How two open innovation programs revealed what's disconnected within.

Open innovation programs almost always launch pointing outward: Find startups. License technology. Bring in what R&D doesn't have.

It's a clean mandate. It's also incomplete.

The real bottleneck to scaling open innovation is rarely the outside world. It's the organization's own systems.

Procurement workflows that exhaust researchers before a single partner is signed. Partnerships scattered across an R&D network that nobody is tracking. Solutions one division already built that another division is still spending years trying to solve.

Kyle Basler-Reeder was hired as a one-person open innovation function at ExxonMobil. His job was to help researchers find external expertise. But researchers had already tried that — and given up. The procurement process was so slow that most scientists would rather solve problems themselves, even knowing a better answer existed somewhere outside.

Bill Roschek Jr. inherited a narrow licensing function at Nestlé Purina. When he started, leadership estimated 30 to 50 external partnerships across R&D. The actual number was closer to 200, with no single team tracking them.

Both were hired to search outward. Both spent their first year fixing what wasn't working internally. And the programs that emerged look nothing like what either organization asked for.

ExxonMobil is one of the world's largest energy companies, with thousands of researchers across dozens of technical disciplines. By 2021, leadership recognized that the volume of external expertise — startups, freelancers, academic research — was growing faster than any internal R&D function could match.

The company created a new role: External Leveraging Coordinator. Kyle Basler-Reeder, a physicist by training, was the sole hire. No team, no tools, no budget for software. And a research culture where scientists had spent their careers solving problems with internal resources — a way of working that had been the norm for decades.

Rather than launching projects immediately, Kyle did something unusual for an open innovation hire. He spent his first months listening.

He ran over 200 one-on-one interviews: 50 with external open innovation practitioners and 150 with people inside ExxonMobil. Engineers. Researchers. Legal staff. He organized every finding by frequency, tracking how many times each issue appeared across conversations.

One signal was louder than everything else. Procurement.

The problem wasn't that researchers rejected external collaboration. Many had tried — joint research, technology licensing, contract work. But the internal procurement process, combined with legal reviews and risk assessments, created so much friction that most researchers had given up. The forms. The reviews. The waiting. The back-and-forth with departments that didn't understand why R&D needed a startup partner in the first place. Researchers described the experience as painful enough that they would rather solve the problem internally, even when they knew a better solution existed outside.

That friction shaped what the program would become.

Kyle used the interview findings to map the internal barriers and build a working knowledge of how to navigate them: which procurement paths moved faster, how legal reviews actually worked, and where past attempts had stalled for fixable reasons.

That examination revealed a pattern worth paying attention to. One research team told Kyle that a request-for-proposals process had failed years earlier. The team's conclusion: RFPs do not work at ExxonMobil.

Kyle investigated. The RFP itself — the scope, the specifications, the outreach — had been well designed. The execution had stalled at a specific procurement barrier. A fixable one. But the team's institutional memory had compressed the entire experience into a single lesson: RFPs don't work here.

Kyle found versions of this across the interview data. Teams carried conclusions about what had failed, but those conclusions often hid the actual reasons for the failure. Untangling the two became one of the program's first and most important jobs.

Where Kyle started from zero, Bill stepped into a function that already existed. The Innovation Partnerships team had been running for years, enabling joint development efforts with startups and external partners to support new product development. The mandate was focused on a single activity: joint development with startups.

Bill's early work revealed the true scale of something leadership hadn't fully seen. Across the R&D organization, there were more than 200 active partnerships spanning suppliers, startups, academic institutions, and collaborations across broader Nestlé businesses. No single function had visibility into the full portfolio. Finance, procurement, and operations each tracked their own slice — but no one had assembled the complete picture.

This reflected a strong culture of external engagement. It also meant that the full value of those partnerships was going unmanaged.

Before expanding the team's mandate, the focus was on getting a handle on what already existed. The initial estimate — that 30 to 50 partnerships needed better coordination — dramatically understated reality. Once the team assembled the full picture, the results were immediate: efficiency improved, partner relationships strengthened, and for the first time leadership could see the actual scope of what the organization was doing externally.

