Welcome to a new edition of Business Case — where we take a fictional but realistic scenario, run the numbers, and show you what the cost of inaction really looks like.
Public administration is heading into its hardest decade in living memory. Application volumes keep climbing as legislators add new entitlements and reporting duties, a substantial share of experienced caseworkers will retire within the next ten years, and citizens — trained by next-day delivery and instant banking — measure their city hall against the best digital service they used this morning. The traditional answer to rising workload, hiring more people, is largely off the table: budgets are tight, and the labour market has no queue of trained administrators waiting to be hired.
What makes this a business case rather than a lament is a simple observation: in most administrations, the actual decision on an application takes minutes. Checking whether a housing benefit claim is complete, whether a trade registration meets the requirements, whether a resident parking permit can be issued — these are short, well-defined tasks performed by competent people. Yet the citizen waits weeks. The gap between minutes of work and weeks of waiting is not caused by lazy officials or complicated law. It is caused by the process around the work: unstructured intake, handoffs between departments that nobody tracks, and a near-total absence of visibility into where applications actually get stuck. That gap has a price, and today we are going to calculate it.
Meet the city of Falkenbrück: a fictional but familiar mid-sized German city of 128,000 residents. Its administration employs 2,700 people across a central town hall and several district offices, runs an annual budget of around 640 million euros, and spends roughly 190 million euros of that on personnel. Every year its departments — the citizens' office, the building authority, the social services office, the foreigners' registration office, the trade office — receive about 240,000 applications and formal requests, from passport renewals and residence registrations to building permits and benefit claims.
On paper, Falkenbrück is a well-run city: a balanced budget, no scandals, satisfaction scores in the middle of the pack. But the numbers underneath the surface tell a different story. The average application takes 34 working days from receipt to decision; digitised peer municipalities manage comparable cases in about twelve. The fully loaded cost of processing one application sits near 48 euros; the benchmark is closer to 25. Forty-two percent of incoming applications are incomplete and trigger at least one formal follow-up request, against a best-practice rate below 15 percent. Every application generates on average 2.7 status inquiries by phone or e-mail. And across all offices, roughly 19,000 cases are open at any given time — a backlog that has grown every year for the past five.
None of these numbers appears in the city council's regular reports. Individually, each one can be explained away — a staffing gap here, a complicated legal change there. Together they describe an application processing landscape that has grown by accretion rather than design, and that nobody fully owns. The chief administrative officer has started asking the question every business case in this series starts with: what is this actually costing us, and what would it take to stop it?
When a small project team traced how applications really move through Falkenbrück's offices, three structural problems came into focus. None of them is unique to this city. They are the standard failure modes of public sector processes that were digitised on the surface — a PDF instead of a paper form — without anyone redesigning what happens behind the intake.
The first problem is unstructured intake. Applications arrive as paper, as e-mails, as PDF attachments, and for some services through a state portal that forwards a form the caseworker prints out anyway. Nothing validates completeness at the moment of submission: no check whether the income statement is attached, whether the ID is still valid, whether the address matches the register. The consequence is that 42 percent of applications enter the queue incomplete, and the incompleteness is discovered days or weeks later, by a human, at the desk. Then a formal follow-up letter goes out, the case waits for the citizen, the citizen responds, and the loop often repeats. Each loop adds around twelve working days and one extra handling touch. What should be the exception has quietly become the standard path — a pattern we looked at in detail in our recent quick tips on handling exceptions without breaking flow.
The second problem is that no end-to-end workflow exists across departments. A building permit touches the building authority, fire safety, the environmental office, sometimes monument protection. A benefit claim moves between intake, assessment, and payment. Each office has its own tray, its own specialist software, its own working rhythm — and the handoffs between them happen by internal mail and e-mail. Nobody owns the case from receipt to decision, a failure mode we have examined before in our business case on end-to-end case management. When the project team measured where the 34 working days actually go, the finding was stark: roughly 29 of them are pure waiting time between desks. The application is not being processed slowly. It is mostly not being processed at all — and nobody notices, because nobody can see it.
The third problem is that nobody reports on patterns. Management sees a monthly aggregate: how many applications came in, how many were closed. Nobody sees which application types stall at which step, which follow-up loops repeat every month, or why one district office resolves the same case type nearly twice as fast as another with identical staffing. And because citizens have no visibility either, they call. Those 2.7 inquiries per application add up to well over 600,000 contacts a year — each one answered by the same caseworkers who are supposed to be working down the backlog the calls are about.
These three problems compound, and the cost surface they create is larger than any single office can see. Start with processing: across 240,000 applications, the fully loaded cost gap against the benchmark is 23 euros per case; applying that gap conservatively to only the 60 percent of volume that is realistically addressable in a first phase yields roughly 3.3 million euros a year in excess handling and rework. Add the inquiry burden: bringing the average from 2.7 contacts per application down to around one — mostly through proactive status information — removes about 400,000 contacts a year, worth roughly 2.4 million euros at a conservative six euros of fully loaded staff time each. Then add the consequences of the backlog itself: overtime, temporary staff, and a growing number of statutory deadline breaches that end in formal inaction complaints with legal costs attached — together an estimated 0.6 million euros annually. The combined cost of Falkenbrück's invisible application process is on the order of 6.3 million euros a year — the equivalent of more than ninety full-time positions, over three percent of the entire personnel budget, spent not on serving citizens but on the friction between desks. And unlike a company, a city cannot lose its customers to a competitor. The cost stays inside, and it shows up as burnout, backlog, and waiting rooms.
