Almost no company fails because it moves too slowly. Yet most startups fail because demand, headcount or geography arrived before the operation could handle it. The pressure to keep growing in the startup world is real and it usually wins the argument. In a stellar career that has seen him be parachuted into markets to build operations from zero to go live within under 8 weeks, startup executive Daniel Azzopardi has seen enough to know that managing the rise of both sides of the marketplace concurrently is key to its ensured success.
His experience speaks for itself. At Rocket Internet he launched the foodtech platform Foodora in London and Hong Kong, both fully operational within eight weeks. He ran operations and key accounts, onboarded 500 restaurants, managed the rider fleet and the P&L, and took the business past 1,000 orders a day with delivery times under 30 minutes.
He joined Hometree, a residential energy installation business, pre-revenue at seed stage and built the sales and operations functions from nothing. He designed the company’s first pricing algorithm for selling heating systems online, moved the entire sales cycle onto Salesforce, and grew the business to roughly £9m annualised revenue with 50 staff and more than 200 subcontracted engineers. Hometree raised an £8.5m Series A off the back of it.
At Honest Food Company, a ghost kitchen enterprise, he joined as MD UK before the business was acquired by Delivery Hero, then took over as Managing Director of the whole company when the founders exited. He grew the team from 25 to 180 people and launched seven markets across the Nordics, Central and Eastern Europe and the Middle East, working to a target of ten weeks from first market meeting to live. Annual GMV went from €2.5m to €50m. He then brokered the sale of several of those markets himself.
He now runs Rendezvous, the AI-native experiential travel platform he founded in 2023, which has powered over 30 premium experiences across four continents, and consults various startups on their go-to-market and scale-up strategies.
We asked him how he sequences a launch, when to deliberately slow down, what breaks between 25 people and 180, and what he would tell anyone starting a company in 2026.
You built Foodora in London from scratch to fully operational in eight weeks. How did you manage to do this in such a short time span?
Speed comes from sequencing, not from asking everyone to work harder. Although at Rocket Internet we had a bit of both.
Before launch, I reduced the business to a small number of critical paths: restaurant supply, rider capacity, technology readiness, customer support and launch demand. Each workstream had a clear owner, a deadline and an explicit definition of “good enough to launch.” Anything that did not directly affect our ability to take and fulfil an order was secondary.
We also resisted the temptation to perfect everything before going live. A launch is not the finished product, it is just the beginning of the learning process. The objective was to create a functioning marketplace quickly, observe where reality contradicted our assumptions and improve it every day.
The other important factor was personal involvement. In an early-stage launch, leaders cannot manage exclusively through presentations and meetings. I recruited riders, onboarded restaurants, negotiated supplier contracts and dealt with operational failures directly. That gave me an accurate understanding of the system before I started delegating it.
In short: narrow the scope, identify the dependencies, put strong people in charge of each workstream and make decisions faster than problems can accumulate.
You managed both operations and key account management at Foodora and onboarded 500 restaurants. How do you sequence supply-side growth against demand growth without one outrunning the other? How do you ensure that growth is sustainable?
Marketplace businesses often fail because they celebrate aggregate growth while ignoring local liquidity.
You do not really have one marketplace. You have hundreds of small marketplaces, defined by geography, time of day, customer behavior and available supply. A city can look healthy in aggregate while individual neighborhoods or meal periods are fundamentally broken.
The correct sequence is to build enough high-quality supply to create a compelling customer proposition, then concentrate demand into areas where you can fulfil it reliably. Expanding supply endlessly before demand exists creates disengaged partners. Driving demand before operational capacity exists creates late deliveries, poor customer experiences and churn.
We therefore watched density metrics very closely: orders per restaurant, orders per rider hour, delivery times, acceptance rates, partner utilisation and repeat behavior. The goal was not simply to add more restaurants. It was to improve the productivity of the network.
Sustainable growth also requires the willingness to slow down. That is unpopular in venture-backed businesses, but sometimes the correct operational decision is to stop acquiring demand until fulfilment catches up.
Growth that damages customer trust, partner economics or contribution margin is not growth. It is a deferred failure.
You took Honest Food Company from €2.5 million to €50 million in annual GMV. What changed about how you ran the business between year one and year three?
In year one, growth depended heavily on individual effort. In year three, it had to depend on the quality of the system.
Early on, the priority was proving that the model worked: launching brands, signing partners, solving supply problems and generating orders. We moved quickly and accepted a certain amount of operational inefficiency because learning was more valuable than optimisation.