In hindsight, both programs were initially designed to help R&D look outward. Both evolved to include strengthening internal enablement to fully realize the value of external collaboration.

With a playbook built from 200 interviews, Kyle's program faced its next challenge: finding research teams willing to try open innovation when the function had no results to show and no mandate requiring participation.

The only KPIs that made sense at this stage were basic: how many teams had the program reached, and how many had launched a project. No financial returns existed yet. Promising any would have been premature.

The teams that engaged earliest shared a trait: they were new. New business units exploring unfamiliar technical areas. Recently formed research groups that hadn't yet built deep internal networks for sourcing solutions. These teams recognized the gap in their own capabilities and were willing to try an unproven function.

Established teams with decades of internal expertise were harder to reach. Their existing workflows and supplier relationships already worked well enough. Adding an unfamiliar open innovation process felt like overhead, not help.

To build credibility with skeptical teams, Kyle needed proof that open innovation could work at ExxonMobil. He didn't have his own results yet.

So he went looking for other people's. He searched across the company for past external engagements that had produced results — some of them years or decades old. A crowdsourcing challenge that had surfaced a useful technical solution. A partnership with an outside research group that had shortened a development timeline. He stripped the dates from these case studies and presented the outcomes in team sessions.

But as teams started reaching out, Kyle faced a different challenge: each stakeholder had a different problem, a different level of urgency, and a different idea of what "external help" meant. Some needed an expert opinion in 48 hours. Others needed to scan an entire field they had never worked in. Others needed a ready-made solution they could pilot immediately.

Kyle relied on three tools to cover the range — each designed for a different type of problem and a different level of urgency.

This scenario comes up often: a business unit moves into a new technical domain, and the expertise to evaluate options, identify risks, or even frame the right questions doesn't exist inside the organization. Open talent solves for that gap by connecting internal clients with external specialists — freelancers, academics, or independent consultants — for short, focused engagements.

The approach Kyle uses most often is the "Hollywood model": assemble a small panel of world-class experts, typically around ten, for a half-day workshop. The internal team presents the problem as they understand it. The experts debate in the room. Months of reviews and exploratory research compress into a single session.

Open talent is the right fit when:

Tip: Ask the internal team to name the three questions they cannot answer with existing expertise. Those questions become the brief for the expert search.

Open talent brings known experts to a known question. Global crowdsourcing works differently. Instead of selecting experts, the innovation team publishes the problem on a public platform with a cash prize and opens submissions to anyone, anywhere. The reach can extend to millions of potential solvers.

Most submissions are unusable. The value comes from outliers: solvers from adjacent industries who recognize a version of the problem they've already worked on. In Kyle's case, that meant medical researchers responding to an energy challenge, or aerospace engineers spotting a problem they'd solved in a different context.

Crowdsourcing is the right fit when:

Crowdsourcing also produces a less obvious form of value. A global challenge that returns no viable solution should not be seen as failure — it should be seen as a definitive scan of what exists externally. When that scan comes back empty, the internal team can stop wondering whether someone else has already solved the problem. That clarity changes decisions and helps reallocate resources.

One research group at ExxonMobil had been working on the same technical problem for 85 years. The same goal had been passed down through successive generations of scientists. The open innovation team ran a crowdsourcing challenge: put the technical requirements in front of a global pool of solvers across industries and disciplines.

Nothing came back. No breakthrough. No adjacent-industry insight. No viable alternative.

That result was one of the most valuable outcomes the program produced. For 85 years, the project had continued because no one could definitively say that a better approach didn't exist somewhere outside. The crowdsourcing challenge provided that evidence. The research group used the result to recommend stopping the work and reallocating resources to other projects.

Permission to stop is a deliverable too.

Tip: Start with a one-page problem statement: what the team is trying to solve, what constraints matter, and what a useful response looks like. That document becomes the challenge brief.

The third tool closes the gap between concept and implementation. The innovation team distributes detailed specifications to a curated list of 10,000 to 20,000 institutions and asks for proposals, not concepts.

The distinction from crowdsourcing matters. Crowdsourcing surfaces directions and ideas. A global RFP returns solutions already developed elsewhere, partnership proposals with timelines and concrete approaches, and options that accelerate programs and reduce costs.

