Case studies
What went wrong, and what a check would have shown
Public failures where the data, or the checking of it, played a part. Each one ends with the questions worth asking before your own version.
Your AI will believe whatever your data says
In 2022 Unity told investors bad data from one large customer had helped cost it about $110 million in revenue. Before you point AI at your CRM, it’s worth knowing what it will read there.
Nobody told them it had to be right
Target Canada loaded 75,000 products into a new system on a deadline. An internal team later put the data’s accuracy at about 30%. A week of checking at the start would have shown it.
The rows that were never counted
In 2020, nearly 16,000 positive COVID tests went missing from England’s daily figures. The records weren’t wrong. They dropped out of a spreadsheet, and for eight days nothing caught it.
The data arrived. That wasn’t the test.
TSB’s 2018 migration moved its customer data onto a new platform. The data made it. The platform didn’t, and the regulators’ fine came to £48.65 million. A migration has two questions, and only one of them is about whether the data moved.
Have a decision coming up?
Book a short call. Tell us what you’re about to decide and what data it rests on, and we’ll tell you honestly whether it’s worth checking.
Book a call