Two years inside Continental AG’s procurement and digitalization machinery showed me something that no MBA case study ever could: the gap between enterprise data infrastructure and what small businesses actually have access to is not just a technology gap. It’s a structural one.
What I saw inside
At Continental, we had dedicated teams for data governance, ERP configuration, compliance reporting, and IT integration. When a new regulation arrived — say, a new sustainability reporting requirement under CSRD — there was a whole choreography of people, systems, and processes that kicked into gear. Not fast. Not always elegant. But resourced.
The data existed. It was messy, distributed across SAP modules, Excel trackers, and legacy databases, but it existed. And we had the people to hunt it, clean it, and format it into something meaningful.
The SME owner I spoke to last month had a spreadsheet.
Not because she was behind. Because that’s the realistic option when you’re running a 20-person clinic and every hour you spend on reporting is an hour you’re not seeing patients or managing your team.
The real gap isn’t the tool — it’s the assumption behind the tool
Most “SME digitalization” software is built with an implicit assumption: that the business already has clean, organised data and just needs a better interface for it.
That assumption is wrong almost everywhere.
What SMEs actually need is software that meets them where their data is — fragmented, partially manual, operationally gathered — and helps them build structure progressively, without requiring a full digital transformation project upfront.
The enterprise approach is: define the data model, then migrate everything into it. The SME approach has to be the reverse: start with what exists, extract patterns, and build toward structure over time.
What this means for what I’m building
The Clinic Management tool I’m developing starts from this assumption. The first question I ask isn’t “what data do you have?” It’s “what decisions do you make every day, and what information do you wish you had to make them better?”
From there, we work backwards to the minimal data capture needed. Not the ideal state — the next useful state.
That’s the principle. I’m still figuring out the execution. But I think it’s the right direction.