Your project management software is built backwards

By Charley Fredaigues

  • Organization
  • Project
project management software upside down

... and that is why your projects fail

There is now more project management software than ways to fail a project. And that says it all. Asana, Monday, Trello, Jira, Wrike, ClickUp, Notion: every quarter brings a new promise of productivity. The board is smoother, the AI more enthusiastic. What is more, the market is worth tens of billions of euros and is growing at double-digit rates. Never has a company had so many means at its disposal. It now knows who must do what, and when.

And yet. Thirty years after the first serious studies on project performance, the verdict has not moved. Most projects still go off the rails. They blow their budgets, deliver less than promised, or simply die. Yet we have digitized the Gantt chart about a thousand times. And the Gantt chart dates from 1910. Despite that, the failure rate has remained Olympian in its calm. There is a paradox here that the software industry does not like to face. What if all this project management software had been, from the start, built backwards?

Thirty years of project management software, zero progress: the paradox nobody dares to name

Let us begin with the only judge that counts: the figures. The Standish Group has published its CHAOS Report since 1994. It is the international reference on IT project performance. In the 2020 edition, the finding comes down to three numbers. First, only 31% of projects are truly successful. In other words, they are delivered on time, within budget, and with a satisfactory result. Second, 50% are "challenged": completed, but late, over budget or stripped of features. Finally, 19% are cancelled before seeing the light of day. Incidentally, the average "challenged" project costs about 189% of its initial budget. You read that correctly: almost double.

These figures bear a strange resemblance to those of the 1990s. Yet at the time, there was no cloud, no SaaS, no animated Kanban cards. Gartner reaches the same conclusion on its side. Large IT projects are judged failures by their own sponsors in nearly 80% of cases. What is more, the probability of failure rises by about 50% as soon as a budget exceeds one million dollars. Three decades of frenzied software innovation, then, for an unchanged score. Instead, a new version of the tool was released.

The conclusion is hard to avoid. If tooling had been the cause of the problem, its explosion should have improved things. It did not. The problem is therefore elsewhere.

The original sin of project management software: everything starts from the task

To understand this failure, we must go back to the elementary building block of this software. This conceptual atom has a name: the task. A task is a thing to do. It is assigned to a person, given a deadline, then dragged from one column to another. Scratch the surface of any project management software on the market. You will always find the same hard core. Every time, it is actions, assignees, dates. And always the same gesture: dragging a card to "Done".

This choice is anything but trivial. It is a worldview disguised as a neutral tool. Building software from the task amounts to laying down a postulate: managing a project would consist of organizing a sequence of actions entrusted to team members. Project management is then reduced to plumbing. It is clean, reassuring, perfect in a sales demo. Above all, it is wrong.

The task, or the art of looking at the finger pointing at the moon

The task is the surface of the project. It is never its substance. It is what can be seen and measured easily. It looks good on a dashboard. The tail lights of a car tell you about the traffic, but nothing about the engine. Thus, you can steer an entire organization by watching cards change columns. And miss the essential entirely. Because the essential hides in three questions. Have we defined together what "succeeding" means? Who decides in case of disagreement? And who answers for the result, and to whom? None of these questions fits in a task. All of them belong to another register. Yet almost all tools have relegated it to the rank of accessory: governance.

What a project really is: a contract, roles, and a single accountable person

A project is not a choreography of tasks. It is a contractual relationship. And this holds, importantly, even when it is internal to the company. Take an example. A business department entrusts a development to an IT department, or two teams commit to each other. Every time, the same thing takes shape. First an agreement on an objective. Then a division of roles and responsibilities. And above all accountability. Yet software has forgotten this word in its dictionary.

This vocabulary is not a consultant’s affectation. It has a name, a grammar, and decades of practice: the RACI matrix. For each deliverable, four roles are distinguished. The R does the work, the C is consulted, the I is informed. And above all, the A is accountable. It is the A who validates, who arbitrates, who guarantees the resources and who answers for the result. The golden rule fits in one sentence. There is one, and only one, A per line. When accountability is shared among everyone, nobody carries it any more. It is mathematical. And it is deadly.

Accountability, the word your project management software ignores

Run the test with your favorite software. Ask it who is accountable for next week’s milestone. Not who has the task, but who answers for it. Ask it also where the decision of the last committee is recorded. On what criteria, and validated by whom? At best, you will get a comment in a card. And minutes that will get lost in emails. Because these tools were designed to orchestrate execution. Never to embody governance. So they can tell you who does what. They cannot tell you who decides, who validates and who answers. In short, they manage the agitation, not the commitment. Yet the causes of failure all live on the governance side. We have therefore built highly sophisticated machines. But they solve a problem that never arose. That is precisely what building backwards means.

Why the best-equipped structures are also the ones that fail the most

If project management software were the key, the best-equipped organizations would be the highest performing. Yet the opposite happens. According to the Standish Group, the success rate collapses as the company grows. In large groups, it barely reaches 9%. Yet these structures are full of certified project managers, colossal budgets and PMOs. Likewise, mid-sized companies set a record for abandonment, around 37%. In other words, the more means, method and software you have, the more you fail.

Gartner observes the same mechanism. A project of more than a million dollars fails far more often than a small one. Yet it benefits from more resources and more supervision. McKinsey drives the point home on large IT programs. Indeed, 17% of them threaten the very existence of the company. And the value actually delivered remains more than half below the promise. The lesson is clear. Adding execution power to a poorly governed setup does not save it. On the contrary, it makes it fail at greater cost and on a greater scale. All with impeccable traceability of its own shipwreck.

