Calculating cost to collect: activity cost and activity effectiveness
Purpose. Cost to collect is not one headline number. It is the result of which activities run, what each unit costs, how often they succeed, and how much rework and delay they create. Use this with NPV in collections so strategy choices are defensible, not guessed.
Why “cost to collect” is often underestimated
In many collections environments, cost to collect is treated too loosely. People often use broad assumptions: internal is expensive, DCA is commission only, legal is high cost. That is directionally true, but not enough for decisioning.
If the organisation wants to use NPV properly, it needs to understand that cost to collect is not only “what did we pay?” It is what we spent, directly and indirectly, to generate recoveries. Cost to collect should be built from the actual work performed, not only from high-level averages.
Cost to collect is built from activities
Collections strategies are made up of activities. Examples include: outbound phone call, SMS or email, letter generation, payment arrangement setup, hardship referral handling, dispute handling, legal pack preparation, DCA placement, DCA recall, payment reconciliation, complaint management.
Each activity has a cost, an effort level, and an effectiveness rate. Two strategies may look similar at portfolio level while the underlying activity mix is very different.
Two core inputs: activity cost and activity effectiveness
A robust cost-to-collect model needs two building blocks.
Activity cost
Activity cost is the cost of performing one unit of work: one outbound call, one letter, one arrangement review, one DCA placement, one legal referral pack. Estimate using labour cost, system cost, vendor fee, and overhead allocation.
Activity effectiveness
Activity effectiveness is how often that activity leads to the intended outcome. Examples: call connects with customer 30% of the time; SMS generates payment 8% of the time; arrangement review yields a sustainable arrangement 45% of the time; DCA1 placement results in recovery 22% of the time.
Without effectiveness, raw cost tells you very little. A cheap activity that never works is not actually cheap. An expensive activity that consistently resolves accounts may be very efficient.
The practical formula
At a simple level:
Cost to collect ≈ Volume of activity × Unit cost of activity
For decisioning you also need:
Cost per successful outcome = Total activity cost ÷ Number of successful outcomes
This is where effectiveness matters.
| Activity | Unit cost | Volume | Total cost | Success rate | Cost per success |
|---|---|---|---|---|---|
| Outbound call | $6 | 1,000 | $6,000 | 10% | $60 |
| SMS | $0.15 | 10,000 | $1,500 | 2% | $7.50 |
| Legal pack | $120 | 100 | $12,000 | 40% | $300 |
This helps answer which activities are cheap but ineffective, which are expensive but worthwhile, and which should be used earlier or later in the lifecycle.
What should be included in activity cost
When calculating activity cost, do not only look at direct transaction cost. A realistic activity cost should include some combination of: staff time, system usage, technology licence allocation, management overhead, QA and compliance overhead, reconciliation effort, vendor fee where relevant, downstream rework where known.
Example: outbound call. Cost is not only telco. It may include agent time, dialler or platform cost, QA sampling, team leader support, call recording and storage. That is why internal collection activity often costs more than people assume.
Direct versus indirect cost
Direct cost is directly attributable to the activity: DCA commission, legal invoice, postage, SMS charge.
Indirect cost is supporting cost required to enable the activity: operations management, governance forums, reporting and MI, compliance oversight, system maintenance.
If you only include direct cost, internal activity often appears too cheap. If you include everything poorly, you can overload the model. The goal is not perfect accounting purity. The goal is decision usefulness.
Activity effectiveness must be measured by outcome, not only completion
A common error is to treat activity completion as effectiveness. That is wrong.
Completion-style metrics: calls made, letters sent, accounts touched. These are volumes, not effectiveness.
Effectiveness-style metrics: payment generated, arrangement agreed, dispute resolved, hardship referred correctly, account cured, balance reduced, recovery achieved.
The model must distinguish “we did the activity” from “the activity worked.”
Levels of effectiveness
Activity effectiveness: did the activity drive a useful step? (Example: letter prompted contact; call produced promise to pay.)
