Scaling a Tech Org without losing CEO oversight
When I joined HelloFresh in May 2020, the CEO used to sit down with every team every 8 to 10 weeks. Over the next two years, tech grew from about 400 to 1,100 people. My job was to make sure we could scale to that size without him losing oversight of how investments in technology capabilities resulted in business outcomes.
Company | HelloFresh, tech organisation |
My role | Head of Product Portfolio Management, reporting to the CEO |
Timeframe | May 2020 to 2022 |
Scale | About 400 to 1,100 people in tech, product and data |
Where we started
HelloFresh had just doubled the number of meals it delivered, and the Tech team was growing fast to keep up with the new demands. I was impressed by how data-driven the company was, and how product teams already worked with outcomes rather than outputs. But there were two main problems: first, there was no easy way to comprehend what teams were doing and how what they were doing would deliver the outcomes. Second, there were many teams joining Tech that were used to working on projects and track outputs.
The CEO's reviews gave him the opportunity to steer the teams, but they also took a lot of his time. With new teams added every few months, that was not going to scale.
What we did
Understand what leaders needed to see. Before building anything, I wanted to understand at what level of granularity leaders wanted to see the work: tasks, stories, epics, initiatives or strategies. Then we agreed on one language: objectives for what we aim to achieve, strategies for the opportunities, initiatives for our bets, and key results for the outcomes.
A six-month strategic cadence. We thought a quarterly plan was too heavy for HelloFresh. Instead, every six months all teams came together and shared their bets, with how much impact, confidence and effort they thought each one had. The first one ran in September 2020, four months after I joined.
A clear role and a few clear artifacts. I defined what the portfolio function was there to do, and the few artefacts that showed what teams were delivering, in the same way for every team.
The hard part
Not every team was used to working this way. For teams that had always worked in projects and were not used to planning in outcomes, it was a big change, and it came on top of other leadership and technology strategy changes they were going through.
Pushing the framework harder would not have helped. For one team, I stepped in for a while to lead their product function myself, partnering with the leadership team until I could write the job description for someone to do it full time.
What changed
One picture of what everyone was doing. The first alignment gave teams and leaders, for the first time, a shared view of what we were all working on and why.
The CEO got oversight without it consuming his time. He started reviewing the six-month plans instead. The team-level conversations moved to product and tech leadership, where they stayed valuable while the business kept changing.
It held as we grew. The same model and language worked as tech grew to 1,100 people.
As one of my stakeholders mentioned:
"The results were profound: stronger cross-functional collaboration, clearer strategic focus, and a culture that prioritized impact."
What came next
The model got us through two years of very fast growth. By 2022, it showed its limits:
Engineering never fully bought in. It was still hard to see what teams were actually delivering, and new technology leadership needed a way to track it.
Product decisions were not documented enough. We could see the bets, but not always who they were for or the business case behind them.
The process was very tech-focused. Business stakeholders were hardly part of it.
Rebuilding the model to fix these three gaps is the next story.
Looking back, the lesson is one I keep seeing: setting up a Product Portfolio or Product Ops function is the easy part. What makes it last is bringing teams along instead of perfecting the frameworks, adapting the function to match the evolving needs of the company, and starting from the outcome you want to create.
If you're working through that right now, I'd be happy to compare notes.
Comments