This format suits internal clients under time pressure. A program that loses its two-year runway and needs results in six months can shift from building internally to scanning globally — testing whether something ready-made already exists before committing to a long build.

Tip: Before launching a global RFP, sit with the internal client and translate the need into a document detailed enough that an external institution can respond with a concrete proposal, not a pitch deck.

With these three tools, Kyle's program gained traction quickly. Open talent workshops gave new business units confidence in unfamiliar technical areas. Crowdsourcing challenges surfaced solutions from adjacent industries. Global RFPs accelerated programs under time pressure. Within 18 months of launching, the program had engaged over 100 research teams and generated more than 250 open innovation projects.

Then, the data revealed something the program had not been designed to find.

Over 40% of those 250 projects did not end with an external solution. Instead, the process of searching externally — surveying what startups, universities, and freelancers were working on in a given technical area — surfaced solutions that already existed inside ExxonMobil. One division had solved a problem that another division was still working on. A technology developed for one application turned out to be relevant to a completely different research group. The solutions existed. They had just never been connected to the teams that needed them.

The open innovation function, with an original mandate to look outward, had accidentally created an internal knowledge map.

At Nestlé Purina, the evolution followed a deliberate progression: building on a strong foundation of external collaboration and scaling it into a more integrated, strategic capability.

Bill's first step was to build coordination and visibility across an already extensive partnership ecosystem. The team deployed a Microsoft Form that every R&D employee filled out when initiating a new external agreement. The form connected to legal and document retention systems and fed into a Power BI dashboard, giving the open innovation team and leadership a unified, real-time view of every external relationship across the R&D network.

The results were immediate: within the first 12 months, the system captured over 180 new agreements, ranging from development partnerships to academic collaborations. For the first time, leadership could see the full scope of external engagement.

But tracking was the foundation, not the destination.

Nestlé Purina's open innovation function evolved through three phases, each expanding the team's mandate:

Since implementing this model, a significant share of R&D projects involve external partners, and external collaboration is expected to contribute meaningfully to future revenue growth.

Both programs started as external sourcing functions. Neither stayed that way.

As the program evolved, several structural enablers proved critical to sustaining impact:

This addressed a vulnerability the program had encountered directly. In the earlier model, the team signed the deal and handed the partnership to an internal project team to manage independently. Projects often stalled at that handoff, because the project team lacked context on the external partner's capabilities and the original intent of the deal. When a single champion left, projects often died with them. The dual-lead model keeps both continuity and context intact through commercialization.

The program could surface a promising solution in a matter of weeks. Getting that solution from a slide deck to a pilot — through the procurement queue, past legal review, and into a team's actual workflow — was a different problem entirely. A harder one.

Kyle addressed the barriers he could control. For procurement, he built a playbook of workarounds, legal tips, and process navigation guides, maintained and updated as new barriers surfaced. For champion turnover, he ensured that project context was documented and shared beyond the individual driving it.

Even with those systems in place, the experience after sourcing varied widely. Some teams found a solution and immediately began testing, moving so quickly that the open innovation function only heard about progress secondhand. Others stalled for months because the pilot required $200,000 — more than any internal grant could cover, and the team had never built a funding case for senior leadership. Some stalled in the legal process: the team wanted to move forward but didn't know how to structure an NDA or development agreement with an external partner. Some simply lost momentum as competing priorities absorbed the team's time.

As the insular culture shifted gradually, something unexpected happened. Champions began running presentations without Kyle's involvement. In one case, a researcher took the program's materials, adapted them for a different research team, and presented them to a group of 20 colleagues. Five new project leads came out of that single session. Other champions did the same across different parts of the organization.

The program had started replicating through informal networks.

As both programs matured, they embraced the opportunity to define how value from open innovation is measured at each stage of development.

At ExxonMobil, this led to a deliberate evolution of KPIs — ensuring that metrics aligned with the maturity of the program and the nature of the work being done. At Nestlé Purina, a tracking infrastructure was established that provided leadership with clear, actionable insights at every phase, from early engagement through to business impact.

