Chapter 5
Activity-Based Costing and Time-Driven ABC
Post Graduate Program in Logistics
Logistics Costing & Budgeting
•
Post Graduate
Activity-Based Costing and Time-Driven ABC
Introduction and context
Logistics costs are often reported by department, facility, vehicle fleet, or accounting ledger. Such reports are useful for financial control, but they may conceal the economic causes of cost. Two customers may generate the same sales revenue while consuming very different amounts of order entry, warehouse handling, transport planning, returns processing, and customer service. A conventional allocation based on shipment value or total units can therefore distort customer profitability, product profitability, and outsourcing decisions.
Activity-Based Costing (ABC) addresses this problem by assigning resource costs to activities and then assigning activity costs to cost objects through cost drivers. Time-Driven Activity-Based Costing (TDABC) simplifies and extends ABC by estimating the cost of supplied capacity per unit of time and multiplying it by the time required for each transaction or order.
In logistics, ABC and TDABC are particularly valuable because operations contain diverse transactions, substantial indirect labour, shared assets, variable service requirements, and measurable process times. The methods are not merely accounting techniques. Properly designed, they support process redesign, service pricing, customer segmentation, capacity planning, make-or-buy analysis, and continuous improvement.
Learning objectives
After studying this chapter, the reader should be able to:
Explain the logic and terminology of ABC and TDABC.
Identify logistics activities, resources, cost objects, and cost drivers.
Construct an ABC model and calculate activity cost-driver rates.
Estimate practical capacity and unused capacity.
Build a TDABC model using time equations.
Interpret cost information for pricing, outsourcing, and process improvement.
Evaluate the limitations, governance requirements, and implementation risks of these methods.
1. Foundations: resources, activities and cost objects
A resource is an economic input consumed by operations, such as warehouse labour, material-handling equipment, information systems, premises, fuel, or supervisory effort. An activity is a work process performed using resources. Examples include receiving, put-away, picking, packing, route planning, delivery, proof-of-delivery administration, and returns inspection.
A cost object is the item for which management wants to know cost. It may be a customer, shipment, order, SKU, route, service level, distribution centre, or contract. A resource driver explains how resource costs are assigned to activities. For example, warehouse labour may be assigned to receiving, picking, and packing according to labour time. An activity cost driver explains how activity costs are assigned to cost objects. Examples include receipts processed, order lines picked, cartons packed, delivery stops, kilometres, or return cases.
The causal sequence is:
ABC causal sequence from resources to decisions
Figure: ABC causal sequence from resources to decisions
Figure: The causal logic linking resources, activities, cost objects, and managerial decisions.
The quality of an ABC model depends on causal validity. A driver should be selected because changes in the driver plausibly cause changes in activity resource consumption—not simply because the driver is convenient to measure.
2. Conventional costing and the case for ABC
Traditional absorption costing commonly applies indirect cost using a broad rate, such as cost per pallet, cost per kilogram, or a percentage of sales. This may be satisfactory where products and customers consume support resources in similar proportions. It becomes problematic when logistics complexity differs substantially.
For example, a full-truckload customer may require one planned dispatch and one delivery. A small customer may require many order lines, special labels, delivery appointments, partial shipments, and returns. Allocating all distribution overhead by tonnes would make the first customer appear too expensive and the second too profitable.
ABC uses multiple pools to reflect heterogeneous work. Its basic logic is:
These formulas assume that the activity pool and driver quantity refer to the same period, scope, and unit definition. If a driver is measured in order lines, the pool must represent resources used for the relevant order-line activity, not total warehouse cost unless the model explicitly justifies that aggregation.
Activity hierarchy
A useful classification is:
Level
Meaning
Logistics example
Unit-level
Per unit handled
Picking one carton
Batch-level
Per batch or order
Setting up a wave
Product/customer-level
Supports a product or account
Maintaining a customer-specific label format
Facility-level
Supports the whole operation
Warehouse management and rent
Unit- and batch-level costs are usually easier to assign causally. Facility-level costs may not have a meaningful short-term driver and should sometimes be reported separately rather than forced into customer costs.
