Time-driven activity-based costing: ABC made simple and scalable
Time-driven activity-based costing, TDABC, is what activity-based costing became once its creators had lived with its weaknesses. Classic ABC was accurate but expensive: it leaned on endless employee surveys, grew unmanageable as activities multiplied, and quietly assumed that everything ran at full capacity. TDABC keeps the goal, costing products and customers by what they really demand, but reaches it with a radically simpler machine. Instead of a cost pool and a survey for every activity, it needs only two things per group of resources: how much it costs to supply a minute of capacity, and how many minutes each transaction takes.
That shift does more than save effort. By measuring the cost of supplying capacity and the time work actually consumes, TDABC makes the gap between the two, unused capacity, visible for the first time, rather than burying it in inflated rates. The method is the engine under most modern profitability and cost-to-serve work, and this profile sets out where it came from, how its two parameters fit together, and the honest limits the later literature has identified.
time-driven activity-based costing (TDABC) is a costing method that assigns the cost of resources to products, services and customers using just two parameters per resource group: a capacity cost rate, the cost of supplying capacity divided by the practical capacity available, and time equations that estimate how many minutes each transaction consumes. It was developed by Robert S. Kaplan and Steven R. Anderson, set out in their 2004 Harvard Business Review article and their 2007 book, building on the capacity-costing ideas in Kaplan and Cooper's Cost and Effect (1998). Because it works in units of time rather than activity surveys, TDABC is far faster and cheaper to build and maintain than classic ABC, scales to enterprise size, draws directly on ERP and CRM data, captures real-world variety through its time equations, and, crucially, surfaces unused capacity explicitly. Critics note that its accuracy depends on the quality of the time estimates and that it is harder to apply to non-routine knowledge work. It is the method most modern customer-profitability and cost-to-serve analysis runs on. ---
Where it came from
Time-driven activity-based costing was developed by Robert S. Kaplan and Steven R. Anderson. Anderson had founded the software and consulting firm Acorn Systems in 1996 and pioneered the time-based approach to activity costing there, working with companies to make activity-based costing practical at scale. The conceptual ground had already been prepared: a chapter of Kaplan and Cooper's 1998 book Cost and Effect set out the idea of costing the supply of resource capacity, the seed from which TDABC grew.
The method was formally introduced to a wide audience in Kaplan and Anderson's article "Time-Driven Activity-Based Costing", published in the Harvard Business Review, Volume 82, Number 11, in November 2004, pages 131 to 138. They then developed it in full in their book Time-Driven Activity-Based Costing: A Simpler and More Powerful Path to Higher Profits (Harvard Business School Press, 2007). The framing throughout is deliberately practical: TDABC is presented not as a rival theory of cost but as the way to get the benefits of activity-based costing without the implementation burden that had caused so many ABC projects to stall or be abandoned.
How it works
TDABC reduces the machinery of activity-based costing to two parameters for each group of resources that does broadly similar work (a department, a process, a team). Everything else follows from them.
Parameter one: the capacity cost rate
The first parameter is the capacity cost rate: the cost of supplying capacity divided by the practical capacity available to do work.
- The numerator is the total cost of the resource group over a period, salaries, supervision, equipment, occupancy, technology, everything needed to make that capacity available.
- The denominator is the group's practical capacity, not its theoretical maximum. People and machines do not work every paid minute; time goes to breaks, training, maintenance and slack. Practical capacity is therefore usually estimated at around 80 to 85 per cent of theoretical capacity.
Dividing one by the other gives the cost of a single unit of capacity, normally a cost per minute. In the original presentation, a resource group costing 560,000 with a practical capacity of 700,000 minutes gives a capacity cost rate of 560,000 / 700,000 = 0.80 per minute.
Parameter two: time equations
The second parameter is a set of time equations that estimate how many minutes each transaction takes. This is what gives TDABC its power to handle variety without a separate activity for every case. A time equation starts with a base time and adds increments for the features that make a particular transaction take longer.
A worked illustration from the original work: a packaging time might be expressed as 0.5 + 6.5 (if special handling is required) + 0.2 (if the order is shipped by air). A standard order takes 0.5 minutes; one needing special handling takes 7.0; one needing both special handling and air shipping takes 7.2. One equation, fed by data the business already records, captures dozens of variants that classic ABC would have had to model as separate activities.
Putting them together
Multiply the time a transaction consumes by the capacity cost rate and you have its cost. Sum across all the transactions a product or customer triggers and you have its total cost. Critically, because you costed the supply of capacity against practical capacity, the minutes actually used will normally be less than the minutes available. The difference, valued at the capacity cost rate, is the cost of unused capacity, reported explicitly rather than smeared into every transaction. That single number, invisible in classic ABC, is one of the method's most valued outputs, and it makes TDABC a natural tool for forecasting resource demand as volumes change.
A worked example
Take an illustrative mid-size B2B distributor we will call CaP Distribution (the figures are illustrative, used only to show the mechanics). Consider its order-processing resource group.
| Parameter | Value (illustrative) |
|---|---|
| Annual cost of the resource group | EUR 480,000 |
| Practical capacity (about 85% of theoretical) | 561,000 minutes |
| Capacity cost rate | EUR 0.856 per minute |
The capacity cost rate is 480,000 / 561,000 = approximately EUR 0.856 per minute. Now suppose the team's order-handling time follows a simple time equation: a base of 3.0 minutes per order, plus 5.0 minutes if the order needs manual credit clearance, plus 2.0 minutes if it ships to an export destination.
| Order type | Time equation (minutes) | Minutes | Cost at EUR 0.856/min |
|---|---|---|---|
| Standard domestic order | 3.0 | 3.0 | EUR 2.57 |
| Order needing credit clearance | 3.0 + 5.0 | 8.0 | EUR 6.85 |
| Export order, credit clearance | 3.0 + 5.0 + 2.0 | 10.0 | EUR 8.56 |
A customer placing many simple domestic orders is cheap to serve; one whose orders routinely need credit clearance and export handling costs more than three times as much per order, and TDABC quantifies it. If, over the year, the orders processed consume 510,000 of the 561,000 available minutes, the remaining 51,000 minutes are unused capacity, worth 51,000 x 0.856 = approximately EUR 43,656, reported as a line of its own rather than hidden in the per-order cost. That is the visibility classic ABC could not give.
