What six public projects teach us about software management
In France, between 2004 and 2025, six major government digital projects accumulated approximately €1.6 billion in costs without delivering a usable result. Three of them deserve a detailed explanation, as they illustrate drift mechanisms that any company can reproduce on a smaller scale.
Scribe-XPN (Ministry of the Interior): launched in 2015 to digitize police procedure writing. Ten years later, the software remains unusable — it takes seventeen clicks to record a single document. Cost: €257 million.
Louvois (Ministry of Defence): a military payroll system deployed in 2011, immediately defective. Some soldiers were overpaid, others not paid at all. The cost of corrections alone reached €200 million per year.
Sirhen (Ministry of National Education): a human resources management system launched in 2007, abandoned in 2018. After more than ten years of development, it covered only 2% of the 1.1 million employees concerned. Cost: €400 million, including €156 million just to oversee the main contractor.
The tunnel effect: enemy #1
The common thread across these projects is not a technological flaw. It is the absence of intermediate deliveries. You specify for months, you develop for years, and you discover at the end that the result does not match the need. The project drifts gradually, but no one sees it, because no one sees the product.
This phenomenon, called the “tunnel effect”, is not exclusive to government agencies. Any SME that outsources a project to a vendor without direct contact with the developers reproduces the same pattern. The client talks to a salesperson, who briefs a project manager, who relays to an invisible technical team. The message gets distorted at every link.
The real cost of software failure
The report on Scribe-XPN introduces a new data point: the time lost by end users. Between 30 and 45 minutes per day per officer, for 150,000 police officers. This cost does not appear in any official report. And because it is a hidden cost, invisible on paper, it is not taken into account. Development continues in the wrong direction, without targeting user needs — the goal should be to make their lives easier.
For an SME, the equivalent will not be a budget overrun of hundreds of millions of euros. But it will be a one-year delay on the competition. A team that loses trust in the IT tool. Unhappy customers. The cost of software failure is not measured only in development invoices.
A different approach: deliver early, adjust along the way
OLATEQ builds custom software for SMEs by applying a simple rule: deliver quickly in visible increments. The client sees, tests, and we get fast feedback. The project never drifts far from its real need, because it is confronted with reality at every iteration.
This approach rests on three principles:
A single point of contact: The client speaks directly to the developers. No intermediary, no filter, no distortion of the message.
Map before coding: Before writing a single line of code, OLATEQ analyses and simplifies the business process to be automated. Automating disorder does not fix it — software does not resolve organisational chaos.
Exactly what is needed, nothing more, nothing less: Many SMEs turn to generic solutions that they twist to fit their business. In the end, the standard solution costs more in customisation, useless features, and imposed updates than software designed for the actual use case.
Custom software is not a luxury
A project that starts at €100,000 ends up at €500,000 not because of technology, but because nothing was seen for 18 months. When the gap between the need and the result is discovered, it is too late to correct without starting over.
The answer is not more budget. It is more frequency. Deliver quickly, deliver often, adjust continuously. This is exactly what the large public projects failed to do, and what OLATEQ does for each of its clients. A deliberately small team, direct proximity, incremental deliveries. Request a free audit.
This article is based on public facts reported in the press and in institutional reports. The figures cited come from published data and documented debates.