The Excel Apocalypse…

Why are trillion-dollar institutions running on a

technological house of cards?

The world’s most powerful enterprises pour millions into cutting-edge AI and blockchain initiatives, but their most critical operations often balance precariously on that most humble of tools: Microsoft Excel spreadsheets. This digital dependency isn’t just risky – it’s already causing spectacular failures.

The horror stories that keep CTOs awake at night

Consider this sobering reality: The Bank of New York Mellon, a financial titan, was using Excel spreadsheets to reconcile 43% of global day trades. Let that sink in. Nearly half of the world’s daily trading activity was being managed through the same software you might use to track your household budget.

But that’s just the tip of the spreadsheet iceberg. The UK’s energy regulator, Ofgem, was managing a staggering £80 billion of market data through a single Excel file with 254 tabs. One wrong click, one corrupted file, one misplaced formula, and the entire British energy market could have been thrown into chaos.

Perhaps most infamously, when a university student simply unhid some rows in Lehman Brothers’ Excel sheets, they accidentally added 779 contracts to the books. We all know how Lehman’s story ended.

The dangerous delusion of control

Excel horror stories reveal a deeper problem: our false sense of security in managing critical data. While organisations sprint toward AI adoption and brace for the EU AI Act’s stringent requirements, they’re overlooking fundamental flaws in their data infrastructure. It’s a very dangerous delusion to think you wield control when your data framework has more holes than a Swiss cheese.

Beyond spreadsheet salvation: real-world solutions

Forward-thinking organisations are already showing us the way forward. Take the gaming industry, where one UK company processes a mind-boggling 100 million behavioural events daily – that’s 400 gigabytes of raw data – with a lean team of just 15 people. Their secret? Abandoning spreadsheet dependency for robust data pipelines and real-time processing systems.

In healthcare, Mayo Clinic’s partnership with Vantiq (intelligent platform pioneers) demonstrates how proper data management can literally save lives. Their system monitors patient vital signs, in real-time, during ambulance transport. It connects physicians virtually with paramedics. And it uses predictive analytics to anticipate patient deterioration. This isn’t just data management. It’s the difference between life and death.

The regulatory tsunami is coming

The upcoming EU AI Act isn’t just another compliance headache. It’s a wake-up call. Organisations will need to quantify algorithmic risk, ensure transparency in high-risk AI systems, and maintain accurate data practices. Those still struggling with basic spreadsheet management will find themselves underwater fast.

The path forward

The maritime industry is an object lesson. Modern ships, equipped with Rolls-Royce marine engines, integrate sensor data with external sources such as weather and tide information. They maintain rigorous data accuracy standards because lives literally depend on it. The lesson? When failure isn’t an option, organisations find a way to move beyond spreadsheet dependency.

The five commandments of data sanity

The choice facing organisations today is clear: continue running trillion-dollar operations on tools designed for personal productivity OR invest in robust data infrastructures that can support modern business demands. Here are five essential guidelines for organisations serious about escaping the spreadsheet trap:

1. Begin with a Minimum Viable Pipeline (MVP), ensuring you understand both data producers and consumers. This means creating a basic but solid foundation that can grow with your needs.
2. Embrace radical transparency about data quality issues. Hiding problems only ensures they’ll explode at the worst possible moment.
3. Demand data models before accepting raw data. This is like insisting on architectural plans before building a house. It seems obvious, yet many – too many – organisations skip this crucial step.
4. Master the maths behind your data. Every calculation, every formula should be understood and documented. No more black boxes.
5. Recognise the crucial distinction between data and information. Data is what you record; Information is what drives business decisions. They are NOT the same thing.

Our thanks to Dan Klein for his guidance and expert contribution to this article. Dan is a Clustre associate and an AI/data expert with a massive global following. Inquisitive, often offbeat but always incisively informed, his tech podcasts have gone straight to the No.1 spot in Apple’s world rankings. 

If you would like to discuss any aspect of this article, you can contact him at: innovation@clustre.net.

Dan will be the keynote speaker at each of our quarterly Data Briefings throughout 2025. Here are the dates and the critical topics he will be covering:

Wednesday 14^th May: Data Security & Privacy

Wednesday 13^th August: Social Expectations & Legal Minefields

Wednesday 12^th November: Technology Adoption & Societal Impact

To register for these briefings simply click here.