Case effectiveness: did the activity contribute to account resolution? (Example: arrangement led to cure; DCA placement led to recovery.)
Portfolio effectiveness: does this activity mix improve overall outcomes? (Example: replacing some outbound calls with SMS reduced cost and maintained cures; DCA2 is more expensive per dollar recovered than debt sale.)
Why this matters for strategy design
Once you understand activity cost and activity effectiveness, you can stop treating strategy as a black box. You can compare internal phone-heavy, digital-first, DCA1 commission, DCA2 specialist, and legal escalation strategies not only on gross recovery but on cost per dollar recovered, cost per successful outcome, speed to outcome, and sustainability of outcome.
Worked example: internal versus DCA
Imagine two strategies for a similar pool of accounts.
Internal strategy: 5 calls, 2 SMS, 1 letter, 1 arrangement review, staff-led payment handling. Total average cost per account: $85. Average recovery: $420.
DCA strategy: placement fee embedded in commission; vendor handles calls, letters, and follow-up; CP handles oversight and reconciliation. Average effective cost per account: $55. Average recovery: $360.
At first glance internal looks better on recovery. Net on a simple cash basis: internal $420 − $85 = $335; DCA $360 − $55 = $305. Internal is still better here, but only slightly. Once timing, overhead, and capacity are included, DCA may become more attractive for some segments. That is why this must be modelled, not guessed.
Cost to collect changes over time
Cost to collect is not fixed across the lifecycle. An account may be cheap to work early, expensive later, then cheap again if sold. Early digital nudges can be cheap; repeated manual outbound is expensive; legal escalation is very expensive; debt sale has no future collection cost but lower realised value. Treat cost to collect as strategy- and stage-specific, not one flat rate per account.
Activity cost must include failure and rework
A mature model should recognise the cost of activities that do not work or create downstream work: repeated failed calls, broken arrangements, recall and reallocation effort, DCA underperformance requiring rework, complaints from poor treatment, credit reporting corrections, reconciliation breaks. If an activity drives high rework, its true cost is much higher than direct execution cost.
Linking activity cost to NPV
NPV should not rely on a generic “collection cost” input only. It should incorporate activity mix, activity unit cost, expected effectiveness, strategy-specific overhead, and switching cost between strategies. That supports a realistic comparison between internal, DCA1, DCA2, legal, sale, and abandon. See NPV in debt collections and DCA decisioning.
Common mistakes in cost-to-collect modelling
- Treating internal collections as “free” because salaries are budgeted. Internal effort has real labour and overhead cost.
- Using only DCA commission as DCA cost. Add oversight, reconciliation, governance, complaint handling, and credit-reporting dependency where material.
- Measuring activity volume, not effectiveness. High contact volume can still mean inefficiency.
- Ignoring timing. An activity that works eventually may still be poor value if it takes too long.
- Ignoring downstream correction cost. Complaints, wrong default listing, and recall failures belong in the model.
How to start if the organisation is not mature
Most organisations do not have perfect activity costing. That should not stop the model. Start with: identify core activity types; estimate unit cost for each; estimate effectiveness by segment; build strategy-level average cost; refine using actual data over time. The first model does not need to be perfect. It needs to be useful and directionally right.
Bottom line
Cost to collect is not one number. It is the result of what activities are performed, how much those activities cost, how often they work, and how much rework and delay they create. A more mature model moves strategy from “this feels like the right channel” to “this is the most economically and operationally effective path for this segment.”
Link to the rest of the pack
Cost to collect is driven by operating design, control strength, DCA performance, data quality, and reporting integrity. Explicit links:
- NPV in debt collections and DCA decisioning
- Allocation strategy and decisioning
- Controls framework
- Recall and reassignment
- Reporting and performance management
- CP / DCA files and reconciliation
Back to pack home · NPV appendix · Effectiveness (internal, DCA1, DCA2)