As the company scaled, that approach became dangerous. Every exception created at small scale becomes a department at large scale. We therefore shifted from asking, “Can we make this work?” to asking, “Can we make this work repeatedly, profitably and without senior management intervening?”
That meant standardizing market launches, simplifying the SKU base, improving forecasting, consolidating suppliers and becoming far more rigorous about unit economics. We also changed how we evaluated growth. More orders were not automatically good orders. We looked at order density per location, ingredient utilisation, logistics cost, partner performance and contribution margin.
The leadership role changed as well. Initially, I was solving problems personally. By year three, my job was to build people, systems and decision-making structures that could solve those problems without me. It took a few psychological hoops to jump through to learn to let go and trust your team to make the right decisions.
The transition from startup to scale-up is the transition from heroics to repeatability.
You expanded across the Nordics, Central and Eastern Europe, and the Middle East under one operating umbrella. How different is the approach depending on the market?
The underlying operating principles are usually more universal than people think. Roughly 80% of the model should be standardised and 20% should be adapted locally. The mistake is choosing the wrong 20%.
The customer journey, KPI architecture, launch process, financial discipline and accountability model should generally remain consistent. That creates comparability across markets and prevents every country team from building its own version of the company.
What changes are the conditions in which that operating model has to function. Labour regulation, supplier sophistication, payment behavior, delivery infrastructure, seasonality, customer expectations and management culture can vary enormously. In some markets, the primary challenge is recruiting reliable partners. In others, it is logistics, regulation or finding managers who can operate with genuine autonomy.
Localisation should be evidence-led rather than cultural theatre. Teams often say, “Our market is different,” when what they really mean is, “We would prefer not to follow the central process.” Sometimes they are right, but the burden of proof should be on the exception.
The central team should define the non-negotiables: economics, customer standards, reporting and brand. Local teams should control how those outcomes are achieved within the realities of their market.
Looking back at both companies, what is one operational decision you would make completely differently if you were starting from scratch today?
I would impose much harder operational and economic gates before expanding into additional markets.
When capital is abundant, geographic expansion can create the appearance of progress. New markets generate headlines, headcount and top-line growth. They also multiply complexity. Every new country adds suppliers, legal structures, management layers, systems requirements and local exceptions. We fell into the trap of chasing vanity metrics because we were under pressure to show as simplistically as possible that the business was doing well. Top line revenue gives that illusion.
The mistake is assuming that a model which works at a small scale is ready to be replicated. In reality, expansion often exports unresolved problems and makes them harder to diagnose.
Today, I would require a market or operating unit to meet clear thresholds before approving the next launch: stable fulfilment, credible contribution economics, high supply-side utilisation and satisfaction, repeatable customer acquisition, documented processes and a leadership team capable of operating without constant central intervention.
I would also close underperforming experiments earlier. Operators can become emotionally attached to solving difficult markets because closing them feels like failure. But capital and management attention are finite. Keeping a structurally weak market alive can deprive a strong market of the resources it needs to become exceptional.
Scale should be earned. It should not be used as a substitute for fixing the core business.
What is your advice to people aiming to start a business in 2026?
Do not confuse the falling cost of building a product with the falling difficulty of building a company.
AI has made it dramatically easier to produce software, content, designs and prototypes. That means supply will explode. It does not mean customers will care. Distribution, trust, operational execution and genuine differentiation will become more valuable precisely because production has become easier. If you don’t know how to point the AI towards the right things, you’re just shipping crappy products faster, wasting everyone’s time in the process (as well as spending AI tokens unnecessarily).
I would advise founders to begin with an expensive, frequent and poorly solved problem – not with a technology they are excited to use. Speak to customers before building. Charge earlier than feels comfortable – revenue is a more honest signal than enthusiasm. Do not trust anyone that says your product is cool. If they’re not prepared to pay money for it, they’re not that excited about it.
Founders should also be much more disciplined about capital. The previous decade taught too many companies to use funding as a substitute for a functioning business model. Raise money when it accelerates something that already works, not when it postpones having to discover whether it works. Great products grow through word of mouth initially – if you’re having to spend marketing money to sell your product when in its infancy, then your product probably isn’t that great.
Finally, choose a problem you are prepared to remain interested in after the novelty disappears. Most businesses are not built through a single brilliant insight. They are built through years of hiring, selling, fixing processes, managing cash and making difficult decisions with incomplete information.
The idea matters. The ability to keep executing after the idea stops being exciting matters far more.
