I was asked to scope infrastructure costs for a photo backup feature. I started with active iCloud users, estimated average library sizes, and multiplied through. We landed on roughly 800 petabytes. Leadership used that figure to negotiate the next storage contract.
Outcome-only narration; no assumptions surfaced and the figure cited is implausibly low.
We ran that estimation for a capacity planning review. I anchored on 1.2 billion active iCloud users, assumed average library size, and built up from there. The infrastructure team accepted the model and used it to size the next procurement cycle.
Names a user-count anchor but leaves the library-size assumption invisible and unverifiable.
I ran the estimate myself — one billion active iCloud users, roughly 2,000 photos each at 4MB average, gets you around 8 exabytes. I flagged it to the infra team. They'd been planning for half that. We revised the roadmap.
Shows every multiplication step, auditable math, and caught a consequential 2× planning gap.