I have managed digital product portfolios that generated close to €2 billion in net sales. I have also inherited teams that were extraordinarily busy — full sprints, packed backlogs, constant releases — with almost no visibility into whether any of it was actually working.
These are not contradictory experiences. They are, unfortunately, common ones.
The most persistent failure mode I see in digital product organizations is not poor execution. It is the absence of a coherent operating model that connects daily delivery to long-term commercial outcomes. Teams produce. They ship. They move fast. But they are building in the dark.
Vision first. Everything else follows.
Before a single sprint is planned, a product team needs to be able to answer three questions with precision: where are we going in three years, what does success look like at the end of this year, and what are we specifically trying to move in the next ninety days?
These are not the same question. The three-year vision gives the team a direction — it is the north star that makes prioritization decisions coherent rather than reactive. The annual objectives translate that vision into concrete milestones that the business can hold the team accountable to. The quarterly goals are where the work actually happens: specific, measurable, achievable in a defined window, directly connected to the layer above.
Without this structure, roadmaps become wish lists. Priorities are set by whoever asks loudest. And when something ships, nobody can tell whether it mattered.
Measure everything. Especially the things that feel obvious.
Releasing a capability is not the finish line. It is the beginning of the question: did this actually do what we thought it would do?
I have seen features that took six months to build, adopted by fewer than 10% of the users they were designed for. I have seen UX changes that seemed minor produce dramatic improvements in completion rates. The difference between knowing and not knowing is measurement — and the discipline to build measurement into the definition of done, not as an afterthought three months later.
Data does not replace judgment. But judgment without data is just opinion. The best product teams treat instrumentation as seriously as development. They define their success metrics before they build. They review those metrics after release with the same rigor they applied to the build itself. And when the data says something inconvenient, they act on it.
Cross-functional is not a process. It is a culture.
The other failure mode I see consistently: product teams that operate in isolation, handing work over to commercial, IT, or market teams rather than building with them. The seams between functions are where value leaks. It is where assumptions go unchecked, where requirements get lost in translation, and where perfectly built products fail in adoption because nobody owned the change.
The best outcomes I have been part of happened when the product team, the commercial team, the data team, and the market teams were working toward the same objective — with shared visibility, shared accountability, and a genuine understanding of what each function needed from the others to succeed.
On simplicity, speed, and AI
One more thing that is non-negotiable in 2026: every digital capability you build must save the user time. That is the baseline expectation. Nobody has patience for complexity. Nobody reads instructions. If a feature requires explanation, it needs to be redesigned.
This applies to AI capabilities as much as anything else. AI is not a feature to add because it is expected. It is a tool to reach for when it genuinely reduces friction, improves a decision, or eliminates a task that was costing someone time they did not have. Used well, it is transformative. Used as decoration, it adds noise.
Build less. Measure more. Make it fast. Connect it to something that matters.
That is the operating model.