The strengths of TDABC are, in large part, the mirror image of ABC's weaknesses, which is precisely why Kaplan and Anderson built it. It is fast and cheap to build and maintain: two parameters per resource group, rather than a cost pool and a survey for every activity, mean a model that can be stood up in weeks and refreshed by updating rates and times rather than re-running a survey programme.
It integrates with existing systems. Because time equations are driven by transaction attributes, order type, customer, channel, service level, the model can be fed directly from the ERP and CRM data a business already captures, rather than from special data-collection exercises. It scales to enterprise size: adding products, customers or variants means extending time equations, not multiplying activities, so the model grows linearly rather than exploding. It captures real-world variety through those same equations, handling the long tail of order and customer types that classic ABC struggled to represent. And it does the two things managers most want: it surfaces unused capacity explicitly, and it supports forecasting of resource demand, because once you know the minutes each transaction takes, you can model what a change in volume or mix will require.
The founding sources for TDABC are, by their nature, advocacy: Kaplan and Anderson set out to show what the method can do, and they do not present a list of its shortcomings. The fair criticisms come from the later literature, and they are worth stating plainly.
First, accuracy depends on the quality of the time estimates. The two-parameter design is an advantage only if the times in the equations are right; poor or stale estimates propagate straight through to the costs, and a model can look authoritative while resting on weak inputs. The discipline of validating times against observation matters as much as the elegance of the structure.
Second, critics note that TDABC is harder to apply to non-routine knowledge work. The method is at its strongest where work is repetitive and time is a sensible measure of resource consumption, order processing, warehousing, claims handling, routine clinical pathways. Where the work is creative, variable and judgement-heavy, the time of a task is a poorer guide to its cost, and the equations become strained. Third, and relatedly, the method assumes time is the dominant cost driver. For most operational and transactional work it is, which is why TDABC fits cost-to-serve so well, but where cost is driven by something other than time, the central assumption holds less firmly. These are limits of fit and of input quality, not flaws in the logic, but they should temper any claim that the method is universally applicable.
Where it fits
TDABC is the method most modern profitability work runs on, and its natural home is customer profitability and cost-to-serve: working out what it truly costs to serve each customer, channel and order type once the demands they place on the business are properly counted. It fits wherever there are large transaction volumes of broadly routine work, which is why it has been adopted across logistics and distribution, healthcare, and financial services, alongside manufacturing and B2B distribution more generally.
Industries with high transaction counts and meaningful variety, many customers, many order types, many service levels, are exactly where its time-equation approach earns its keep, and where the unused-capacity output is most actionable. It is the engine beneath cost-to-serve analysis and the whale curve of customer profitability, and it is the natural choice when a model has to stay live, scale with the business, and draw on transactional data rather than periodic surveys. Where the work is non-routine and judgement-heavy, or where time is not the dominant cost driver, its fit is weaker and the time estimates need particular care.
FAQ
What is time-driven activity-based costing in simple terms?
Time-driven activity-based costing is a simpler version of activity-based costing that works out what products and customers cost using just two pieces of information for each group of resources: how much it costs to supply a minute of capacity, and how many minutes each transaction takes. Multiply the minutes by the cost per minute and you have the cost. Because it counts the cost of supplying capacity against the capacity actually available, it also shows, as a separate figure, how much capacity is going unused.
Who developed TDABC and when?
Time-driven activity-based costing was developed by Robert S. Kaplan and Steven R. Anderson. Anderson had pioneered the time-based approach at his firm Acorn Systems, which he founded in 1996, building on capacity-costing ideas in Kaplan and Cooper's 1998 book Cost and Effect. The method was introduced in their Harvard Business Review article in November 2004 and set out in full in their 2007 book Time-Driven Activity-Based Costing: A Simpler and More Powerful Path to Higher Profits.
What are the two parameters in TDABC?
The two parameters are the capacity cost rate and the time equations. The capacity cost rate is the cost of supplying a resource group divided by its practical capacity, usually expressed as a cost per minute, where practical capacity is typically estimated at about 80 to 85 per cent of theoretical capacity. The time equations estimate how many minutes each transaction takes, starting from a base time and adding increments for the features that make a particular transaction take longer.
How does TDABC reveal unused capacity?
TDABC costs the supply of capacity against the practical capacity available, then measures the minutes that work actually consumes through its time equations. Because the minutes used are normally fewer than the minutes available, the difference, valued at the capacity cost rate, is the cost of unused capacity. Classic activity-based costing tended to assume full capacity and roll any idle time into its rates, which hid this figure; TDABC reports it as a line of its own.
Is TDABC better than ABC?
For most modern applications it is more practical: it is faster and cheaper to build and maintain, scales to enterprise size, draws on existing ERP and CRM data, and surfaces unused capacity, which classic ABC hides. It is not a different theory of cost so much as a simpler, more maintainable way to deliver the same insight. The honest caveats are that its accuracy depends on good time estimates and that it fits routine, transaction-heavy work better than non-routine knowledge work. See the full comparison at /tdabc-vs-abc/.
Go deeper
Want to see what it really costs to serve each of your customers? Start with a Profit Check.
Start a Profit CheckExplore the Profitability Lab