3. Designing an ABC model
Step 1: Define the decision purpose and boundary
The model should begin with a decision, such as identifying unprofitable delivery patterns, pricing premium services, or comparing internal warehousing with a third-party logistics provider. Define whether the model covers one warehouse, a network, a contract, or an entire supply chain. A model designed for strategic network design may include depreciation and facility costs; a short-term pricing model may focus on avoidable and incremental costs.
Step 2: Identify activities
Activities should be sufficiently detailed to reveal meaningful variation but not so detailed that data collection becomes unmanageable. A practical activity dictionary records the activity name, start and end points, responsible department, outputs, resources consumed, frequency, and proposed driver.
Step 3: Assign resource costs to activity pools
General ledger costs are traced directly where possible. Shared costs are assigned using resource drivers. For example, warehouse supervisors may be assigned using observed supervision time; equipment depreciation may be assigned using equipment hours; information-system costs may be assigned using transaction volume or licences.
Step 4: Select cost drivers
A strong driver has three properties:
Causality: it explains resource consumption.
Measurability: reliable data can be collected.
Behavioural acceptability: employees understand and cannot easily manipulate it.
Transaction drivers are simple but assume similar resource consumption per transaction. Duration drivers are more accurate where transaction complexity varies. A cost driver should not be selected solely because it is available in an enterprise system.
Step 5: Calculate rates and assign costs
Rates are calculated for each activity pool. Costs are then assigned to customers, products, orders, or routes. The model should reconcile assigned costs to the activity pools and ultimately to the costs included in the model boundary.
4. Worked ABC example: distribution-centre costing
The following is an illustrative example. A distribution centre has the following annual activity pools:
Activity
Annual pool
Driver volume
Receiving
$180,000
12,000 receipts
Picking
$420,000
210,000 order lines
Packing
$240,000
80,000 parcels
Dispatch administration
$160,000
16,000 shipments
The rates are:
Receiving rate = $180,000 ÷ 12,000 = $15 per receipt
Picking rate = $420,000 ÷ 210,000 = $2 per order line
Packing rate = $240,000 ÷ 80,000 = $3 per parcel
Dispatch rate = $160,000 ÷ 16,000 = $10 per shipment
Customer X generates, during the year, 40 receipts, 6,000 order lines, 2,400 parcels, and 300 shipments. Its assigned cost is:
Activity
Quantity
Rate
Assigned cost
Receiving
40
$15
$600
Picking
6,000
$2
$12,000
Packing
2,400
$3
$7,200
Dispatch administration
300
$10
$3,000
Total
$22,800
This figure is not necessarily the customer’s total economic cost. Facility-level costs excluded from the model remain unassigned, and transport costs may require a separate activity pool. The interpretation must therefore state the model boundary.
5. Time-Driven Activity-Based Costing
Traditional ABC can become expensive when many activities and drivers are required. TDABC uses two principal estimates:
the cost per unit of practical capacity supplied; and
the time required to perform one transaction or service.
If supplied capacity is measured in minutes, the rate is currency per practical minute. The cost of a transaction is:
Transaction cost = Capacity cost rate × Time required for transaction
For multiple resource groups:
Transaction cost = ∑(Capacity cost ratei × Timei)
Here, subscript i identifies a resource group such as receiving labour, picking labour, or customer service. The time estimate must use the same unit as the capacity denominator.
Practical capacity
Theoretical capacity assumes continuous productive work and is unrealistic. Practical capacity deducts normal allowances for breaks, meetings, training, maintenance, absenteeism, equipment downtime, and unavoidable interruptions.
A simple estimate is:
Practical capacity = Theoretical working time × (1 − Normal non-working proportion)
For example, a worker scheduled for 2,080 hours per year with a normal non-working proportion of 20% provides:
2,080 × (1 − 0.20) = 1,664 hours
This is an estimate, not a universal standard. Capacity should be validated using observed operations and local labour arrangements.
Idle capacity
If supplied practical capacity is 100,000 minutes but actual time used by transactions is 82,000 minutes, unused capacity is:
Unused capacity = Supplied practical capacity − Used capacity
Thus, unused capacity is 18%. It should normally be reported transparently rather than automatically loaded onto transactions. Assigning idle capacity to customers can make existing customers appear responsible for resources they did not consume.
