Activity-based costing software: what it must do in 2026

What must ABC software actually do in 2026?

Work at transaction level. Averages are where profitability goes to hide. If the tool cannot ingest order lines, shipments and invoices, it will average away exactly the variety you bought it to reveal. For scale reference, our largest CostCtrl model processes 525,000 transaction rows for a logistics operator.

Model time, not survey percentages. Classic ABC asked staff what percentage of time each activity takes, then went stale within a year. Time-Driven ABC (TDABC) replaces surveys with time equations: minutes per task, cost per minute of capacity. In 2026 this should be the default, not the premium option. See TDABC vs ABC for the full comparison.

Show unused capacity. A tool that spreads 100% of cost across products is hiding your idle capacity inside the rates. TDABC separates the cost of work done from the cost of capacity unused, which is where many of the decisions live.

Produce decision outputs, not just allocations. Whale curves, customer and product P&Ls, cost-to-serve by order type. If you still need a separate BI project to see the answer, the tool is half a tool.

Load from what you already have. CSV exports and, in Europe, SAF-T files. An integration project before the first insight is a 2010 pattern.

Be maintainable by finance. The model dies when the only person who understands it leaves. The maintenance test is simple: can your controller update a time estimate without a consultant on the phone?

Should you build or buy?

Build (in spreadsheets or a warehouse) when the model is small, the analysis is one-off, and someone in-house genuinely wants to own it. Plenty of good first models started life in Excel, including some of ours.

Buy when any of these are true: the spreadsheet has passed roughly 50,000 rows or crashes under refresh; more than one person needs to trust the numbers; the model must survive its author; or you refresh more often than quarterly.

The quiet failure mode of build is not the first version. It is version twelve, eighteen months later, when nobody can explain a rate.

What belongs on the evaluation checklist?

Score each candidate against these ten, and make the vendor demonstrate rather than assert:

1. Ingests transaction-level data (order lines, shipments, invoices) at your real volumes.
2. Uses time equations, not percentage surveys, as the core allocation logic.
3. Separates used from unused capacity in every output.
4. Produces a whale curve and per-customer, per-product P&L natively.
5. Loads from CSV and SAF-T without custom integration.
6. A controller can change a time estimate and re-run the model unaided.
7. Model logic is inspectable: you can trace any cost figure back to its equation.
8. Handles your growth: test with double your current row count.
9. Pricing is transparent before the pilot, subscription rather than open-ended services.
10. Someone accountable helps you build the first model, then leaves you autonomous.

Any tool that fails 2, 6 or 7 will become the next spreadsheet problem, just more expensive.

When do you NOT need ABC software?

Honestly: more often than software vendors say.

You do not need software if the question is one-off. "Are our small orders profitable?" can be answered with a two-week diagnostic on exports, no licence required. That is literally what our free Profit Check and ProfitAudit 360 exist for.

You do not need software if you have fewer than a few thousand transactions a year and low variety. A well-built contribution margin analysis may carry you for years.

And you do not need software before you have a method. Buying a tool to compensate for an undefined costing approach automates the confusion. Method first, tool second.

The honest sequence: diagnostic, then method, then software if the cadence and volume justify it.

Where does CostCtrl fit?

CostCtrl is our TDABC platform: time equations, capacity cost rates, whale curves, transaction-level datasets loaded from CSV or SAF-T, subscription pricing. It is built for SMEs and mid-market companies, not for enterprise chargeback across hundreds of departments; heavier platforms exist for that, and we compare ourselves to them openly (see the SAP PCM and PaPM comparison).

The distinguishing choice is consulting embedded at the start. We run the Profit Check, build the first model with your team, and hand it over. In one distributor case that first model surfaced roughly EUR 1.335M of hidden cost-to-serve across 1,951 accounts; the finance team still runs the model.

Fair questions.

What is the difference between ABC software and TDABC software?

Classic ABC software allocates cost using surveyed percentages of staff time; TDABC software uses time equations and capacity cost rates. TDABC is faster to build, cheaper to maintain and shows unused capacity. Most serious tools in 2026 are time-driven or moving that way.

Can Excel do activity-based costing?

Yes, up to a point: small models, one-off analyses, a patient owner. Excel fails on refresh cadence, row volume and survivability. If the model matters monthly, it needs a home that is not a workbook.

How much does ABC software cost?

Ranges vary widely by segment, from SME subscriptions to enterprise platforms priced per implementation. CostCtrl is subscription-priced and scoped transparently against your data before you commit. For competitors, verify pricing directly with each vendor.

How long does implementation take?

With TDABC and standard exports, a first working model in weeks. Survey-based ABC or integration-heavy enterprise tools typically run months to quarters.

Do we need clean data first?

Cleaner is better, but a costing project is one of the best data-cleaning forcing functions there is. Start with the exports you have; the first model will tell you which fields matter.