In both cases, the focus was on building a measurement approach that reflects how innovation actually progresses. Capturing leading indicators early. Scaling toward financial impact over time. Rather than over-indexing on immediate financial outcomes, both programs emphasized transparency, credibility, and stage-appropriate metrics to demonstrate progress and build confidence.

Two programs, two starting points, one pattern:

There is a quiet paradox at the core of both of these stories: the function built to look outward becomes most valuable for what it reveals inward.

At ExxonMobil, 40% of projects surfaced solutions that already existed inside the company. At Nestlé Purina, the first act of building external partnerships was discovering 200 relationships no single team had visibility into. Both programs delivered on their external mandates — but the unexpected payoff was forcing the organization to see what it already had.

The organizations that will get the most from open innovation over the next five years are the ones that treat the function not as a sourcing channel, but as a connective layer: one that maps internal expertise, coordinates fragmented partnerships, and translates between R&D and the commercial side of the business simultaneously.

The outside search begins with a walk through your own building first.

Most organizations pick their open innovation engine the wrong way.

They copy what a peer company launched. They go with whatever generated buzz at the last conference. They choose the model before they've honestly assessed what their own organization actually needs.

Diana Joseph, author of Making Open Innovation Work, calls this engine-reality mismatch — and it's the second most common reason open innovation programs stall, right behind internal alignment failures.

Her fix is a forcing function: score seven dimensions of your organization's reality before you look at a single engine profile. The sequence matters. Scoring first prevents reverse-engineering the answers to fit a model you've already fallen in love with.

Score each dimension 1–5 before reading further:

These dimensions aren't independent — and the tensions between them are where the real work happens.

A program that scores high on both visibility and learning is committing to exploring unknown territory in public. High corporate control limits the diversity of partners willing to engage; startups and academic institutions rarely accept heavy governance. High speed compresses the time available for learning — so an organization that needs fast financial returns and deep capability-building from the same program may be setting contradictory goals.

These aren't problems to solve. They're design constraints to acknowledge — before you pick anything.

Once you've scored all seven dimensions, hold that picture up against each engine, one at a time. Diana defines six engine types in Making Open Innovation Work:

Each engine sits at a fixed position on the same seven scales. Place your scores next to an engine's profile, dimension by dimension, and look for gaps. The book provides a physical overlay tool for this (the cover is designed for dry-erase marker; the gatefold folds over so the two sets of scales line up side by side), but a table with two columns works the same way.

Example: overlaying against the incubator

Corporate control and cultural impact are mismatches in the wrong direction: the organization needs far more of both than an incubator provides. Running the same comparison against the accelerator and prove-out journey produces a different pattern of gaps each time.

When no single engine matches cleanly, the answer is often a hybrid. An accelerator for the cohort experience combined with prove-out journeys for deeper partnership work is one combination that appears regularly.

The engine comparison clarifies which model fits. But the mismatches are just as valuable — they become the starting points for the alignment conversation that needs to happen before any program launches.

Not every mismatch is a problem. Diana distinguishes between a mismatch in the wrong direction (the organization needs high control but the best-fit engine requires low control — a genuine misfit that forces a choice) and a mismatch in the correct direction (the organization has more resources than the engine requires, which is comfortable surplus). Reading the gaps this way keeps the conversation focused on the dimensions that actually need resolution.

The exercise works two ways:

As a workshop — get enough of the right people in a room to score the dimensions together. The conversation surfaces disagreements on specific dimensions rather than producing a general endorsement of "innovation."

As an informal test — draw the seven scales into a one-on-one with a sponsor or business unit leader and walk through the positions together. That conversation tends to land better than a slide deck, because it asks for commitments on specifics rather than general enthusiasm for "innovation."

The mismatches will surface eventually. This exercise determines whether they surface in a planning conversation — or in a quarterly review after the first program stalls.

That’s it for today.

Curious to hear how Kyle and Bill’s experience compares to your reality. Next time, we’ll look at why your strongest innovation capability might be your most expensive blind spot.

Hans Balmaekers
Founder, the Compass and Chief @ Innov8rs

PS- want us to cover a particular topic, or would you like to be featured sharing your approach and lessons learned? Keen to hear from you.

PPS- feel free to forward this newsletter to all the innovation leaders in your company and network. Sharing is… indeed!