Proof by absurdity: Louvois, Berlin and the ghosts of steering

Statistics convince the head. Stories convince the gut. Here are two that no budget will ever make up for.

The first is French. Its acronym is almost comical given the disaster: Louvois. It stands for "Logiciel Unique à VOcation Interarmées de la Solde" (a single pay software for all the armed forces). It was meant to pay the military of all three services. Put into service in 2011, it immediately sank into surrealism. More than 160,000 soldiers had their pay wrongly calculated. Some received 90 euros a month. Others received fabulous sums, which would be reclaimed from them years later. But the money had already been spent. In total, the overpayments exceed half a billion euros. At the end of 2013, the ministry therefore announced the outright abandonment of the system. But the most instructive thing is not the amount. It is the diagnosis. Indeed, the parliamentary reports do not point to a bug. They point to failures in the steering of the project. Concretely, a multitude of departments shared the subject. No authority was clearly accountable for the overall result. And nobody could even stop the project, for lack of a pilot to decide. Thus, the code merely made visible a governance bankruptcy, on a national scale.

Berlin, or when project management software has nothing to do with it

The second story is German. And it has a formidable virtue: it contains not a single line of project management software. Berlin-Brandenburg airport was due to open in 2011, for about two billion euros. In the end, it opened on October 31, 2020. That is nine years late. In the meantime, six opening dates were announced and then abandoned. What is more, the bill doubled, then tripled, to exceed six billion. A former manager was even convicted of corruption. Moreover, the airport’s chief admitted that the project had made Germany a laughing stock. A "solution" was even considered for the fire protection system. It consisted of posting hundreds of people on stools, phone in hand. Because when governance collapses, no technical feat fills the hole. You only invent ever more humiliating patches. Everyday business replays the same scenario, in a quieter voice. Think of the contractor whose fixed-price projects end up in court. Or the bank that piles up "task forces" without ever touching the management.

The hidden cost of the illusion: what nobody puts in the business case

We like to believe that the cost of a failed project stops at its budget. Charming naivety. The Consortium for Information & Software Quality has put a figure on the cost of poor software quality. In the United States alone, it reaches about $1.82 trillion a year. Part of it comes from the failures themselves. But most of it comes from the operational failures that follow. In short, a failed project is not a line crossed out in a spreadsheet. It is a debt that runs through the organization for years.

And then there is the cost that nobody writes down in black and white: the human one. Gallup measures it in its State of the Global Workplace 2026. In Europe, the employee engagement rate peaks at 12%. It is the lowest in the world, behind every other region. What is more, disengagement costs the global economy the equivalent of 9% of GDP. Yet projects that go off the rails are precisely what produces disengagement. There is overload, simmering conflict, the feeling of absurdity. And weak signals that nobody picks up in time. Moreover, human drift always precedes milestone drift. But it happens out of the field of view of the software, which is looking elsewhere.

Agility will not save you (and neither will the next license)

At this point, the same objection always arises. "We do Agile." That is cute, and it is not enough. The PMI shows this in its Pulse of the Profession. Predictive, agile and hybrid approaches broadly perform alike. In other words, the massive adoption of agility has changed nothing in the success rate. Because the method, like the tool, is only an instrument. Applied to the wrong problem, it does not fix it. On the contrary, it ritualizes it.

Should we then throw the software out of the window? Not that either. PWC observes a clear correlation. 77% of high-performing projects rely on project management software. Only 23% do without. Tooling therefore matters. But it matters as a servant of governance, never as a substitute. Thus, what correlates with performance is the tracking of decisions and traceability. It is also cross-functional steering and attention paid to the human side. It is not the umpteenth drag-and-drop feature. In short, neither method nor tool makes a project succeed. Governance does.

Neither project management software nor AI will replace the pilot

What remains is the big promise of the moment. Artificial intelligence, at last, would settle the question. Copilots, automatic minutes, predictive dashboards, agents that send reminders on your behalf. It is appealing, and sometimes useful. But AI rests on the same inverted foundation. It speeds up the surface. It moves the card faster. It summarizes the meeting and anticipates the schedule slip. However, it does not govern. And for a simple reason: it cannot.

Accountability cannot be delegated to a model

Because accountability (the "A" of the RACI matrix) is not a piece of data to be computed. It is a moral and contractual position. Being accountable means answering for the result. It means arbitrating between opposing interests, and bearing the consequences in front of the customer. And you do not delegate that to a model. A model signs nothing, risks nothing, and is dismissed by nobody. Gallup puts it bluntly. Winning the AI revolution will depend less on technology than on the quality of leadership. Researchers at Stanford, Harvard and MIT had already measured it. Management practices weigh more heavily than technology on productivity.

The misunderstanding, from the start, comes down to one word. Task management has been confused with real project management. The former, software does well, and AI will do even better. The latter is governance management. Concretely, that means framing an objective and distributing accountabilities. It also means reading weak human signals. It means, finally, holding a go/no-go and defusing a conflict in time. And that cannot be bought under license. It is acquired in the field, in projects that are on fire. In short, it is a rare skill, expensively paid for, and irreducibly human.

In incompetent hands, the most powerful tool does not erase chaos. On the contrary, it industrializes it. It then produces superb dashboards of a sinking ship. Just faster. Hence an uncomfortable and liberating conclusion. No project management software, present or future, will steer your projects for you. The pilot is, and will remain, human. That is therefore where to invest. Not in the next license.

Sources

Reference studies and data

Costs, management and cases cited

Ready to improve your operational performance?

A first strategic interview, free and with no commitment.

Free strategic interview