.

Test the data tool against a difficult trip

Evaluate a prospective data tool using the inventory's hardest reconciliation case, not a polished total from a clean demonstration file. Ask it to connect an agency booking and an expense receipt to the same leg, then replace the booked itinerary with the travelled one while preserving the earlier evidence.

The useful acceptance questions concern identity, method and correction history. Can the team see which activity is counted and which rows support it? Can an unresolved match remain visible without being counted twice? Can a factor change be distinguished from a distance correction? Can another analyst reproduce the period result from the retained input and method versions?

These are requirements to test, not promised capabilities of any named supplier. A carbon-data platform does not automatically provide booking integrations, current travel factors or assurance of a corporate inventory. Define which travel or finance system remains authoritative for the underlying event, and who resolves inconsistent records. A modest controlled register can be a better starting point than a broad integration whose matching decisions nobody can explain.

The central control is one governed activity identity with traceable evidence and a compatible calculation. A second document can strengthen that evidence without becoming a second journey. An unmatched document can reveal missing travel rather than being treated as a nuisance to discard.

Start by testing a small period containing a normal return trip, a shared vehicle journey, a changed booking and a direct expense absent from the agency feed. Check whether another analyst can reconstruct what was counted, why other rows were not counted, and which assumptions remain unresolved. Use that test to improve the collection and reconciliation process before expanding the import volume.

Choose further guidance according to the unresolved decision. A detailed Category 6 boundary guide can help with classification; an air-travel calculator guide can address flight-specific method inputs; a corporate travel-footprint plan can support management choices once the inventory is understood. None replaces the underlying trip record. Keep the fictional coefficient out of operational reporting, and do not interpret a well-reconciled total as proof that every journey or factor assumption is complete.