Dell/ Feature change during product development phase. Intel CPU late spec change caused product architecture level impact. Saved $90M future sales Customer obsession, Ownership, Think big, Be right a lot, Bias for action, Deliver results, Dive Deep - going above and beyond to meet customer needs / when you had to pivot 75% into the project development S - To compete better with AMD, Intel changed CPU spec to increase max performance in the middle of product development cycle 3 months before my product launch. This change would lead to a change at system level for PSU and thermals and such drastic change was never expected. Up to 10% performance bump was observed. Intel changed their CPU turbo duration to compete better with AMD. T - Complication was that if Dell doesn't make this change, customers will perceive that our product runs slower and they wouldn't buy Dell plus it’ll drop our NPS by 3pt. A - - Measured customer impact by checking sales bids and mention of CPU benchmarks to understand how many bids will be affected - 20% contracts were affected cause they mentioned performance benchmark as #1 priority to win bid. - Negotiated with engineering teams to allocate resources for re-tuning of system with better cooling. - At the time, engg team was involved with performance benchmarking for graphics cards used for sales collateral since we were getting ready for sales training. - Went through Sales and Ops management and exception approval process to adjust operations milestones while keeping launch timeline intact. (milestone for example: reduce pilot production qty and time, delay chassis validation testing) R - able to save 20% of cases ($1.5M per month margin impact) where we would be exposed, and met product launch timelines. NPS did not drop. what did you deprioritize? how did you decide? Performance engineering was doing graphics performance testing that I can use to create sales education material. Ops argument: Risky to reduce prod qty, but it was still within statistical limit Sales argument: Risky to delay sales training, but graphics performance data could be benchmarked based on publicly available info for initial 2 months.