6. Time equations and operational variation
A central advantage of TDABC is the time equation, which represents how transaction time changes with observable attributes. A general form is:
t = β₀ + β₁x₁ + β₂x₂ + ⋯ + βₙxₙ
where t is estimated time, β₀ is base time, βᵢ is the incremental time associated with condition xᵢ, and xᵢ is normally a count or binary indicator. Conditions should be mutually defined so that time is not double-counted.
Suppose a picking task requires 4 minutes per order line, plus 6 minutes when the order is urgent, plus 3 minutes when a special label is required. Let urgent and special-label indicators equal 1 when present and 0 otherwise:
t = 4L + 6U + 3S
For an order with 5 lines, urgent status, and a special label:
t = (4 × 5) + (6 × 1) + (3 × 1) = 29 minutes
The equation is applicable only within the process conditions from which its parameters were estimated. If very large orders require a different picking method, a single linear equation may become inaccurate.
TDABC worked example
A customer-service team costs $312,000 annually. Its practical capacity is 156,000 minutes, giving:
where 8 is base processing time, E is the number of external contacts, and C equals 1 if carrier escalation is required and 0 otherwise. A case involving two external contacts and carrier escalation requires:
t = 8 + (5 × 2) + (12 × 1) = 30 minutes
Its assigned cost is:
$2 per minute × 30 minutes = $60
If 4,000 such cases are processed, they consume 120,000 minutes and cost $240,000. The remaining 36,000 minutes represent unused capacity or capacity used by other activities, depending on the model boundary.
7. ABC versus TDABC
Issue
Traditional ABC
TDABC
Main driver
Activity transaction or duration driver
Time equation and capacity rate
Data burden
Can require many activity pools
Often fewer estimates and equations
Variation
Requires additional pools or drivers
Captured through transaction attributes
Idle capacity
Often hidden in rates
Explicitly visible through practical capacity
Updating
May require repeated surveys and reallocations
Update time equations or capacity rates
Risk
Over-complexity and subjective allocations
Inaccurate time estimates or omitted conditions
Neither method is automatically superior. ABC may be appropriate where activities and transaction counts are stable and easily measured. TDABC is attractive where transaction complexity changes frequently, process times are observable, and managers need capacity information.
8. Using results for managerial decisions
Process improvement
ABC identifies activities that consume resources; TDABC identifies the time consumed by process variants. Managers can ask why a high-cost activity exists, whether it adds customer value, and whether its frequency can be reduced. Typical interventions include standardising packaging, improving master data, reducing order fragmentation, redesigning delivery windows, and integrating electronic proof of delivery.
Cost information should be combined with quality and service metrics. Eliminating a low-cost activity that prevents major errors may reduce apparent cost while increasing total cost of failure.
Pricing and customer profitability
A logistics service price should reflect the resources required by service attributes such as urgency, low minimum order quantities, special handling, narrow delivery windows, and returns. ABC/TDABC can support a differentiated price schedule. However, assigned cost is not the same as a recommended price: price also depends on willingness to pay, competition, capacity constraints, strategic account value, and avoidable cost.
Customer profitability is commonly represented as:
The term “relevant” is important. Long-run strategic profitability may include allocated facility costs, whereas a short-run acceptance decision may consider only incremental costs and available capacity.
Outsourcing and make-or-buy
For outsourcing, compare the supplier’s total evaluated cost with the internal cost that would change as a result of the decision. Allocated fixed costs that remain after outsourcing are not avoidable savings. Conversely, opportunity cost should be included if released capacity can serve profitable demand.
A robust comparison includes price, transition cost, governance, service-level penalties, resilience, data security, labour implications, and switching costs. ABC can reveal the internal resource consumption; it cannot by itself determine strategic suitability of a supplier.
Network and capacity decisions
TDABC makes the capacity consequences of demand visible. A low-margin customer may be attractive when idle capacity exists but unattractive during peak periods. Capacity cost rates should therefore be interpreted alongside utilisation, bottlenecks, peak demand, and service commitments.
