Wei Lin Tan rebuilt her financial close dashboard after it crashed spectacularly during Q2 close last year. The failure happened at 11 PM when her team was racing to finalize board reports, and the automated system simply stopped refreshing. Two previous consultants had blamed insufficient server resources and added more computing power that didn't help.

The Real Problem Hidden in Data Pipelines

Wei discovered that her ERP system locked certain tables during depreciation calculations, which happened to run exactly when her dashboard tried to pull data. The automation worked perfectly 28 days per month but failed when it mattered most. Previous fixes focused on the dashboard software when the collision was happening at the database level.

Rebuilding Around System Constraints

She restructured data pulls to happen in sequence rather than simultaneously. The dashboard now monitors for locked tables and queues requests instead of timing out. Her team closed the books 18 hours faster last month because reports updated throughout the close process instead of requiring manual regeneration when automation failed. The solution required understanding how systems interact under load, not adding more automation features.