Open innovation: why 3 in 4 pilots fail to scale

A 500 Startups survey of more than 100 corporate innovation executives found that the vast majority see fewer than 25% of their initial startup pilots reach the market. The study dates from 2017, but a similar ratio shows up in Medellín. In April 2023, Ruta N’s Open Innovation service reported that since 2020 it had published 140 challenges, connected 113 challenger companies with 867 solution providers and closed 35 successful business cases. The usual explanation blames slow corporations or immature startups. In The Seat · EP 08 — Juan Manuel Bottía / Las buenas ideas son las que menos amigos tienen (episode in Spanish), the Startus podcast hosted by Matt Carrasco, Bottía offered another one, based on a mistake he admits was his own.

Where do open innovation programs get stuck?

Most challenges never become a business. The global figures and the Colombian ones come from different contexts, but they agree on the order of magnitude.

Our reading: 35 business cases out of 140 challenges is a one-in-four ratio. 500 Startups notes that more than ten departments are involved in each startup deal, and that the people, goals and resources involved need to be aligned early. Before the first prototype, someone inside has to want the solution.

The mistake Bottía admits was his own: opening challenges nobody inside cares about

Matt asked him what he would recommend to someone in a role like his who is trying to get open innovation off the ground and can’t. Bottía didn’t talk about budgets or methodologies:

“Let’s identify the issues the organization genuinely wants to solve, because here’s what happens: if you start from a topic that isn’t attractive enough, the organization’s internal teams won’t dedicate time to working on it. And it has happened to me.”

He explained how it plays out. Eager to show that a program is running, the innovation team launches topics that interest the innovation team. Then the department that actually knows the process finds the solution too generic. At the end of the episode, when Matt asked what advice he wishes he had received when he started, he came back to the same point: “the most important advice is that the topics you open to the outside should be topics that matter to the people inside.” He said that early on he chose topics without knowing the business operations in detail, and the owners of those problems showed him why the solutions weren’t strong enough for them.

A challenge with no internal owner gets no time, even with a good solution

Bottía closed that answer with the line that sums up the argument: “people dedicate time to what matters to them, and not to the rest.” For him, that is why you have to work with the problem owners from the start and bring them into the process.

Some programs already build this into their design. Ruta N helps selected companies define and structure their challenge, and asks them to appoint a team of at least three people to work through the methodology, define the challenge profile and attend the meetings with potential solution providers.

KPMG’s Colombia Tech Report 2024–2025, published in April 2025, asked 103 corporations about their main challenges when working with startups. The top one was the time internal processes take (40%). Tied for second place, at 39%, were aligning priorities across the departments involved and understanding startups’ business models. The same report points to a paradox: 80% of corporations say they have strategic alliances with startups, but more than 90% had fewer than five startups among their suppliers in 2024.

Our reading: an internal owner is what gives a pilot time, information and someone to defend it when priorities are set. A good challenge doesn’t make up for not having one.

Counting hackathons is not the same as counting implemented projects

Bottía also talked about the risk of measuring innovation by volume of activity. He admitted it can happen in his own work: “you fall into activism, into doing for the sake of doing.” As examples of those metrics he cited saying you delivered 100 hours of training, ran so many hackathons or received so many solutions.

He doesn’t dismiss that stage: in his view, at the beginning you need to generate activity. What he proposes is a progression. As the program matures, the question shifts from how many ideas came in to how many projects were implemented, and then to what impact they had. He summed it up with a phrase he uses internally: “we don’t work in order to innovate, we innovate in what we work on.”

He speaks from a measurable track record. According to what he said in the episode, over five years Colsubsidio went from 26th to 11th to 4th in the ANDI innovation ranking, reached first place and placed sixth this year. He made it clear he never set out to be number one: “what I set as my goal was to improve innovation management.” The edition presented on May 19, 2026 measured 303 companies across 50 subsectors and 13 departments. Corona Industrial came first, followed by Alpina and Fundación Cardio Infantil.

For startups, the first step is understanding the challenge, not pitching the product

The argument has a mirror image on the startup side. Bottía’s advice: “A startup should come to understand my challenge and my problem, and from there show me it has the capability and experience to help me build a solution.” In his view, many startups make the mistake of assuming that what worked with one client will work with all of them.

He also described a pitching problem. Many startups present to a corporation the way they would to an investor: they talk about the market, the competition, their rounds and how much they’ve raised, but they don’t explain how the solution helps the corporation. His explanation of why that doesn’t work: “The corporation isn’t necessarily after a financial return.” In his view, a corporation invests when the startup helps it reach a strategic goal of its own.

The advice carries weight in Colombia, where most startups sell to businesses. According to KPMG’s Colombia Tech Report 2024–2025, 90% of the startups in its sample have a fully or partly B2B model, and almost half are B2B only. Ana María Mayorga, from KPMG, talked about that report on The Seat.

Our reading: if the challenge has no owner inside, even the best pitch finds no one to convince. Both recommendations are the same idea seen from each side of the table.

The open question

If one in four relationships becomes a business, how many open innovation programs track how many of their challenges had an internal owner before they were published?

The next guest on The Seat

Bottía offered to make the introductions and recommended five people and organizations: Óscar Emir Rincón, from the innovation team at Alpina; Carla Podestá, who led the innovation vice presidency at Grupo Nutresa; Juan Camilo Caicedo, from Fundación Cardioinfantil; ALSEC, a company based in Medellín; and Tatiana Acero, from Ecopetrol. Alpina and Fundación Cardio Infantil ranked second and third in ANDI’s 2026 Business Innovation Ranking.