Background. NorthStar, a fictional third-party logistics provider, serves retail and e-commerce customers from one regional centre. Management used a single handling rate per parcel. E-commerce Customer A generated many small orders, urgent dispatches, address corrections, and returns. Retail Customer B generated fewer, larger orders with scheduled deliveries.
Problem. The single rate showed similar cost per parcel for both customers. Management was considering a price reduction for A because its reported gross margin appeared high.
Analysis. A project team mapped receiving, storage administration, picking, packing, exception handling, and returns. It created a TDABC model with resource groups for warehouse labour and customer service. Practical capacity was estimated from scheduled time less normal breaks, meetings, training, and equipment downtime. Time equations included order lines, urgent status, special packaging, address correction, and returns inspection.
The analysis showed that A consumed 38% of warehouse practical minutes but only 24% of parcels. Its frequent exceptions and returns generated high customer-service time. B consumed 18% of practical minutes and 27% of parcels. The old parcel rate had therefore cross-subsidised A through B.
Approach and solution. NorthStar introduced: (1) a base fulfilment price, (2) an urgency surcharge, (3) a fee for address correction, and (4) a returns-processing charge. It also offered A a lower price for consolidated order cut-off times and standard packaging. Internally, NorthStar redesigned the address-validation interface and created a returns triage station.
Outcome. The revised model did not simply increase prices. It linked charges to controllable service attributes and identified process improvements. Management retained A but required operational changes and a contract review based on measured resource consumption.
Learning points. A broad average rate can hide cross-subsidies. TDABC can connect customer behaviour, capacity consumption, and commercial terms. The model must be used with judgement: some strategic customers may justify temporary subsidies, and cost transparency must not undermine service relationships.
10. Implementation, governance and technology
A credible implementation requires a cross-functional team involving finance, operations, IT, commercial managers, and frontline employees. Process maps should be validated through observation and operational data rather than relying solely on interviews. Time estimates may be obtained from work measurement, warehouse-management-system timestamps, telematics, barcode scans, or statistically designed samples.
Technology can automate driver extraction and update time equations. Machine-learning systems may predict handling time from order attributes, but governance remains essential. Managers must understand the features used, detect bias, monitor model drift, and distinguish correlation from causation. Automated cost estimates should not penalise customers or employees without a review process.
Data governance should address:
consistent definitions of orders, lines, parcels, stops, and returns;
reconciliations to the general ledger;
version control for rates and equations;
audit trails for changes;
access controls for commercially sensitive customer data;
privacy requirements when worker or driver-level data are used.
Ethically, ABC information should not become a crude employee surveillance system. Time standards should reflect normal variation and safe work practices. Incentives based only on speed may encourage unsafe handling, mis-scanning, concealment of errors, or deterioration in customer service.
11. Limitations and critical evaluation
First, causality is difficult in shared facilities. Rent, security, and management costs may support all customers but have no credible short-term driver. Forced allocations can create false precision.
Second, time equations may be unstable when processes, technology, layouts, product mix, or labour agreements change. Calibration and periodic validation are required.
Third, capacity is not always divisible. A warehouse may need a full shift or vehicle even when demand falls slightly. Linear cost assumptions should therefore be supplemented by step-cost analysis.
Fourth, ABC is mainly a cost measurement system, not a complete performance-management system. It does not automatically capture service quality, resilience, carbon emissions, innovation, strategic relationships, or risk.
Fifth, employees and customers may resist the model if they believe it is intended for blame or price increases. Transparent definitions, pilot testing, sensitivity analysis, and communication improve acceptance.
A useful practice is to report ranges rather than spurious precision. If a picking time is estimated as 2.5 to 3.0 minutes depending on congestion, decision makers should understand the uncertainty and test whether the decision changes across the range.
Best-practice checklist
Start with a clearly stated managerial decision.
Define the model boundary and cost horizon.
Prefer causal drivers over convenient accounting bases.
Separate practical capacity from theoretical capacity.
Report unused capacity explicitly.
Use time equations when transaction attributes explain variation.
Reconcile the model to financial accounts.
Validate estimates with operational observation and system data.
Combine cost with quality, risk, service, and sustainability measures.
