From cloud skepticism to global ERP

In the late 1990s, putting business-critical data on the internet was widely considered unsafe. Today, global project businesses operate across entities, currencies, and continents from a laptop.

Following Leo Koster’s recent Medium article reflecting on 25 years in SaaS, we asked him to go deeper into the operational decisions behind VOGSY.

Over multiple technology cycles, from early cloud resistance to legacy ERP complexity and now enterprise AI experimentation, one theme remains consistent: infrastructure decisions shape long-term business outcomes.

In this full Q&A, he explains:

• Why most ERP systems carry a hidden innovation tax

• Why simplicity beats feature checklists

• Why VOGSY chose Google Cloud over larger competitors

• What zero downtime actually means in practice

• Why explainability matters more than intelligence in enterprise AI

For international, project-based businesses, these decisions directly impact rollout speed, financial visibility, and long-term resilience.

Early cloud skepticism and timing

You were building online software when people said it was impossible. What did they get wrong?

They confused physical control with security. In the late nineties, everyone thought if they couldn’t touch their server, their data wasn’t safe. What they missed was scale. A global platform like Google can hire more security engineers than any single company ever could.

In the late 90s, why didn’t people trust putting business data on the internet?

It was fear of the invisible. Leaders were used to filing cabinets and locked doors. The internet felt chaotic. The problem was not the internet itself, but the immature infrastructure.

Looking back, were the skeptics wrong or just early?

Skepticism was healthy. The risk was real. The mistake was thinking the solution was to stay offline. The solution was partnering with infrastructure stronger than anything you could build yourself.

How do you know when the world isn’t ready yet versus when you are just too early?

If the problem is inevitable, you persist. We knew disconnected software and emailed spreadsheets would not survive in the long term.

The innovation tax and operational breaking point

What was the moment you realized you were spending more time maintaining software than improving it?

We introduced Scrum and failed every sprint. Developers were drowning in support tickets. They were afraid to touch legacy code because fixing one issue created two more. We had become a repair shop.

Why is infrastructure a hidden tax on product teams?

Infrastructure demands immediate attention. Every minute spent fixing servers is not building value. Leaders underestimate this cost.

What’s the cost of keeping the lights on?

It’s the innovation tax. If 80 percent of your budget goes to maintaining old ERP systems, only 20 percent is left to compete. It also drives away top engineers who want to build, not patch.

Choosing simplicity over complexity

Why did simplicity matter more than raw power when choosing a cloud platform?

Simplicity is power. Complexity slows you down. If you need a PhD to turn your infrastructure on, you are focused on the wrong problem.

You spoke to Microsoft, Amazon, IBM, and Salesforce. What made Google different?

Google offered a true platform. They do the heavy lifting. Strategically, they also have no ambition to build a B2B ERP. They provide the engine. We built the car.

Why do platforms confuse complexity with capability?

Procurement departments love feature checklists. Users do not. We focus on flow and profitability for service businesses, not feature volume.

How do you decide what not to build?

We don’t copy, we connect. If best-in-class tools already exist, we integrate with them and focus on what makes project businesses profitable.

Betting on physics and building trust

How do you evaluate risk in a 10-year technology decision?

You distinguish between fashion and physics. Fashion changes. Physics does not. The move to the cloud was gravity, not hype.

What does trusting a platform mean in practice?

It means sleeping at night. When a major security issue hits the news, you know hundreds of engineers are already fixing it. Trust is outsourcing your anxiety.

When did you know the bet had paid off?

When a customer said they didn't need training. In ERP, that is rare. The software became natural behavior.

Small team, global scale

Why do people find it hard to believe you run a global ERP without exploding headcount?

They are conditioned to the army of consultants model. Software is leverage. Automate onboarding, simplify the UI, and use serverless infrastructure. You do not need an army.

Why is revenue per employee a better goal than headcount?

Headcount is ego. Revenue per employee proves the software is doing the heavy lifting.

How has the cloud changed scale?

Scale used to mean more offices. Now it means more reach. We can serve customers globally without increasing physical footprint.

Stability, calm, and responsible AI

What does zero downtime mean for customers?

It means trust. The system is like electricity. You only notice it when it’s gone. With VOGSY, it isn’t.

Why is calm underrated in enterprise software?

Professional services firms already operate in high-pressure environments. Their software should not add stress. Calm is a competitive advantage.

Why does AI feel similar to the early days of cloud?

It’s the same mix of hype and fear. The winners will solve real operational problems, not showcase tricks.

Why is explainability more important than intelligence?

Enterprise decisions require accountability. Leaders must understand why an AI recommendation was made.

What does AI assist, not replace judgment, mean?

In professional services, the product is trust. AI can remove repetitive work and surface insights. Judgment remains human.

Closing

Technology cycles repeat. Infrastructure matures. New tools emerge. The constants remain discipline, long-term thinking, and trust.

For international project businesses operating across entities and currencies, the foundation matters. Stability is not the opposite of innovation. It is what makes innovation sustainable.

Read Leo's post on LinkedIn here.