Fast Delivery or Deep Learning? Dilemma in Ecosystem Exploration Projects

Companies developing new technologies, services or business models face a core dilemma: push ahead fast enough to secure continued funding, or step back to properly understand what they are building – even if that threatens future support.

In our research, co-authored with Christophe Midler and Lorenzo Fioni (Ecole polytechnique) we show that the way these so-called “exploration projects” are split between partners can determine how fast companies move, and how much they actually learn.

In the study’s case project, a European carmaker handed the energy system for an electric mobility service in rural Africa to an external partner to move faster. But when that partner pulled out, the company spent 45 days rebuilding knowledge internally because the work had effectively become a “black box”.

To avoid that, companies need to actively manage so-called “project granularity”: how projects are carved into bigger or smaller pieces, and how responsibility is split between partners, including whether it is outsourced or kept in-house.

How you split the work shapes what you learn

At its simplest, it is a trade-off between speed and understanding.

Handing big chunks of work to external partners speeds things up and makes projects easier to manage, but leaves companies in the dark. Breaking work into smaller pieces slows delivery and makes the work harder to manage, but forces companies to understand what they are actually building.

That trade-off played out in real time in my case study. The carmaker hit problems between its solar power and battery systems, and had to work out how the whole system actually fit together.

That kind of setback is exactly what makes these projects so hard to sustain.

Where projects become vulnerable

Exploration projects are constantly at risk of being killed off because they compete with more profitable work, while carrying high uncertainty and costs and delayed returns. Managers often default to prioritising short-term performance, making them reluctant to back projects that may take years to pay off.

The problem is even sharper in so-called “ecosystem exploration projects”, where companies are trying to build new systems with multiple partners at once. Here, everything depends on whether partners stay in. When they walk away, projects can grind to a halt or collapse altogether.

There is no single “best” way to structure these initiatives. Companies have to keep changing how they are set up – pushing for speed when they need to prove progress and secure resources, then shifting towards deeper learning as problems surface.

Why standard approaches break down

Standard project management falls apart here. Exploratory work does not follow a fixed plan; it throws up problems, gaps and unknowns that force companies to adapt as they go. That is because these projects start without a clear end point.

Unlike conventional development projects, where the goal is clear from the start, exploration projects deal with unfamiliar technologies, markets and business models with uncertain commercial outcomes.  

In practice, companies are trying to make unfamiliar partners, technologies and ways of operating work together at the same time. That matters more as companies take on bigger, messier projects in areas like digital tech, electric mobility and clean energy.

And it is not just about delivering something at the end. These projects create two types of value at once: the product or service itself, and the knowledge companies build about how these systems actually work.

In the case study, the carmaker kept changing how the project was set up; first handing the energy system to an external partner, then breaking it into smaller pieces across several partners as problems surfaced.

Ultimately, the project was killed off after the war in Ukraine pushed management back towards short-term pressures. The work was too uncertain, too slow to prove itself, and too easy to cut – exposing just how fragile these projects are when conditions tighten.  

It is the dilemma in full: move fast and learn less, or slow down and risk being cut.