Review equations, rates, and assumptions after process changes.
Use sensitivity analysis for uncertain time and capacity estimates.
Protect employee and customer data and provide governance over automated decisions.
Chapter Summary
Activity-Based Costing assigns resource costs to activities and then assigns activity costs to cost objects through causally meaningful drivers. It is valuable where customers, products, or shipments consume logistics support resources in different proportions. A well-designed ABC model requires a defined boundary, activity dictionary, resource pools, driver selection, rate calculation, assignment, and reconciliation.
Time-Driven ABC simplifies the model by estimating a capacity cost rate and the time required for each transaction. Time equations capture variation caused by order lines, urgency, special handling, exceptions, and other observable attributes. Practical capacity deducts normal non-productive time from theoretical capacity, making unused capacity visible rather than hiding it in transaction rates.
ABC and TDABC support customer profitability, service pricing, process improvement, capacity planning, and outsourcing analysis. They do not replace managerial judgement. Allocated cost differs from avoidable cost, and cost information must be interpreted with service quality, strategic value, risk, sustainability, and capacity constraints. Strong governance, data validation, ethical use of employee information, and periodic model updating are essential for reliable decisions.
Key Terms
Resource: An economic input consumed by logistics operations.
Activity: A process performed using resources.
Cost object: The item for which cost is measured, such as a customer or shipment.
Resource driver: A basis for assigning resource costs to activities.
Activity cost driver: A factor used to assign activity cost to cost objects.
Activity cost pool: Costs grouped for an identifiable activity.
Activity-Based Costing (ABC): A costing method that assigns resource costs through activities and drivers.
Time-Driven ABC (TDABC): An ABC method using capacity cost rates and estimated transaction time.
Practical capacity: Realistic usable capacity after normal interruptions and non-working time.
Theoretical capacity: Maximum capacity assuming continuous productive operation.
Capacity cost rate: Cost of supplied capacity divided by practical capacity.
Time equation: An equation estimating transaction time from transaction attributes.
Unused capacity: Supplied practical capacity not consumed by measured transactions.
Cost causality: The relationship by which a driver explains resource consumption.
Cost object profitability: Revenue less relevant costs associated with a customer, product, or service.
Relevant cost: A cost that changes as a result of a decision.
Cross-subsidy: A situation in which one cost object bears costs caused by another.
Step cost: A cost that changes when activity crosses a capacity threshold.
Driver governance: Controls over definitions, data, assumptions, updates, and auditability.
Review Questions
Short Answer Questions
Distinguish a resource, an activity, and a cost object.
What makes a cost driver causally meaningful?
Define practical capacity and explain why it differs from theoretical capacity.
What is the purpose of a TDABC time equation?
Why should unused capacity normally be reported separately?
Long Answer Questions
Explain the complete process of designing and implementing an ABC model in a distribution centre.
Compare traditional ABC and TDABC in terms of data requirements, variation, capacity information, and limitations.
Discuss how ABC information can support logistics pricing and customer profitability analysis.
Explain the difference between allocated cost, avoidable cost, and opportunity cost in outsourcing decisions.
Critically evaluate the governance and ethical risks of using operational data and AI to estimate logistics costs.
Analytical / Application Questions
A warehouse has an annual receiving pool of $240,000 and processes 16,000 receipts. Calculate the receiving rate. A customer generates 75 receipts; determine its assigned receiving cost and state the unit assumptions.
A resource group costs $450,000 and supplies 225,000 practical minutes. A transaction takes 18 base minutes, plus 7 minutes for urgency and 4 minutes for special packaging. Calculate the cost of an urgent transaction with special packaging and explain the result.
A logistics provider’s model shows that Customer P consumes substantial practical capacity but has a low reported cost under a parcel-based rate. Design an analytical investigation using ABC or TDABC, and recommend at least three operational or commercial actions.
Multiple Choice Questions with Answers
Question 1
Which sequence best represents the causal logic of Activity-Based Costing?
A. Cost objects → resources → activities → drivers
B. Resources → activities → cost objects → decisions
C. Activities → revenue → resources → capacity
D. Drivers → cost objects → resources → activities