by Stefan Degen (ERNI Switzerland)
I see it regularly: a project is completed. On time, on budget, within the agreed scope. The team is satisfied, and the client has given the green light. And then, a few months after go-live, it turns out that the product isn’t delivering the results everybody was hoping for. Customers are hesitant to use it. Internal processes work exactly as they did before. The promised efficiency gains remain untapped. The project was a success. The product was not. This contradiction is not an exception. It shows up in many digital initiatives, across industries and types of organisation.
Project logic and product logic are fundamentally different
A project measures its success against three parameters: time, budget and scope. These are legitimate criteria for project management. At the end, ‘delivered’ counts as success: project closed, budget met, box ticked. But delivered does not mean successful.
No user clicks for the first time and wonders, ‘Was this delivered on time and on budget?’ A user asks, ‘Does this solve my problem? Is it easier than before? Does it make my everyday life better?’
A product measures its success by something else: Is it being used? Does it solve the right problem? Does it create value for customers, users or the organisation itself?
These two logics do not contradict each other, but neither does one guarantee the other. A project can be executed perfectly and still deliver a product that doesn’t work in the market or within the organisation.
The bottleneck lies before implementation
Why does this happen? Often not because of the implementation itself. Experienced engineering teams can only work effectively if the problem, the target vision, the user context and the technical context and constraints are clarified early enough. The problem lies in the phase before: in the clarification that is either skipped or rushed.
People know the industry. They know the internal processes. They have experience, so they believe they know what is needed.
Industry knowledge is no substitute for validation.
What flows into the requirements as a self-evident assumption is often untested. Who the target group actually is. Which problem really needs to be solved. Whether the intended solution works in everyday use: not in the workshop, but in the real situation in which it will be used.
A concrete example: a company is planning a digital service portal for its internal employees. The requirements have been drawn up and the roadmap is in place. Only in conversations with the future users does it become clear that the problem isn’t the lack of a digital channel, but unclear responsibilities between two departments. The portal would be built. But it would not solve the underlying bottleneck.
When assumptions meet implementation
What happens when untested assumptions flow into an implementation project? At first, nothing obvious. The team starts developing and the first sprints go well. Then the first contradictions emerge. A requirement turns out to be unclear. A user group behaves differently than expected. An integration proves more complex than anticipated.
Now corrections are needed. New stories have to be written. Dependencies shift. What starts as a minor adjustment ends up costing weeks and budget. And at the end of the project? There is a product that works technically, but doesn’t solve the actual problem.
AI accelerates the problem, not just the solution
Anyone who believes generative AI will solve this problem is overlooking a crucial mechanism. AI makes development faster. Considerably faster. Prototypes are created in days, and code is generated and tested more quickly. This is real and significant. But building the wrong thing faster is no victory. The damage occurs sooner and on a larger scale.
The real strength of AI in product development lies elsewhere: in the early phase. AI can analyse user data, recognise patterns in feedback and visualise initial concepts quickly. Those who know what should be built, and why, can take full advantage of the speed of AI-assisted development. The bottleneck remains clarity, not speed.
What helps: Three questions before implementation
There is one fundamental change in ways of working that makes the difference: deliberate clarification before the main implementation team gets started. This doesn’t mean spending six months on research. It means answering three questions robustly:
- Why this problem? Is the problem to be solved actually the relevant problem for the target group, or merely the most obvious one? Validated with real feedback, not internal assumptions.
- Why this solution? Does the intended approach work in practice? Not on paper, but tested with real users in a real or simulated situation. Several options explored, the most promising one validated.
- Why is it feasible? This is where the bridge between product thinking and engineering is built. A good product decision takes into account not only user needs and business objectives but also architecture, data flows, interfaces, operations and scalability. Otherwise, what looks convincing on the whiteboard fails because of legacy systems, missing APIs or operational requirements, and this usually only becomes apparent once implementation is already under way.
Those who separate and examine these three questions before implementation don’t reduce the pace. They reduce the guesswork.
The first real value is created before go-live
When this phase is taken seriously, something tangible emerges: a validated concept based on real feedback. A basis for decision-making that allows stakeholders to act on insights rather than assumptions. And a clearer starting point for implementation: the team knows from day one what it is building and why. This is the first real value of a digital initiative. It is created before a single line of production code is written.
This requires interdisciplinary collaboration: business, the user perspective, architecture and engineering need to come together earlier, not only once implementation is already under way. Those who answer these three questions early don’t start more slowly. They start with fewer assumptions, better decisions and an implementation team that knows what matters.
No time for clarification?
The obvious objection: there’s no time for upfront clarification, because the organisation expects results. But the argument works the other way round. The corrections that result from untested assumptions cost many times what the clarification would have required. If you don’t have time for clarification, you certainly don’t have time for the detour.
Conclusion
Completing projects successfully is important. But it isn’t enough. Organisations undergoing digital transformation build products and services, not projects. The difference doesn’t lie in the method or the team. It lies in the question asked at the outset.
Not: ‘How do we implement this?’ but: ‘What should we actually be implementing, and why?’ Clarifying this question early and in a structured way is not a delay. It is the step that accelerates everything else. To close, a question every organisation can ask itself: with your current digital initiative, do you know whether you are solving the right problem – or are you just assuming you are?
About the author
Stefan Degen is an Expert Digital Product Architect at ERNI Switzerland. He combines product strategy, the user perspective and technical feasibility to clarify digital initiatives early and turn them into effective products, from the initial idea through to the implementation decision.
Do you have a digital initiative that isn’t quite clear yet? Get in touch.