Daniela
Rubio Olaya

As Growth Manager I owned retention for a product with no built-in way to reward a returning customer: a vending machine. The only lever was the purchase data and app already in front of users.

1. 01What brings someone back, without paying for it directly?
2. 02What tone actually earns this audience's trust?
3. 03What works at every machine, with zero hardware spend?

Field research surfaced three buyer types: singles buying out of necessity, mothers who forgot lunch, families topping up a meal. All bought out of a specific, occasional need. The job was to make it a habit.

A reward on every purchase, built to pull toward the next one. Tom, a family-warm mascot instead of a youth-coded loyalty voice, since 80% of users were families. A streak-and-points system converting to cashback, running 100% in-app so it launched identically across every machine.

Alpina came to Rockstart to evaluate a venture investment in Europe. That assessment surfaced a bigger opportunity: running that same discipline internally. Rockstart brought me in to lead the shift, and I took ownership of redesigning Alpina's internal innovation process end to end.

Ten teams, roughly 350 people, had no shared way to prioritize new products, track them, or decide what happened next. Promising projects kept stalling at the finance review because teams and finance were not speaking the same language.

I built a structured methodology across every team: clear KPIs and timelines, plus a data-based reporting process and dashboards so teams could show a project's potential in numbers finance trusted. Working with the product team, I ran design thinking and agile sessions to turn existing assets into new seasonal products grounded in real user knowledge.

Same system as How I Work →

As co-founder I owned brand creation and positioning for a scientific, PCR-based fertility diagnostic sold through doctors and labs, without losing sight of the woman actually going through it.

Women didn't want technical terms or an explanation of what a PCR test was. They wanted to feel supported, met with empathy on a hard path.

Two approaches, one brand: a consumer voice built on feeling, and a clinical voice built on evidence.

Planning's process knowledge lived in scattered documents and specific people's heads, slowing onboarding and creating repeated questions, with real risk if someone left or changed roles.

A two-part system: a SharePoint built as the single source of truth for every process, and a Planning Agent inside Marcel AI, trained on that documentation, answering operational questions in real time instead of routing them to a person.