Perspectives
Eight ideas about the economics of back-office work.
These are the positions behind everything JC360 says in the US market, and behind the comparison pages. Each one rests on a public fact. None of them depends on believing a vendor, including us.
-
A productivity percentage is not a denominator.
Every vendor in this category reports in percentages: more staff in the optimal band, less non-core time, a double-digit productivity uplift. What none of them publishes is what one unit of work costs.
A percentage with no denominator is a claim you cannot check. The back-office benchmarks that circulate in this market give an uplift range and no currency, which tells you they were written to be quoted rather than to be used.
-
Cycle time is not unit cost.
Faster is not cheaper until someone measures what the remaining human handling costs.
A large US mortgage servicer told investors this year that cycle times fell by 40 to 80 percent where automation was deployed, and in the same quarter stated a direct servicing cost per loan of $89. Two numbers, no bridge between them — and the bridge is the only thing a CFO actually needs.
-
Every AI-in-operations claim needs a baseline underneath it.
A fraction of a point off a loss-adjustment expense ratio after a year of AI. A 40 percent reduction in reconciliation work. An 80 percent reduction in revenue-cycle staffing workload. Each is a public statement from a large US operation this year, and each may well be true.
But none can be attributed to the AI without a transaction-level baseline taken before it was switched on. Without that, the number is a before-and-after with no before.
-
Evidence about the work is not oversight of the people.
What a transaction costs and where capacity sits is evidence about the work. It is not an application that watches people or decides anything about them — and that distinction is the whole difference between being welcome inside a regulated custodian and being shown the door.
When a chief executive says at an annual meeting that they would rather build technology than buy it, the reply is that evidence is not technology. Nobody builds their own baseline; they either have one or they do not.
-
When headcount stops growing, capacity beats cost as the argument.
A large insurer has said its claims call-centre population is down by a third and four centres are becoming two. A health system has committed to serving significantly more customers with no increase in headcount.
In both, the question is no longer what the work costs. It is what the remaining team can absorb, and when the next surge arrives — a capacity question, answered by measurement rather than by a budget line.
-
A greenfield floor is the cheapest moment to instrument.
There is no existing baseline to defend, and nobody has yet written the report that will later be quoted back at you. A capability centre building toward three thousand seats; a health system bringing fifteen thousand revenue-cycle roles in-house from a vendor.
Instrument the work during the build and the objection “we already measure this” never gets the chance to arise. Wait a year and it is the first thing you hear.
-
Every acquisition imports a different definition of a productive hour.
Four merged companies plus a nine-hundred-person bolt-on, run by four presidents. Two thousand people absorbed into four-country delivery in a single quarter.
Roll-ups run on inherited reports that each count a productive hour differently, which means the consolidated number is not one number. It is five numbers added together as if they were the same unit.
-
Timesheets and workforce-management tools measure time, not cost.
Handle time, adherence, capacity, throughput — all real, all useful, and the category’s own product pages describe exactly that and then stop.
“Our workforce-management licence already covers this” is true right up until someone asks what one claim costs to work. Then it is not, and the licence was never going to be.
Where these come from
JC360 is where these ideas were formed and where they are tested. Fifteen years in production at 150+ enterprises across 17 countries, measuring back-office work at the level of the transaction, the process step and the person — automatically, with no timesheets.
One telecoms operator measured 250+ business processes across 2,000+ employees at a 100% sample rate, recovered 6,800 hours a month, and paid back the investment inside the first month. On-premise or cloud. GDPR readiness certificate and ISO 27001.
To see the ideas applied to a specific vendor evaluation: the comparison pages.
Start with a pilot
See your operation the way JC360 sees it.
A two-to-four-week proof of concept, at no cost, configured to your processes — on-premise or hosted. You keep the findings either way.
Book a walkthroughOr email peter.varga@jobctrl.com directly.
Comparing vendors? The five questions worth asking any of them — and JC360’s answers.