The path forward in Falkenbrück is not a decade-long IT megaproject, and it is not a reorganisation. It is three integrated levers that turn citizen services from a chain of disconnected trays into a structured, visible, fast process. Each lever maps directly onto one of the three problems above.
The first lever is digital, structured intake with completeness validation. Applications are submitted through online forms that check required documents and plausibility at the moment of submission: the upload is verified, the fields are validated against register data where legally permitted, and the applicant gets immediate feedback on what is missing. An incomplete application simply cannot enter the queue — the follow-up loop moves from week three to minute one, and from a formal letter to an on-screen message. In Falkenbrück's pilot, covering the citizens' office and the housing benefit section, the incomplete rate fell from 42 percent to under 15 percent within months. Applied to the addressable volume, this lever recovers roughly 1.7 million euros a year. It costs about 240,000 euros in the first year for form design, register integration, and legal review, and around 80,000 euros a year to run.
The second lever is an end-to-end workflow across departments with owned deadlines. Every application becomes a case in one workflow, with defined steps, responsible roles, and target times per step. Handoffs are system transitions rather than envelopes: when the building authority finishes its part, the case appears in the fire safety office's queue the same second, with the full history attached. When a step exceeds its target time, the case escalates visibly instead of silently ageing in a tray. In the pilot, average processing time fell from 34 to around 14 working days — not because anyone worked faster, but because the idle time between desks collapsed. This lever recovers roughly 1.9 million euros a year, at about 160,000 euros in year one and 70,000 euros annually thereafter.
The third lever is process mining on the case data, feeding a monthly steering review — combined with proactive status updates to applicants. Once cases flow through one system, the patterns that no report would surface become visible: which application types stall at which step, which loops repeat, which offices process identical cases at different speeds and why. This is the domain of process intelligence tooling — the category in which noreja operates, catalogued alongside peers on directories such as topai.tools. The findings drive a focused monthly meeting where the worst patterns are fixed by name rather than in aggregate. At the same time, automatic status notifications — received, in assessment, decision expected by a stated date — remove the reason for most inquiries before the phone rings. This lever recovers roughly 1.6 million euros a year, on tooling and governance costing about 80,000 euros fully loaded.
Combined, the three levers recover roughly 5.2 of the 6.3 million euros of annual cost in steady state — about four-fifths of the leak. First-year implementation, including training and process redesign, comes in at approximately 520,000 euros, with annual run costs around 230,000 euros thereafter. Payback is reached well within the first year, typically in the first few months after rollout. And one effect no business case fully captures: a citizen who receives a decision in two weeks instead of seven, and who can see its progress along the way, stops experiencing the administration as a black box — which is worth more to public trust than any campaign.
If you could see only the applications with a complete digital trail — received, validated, routed, decided — how much of last month's caseload would still be visible?
Of the working days an application spends in your organisation, how many involve someone actually working on it — and who in your management team could answer that question today?
What share of your inbound calls and e-mails are people asking where their case is, and what would it cost to answer that question before it is asked?
If two of your offices process the same application type at very different speeds, is that a staffing difference or a process difference — and how would you tell the two apart?
What would your administration do with the equivalent of ninety recovered full-time positions?
Slow citizen services are rarely the fault of the people delivering them. They are the predictable result of applications that enter the building unchecked, travel between departments unseen, and age in queues that no one measures. The fix is not more pressure on caseworkers, and it is not a bigger IT budget spent on digitising the same broken sequence. It is a structured intake that keeps incomplete applications out of the queue, one workflow that owns the case across every desk it touches, and process mining that shows management where the days actually go. Once those three are in place, application processing stops being something the administration apologises for and becomes something it can steer.
We invite you to look at your own numbers with Falkenbrück's lens. Measure the true end-to-end time of your three highest-volume application types, count how many touches and status inquiries each one generates, and put a price on the gap between your process and the benchmark. Then decide whether that gap should survive another budget year.
Rarely because of the actual work: the decision on most applications takes minutes. The time is lost around the work — incomplete submissions that trigger follow-up loops, handoffs between departments by internal mail, and cases idling in queues nobody measures. In typical administrations, the large majority of end-to-end time is pure waiting.
Structured digital intake means applications are submitted through online forms that validate completeness and plausibility at the moment of submission, so an incomplete application cannot enter the queue. Since incomplete submissions are the single biggest source of loops and delays, fixing intake typically removes weeks from processing times before any other change is made.
An end-to-end workflow turns every application into one visible case with defined steps, responsible roles, and target times — across all departments it touches. Handoffs become system transitions instead of envelopes, and overdue steps escalate automatically. Processing time falls mainly because idle time between desks collapses, not because anyone works faster.
Process mining analyses real case data to reveal patterns no standard report shows: which application types stall at which step, which follow-up loops repeat, and why identical offices work at different speeds. It gives public sector leaders the evidence to fix the worst bottlenecks by name and to verify that changes actually work.
For a mid-sized administration, first-year implementation costs are typically a fraction of the annual cost of inaction, so payback lands well within the first year — often within months of rollout. The recovered capacity shows up as reduced backlog, fewer status inquiries, and staff time returned to actual casework.