Dare to be 10x better

Around December, after reInvent, teams in AWS start to set the goals for next year. I’ve built a habit of challenging myself and each team reporting to me: “what is the one thing you can do, that you can make something 10x+ better than this year?” 10x better?! “It is too bold. It is unrealistic, it is unachievable...”, you may question. “It is not even a SMART goal. Why 10x, why not 9x, 8.5x? Is it because we humans have 10 fingers accidentally?”, the deep analytical mind of you start to doubt. You are all correct. But I still insist there must be at least one thing we CAN do, in the coming 12-18 months, that can be 10x+ better. If we put our heart out and we end up with achieving only 7x, well, it is still a huge win. It is better to overshoot than undershoot in setting up bold goals. Because they force us to move out of our comfort zone, think outside the box and challenge the status quo. Not every goal should be set as 10x+. SMART - Specific, Measurable, Actionable, Relevant and Time Bounded - is still the best practice of setting goals. If we set a goal to increase KMS revenue by 10x over one year, we are definitely smoking something weird. But I want to challenge every team, to pick one thing, one problem, or one customer pain point, that they can make a 10x+ difference. In the last couple years we raised KMS’s symmetrical cryptography API’s default quota 40+ times higher in large regions. It wasn't done in one year but we set the goal at the beginning: “we will be 10x+ better. We will not throttle customer traffic going through organic growth.” We even removed a resource quota called “Grants Per Grantee Principal” completely through architectural innovations. So it is not 10x better in this case, it is infinitely better! The 10x goal don't have to be huge and lofty, it can be set in a small scope and very targeted problem space. it just needs to: 1. Solve a real problem that make some stakeholders’s life in that area 10 times better 2. Be measurable or at least you can estimate the impact convincingly. So why 10x, not 8.5x? Well, 10x sounds better indeed. It represents a qualitative change, not quantitative incremental changes. When you make a 10x difference, you change how the game is played. For example, we are introducing some hardware device management automations so that the rebooted devices can be processed by machine driven workflows, instead of paging human operators in the middle of the nights. This will certainly be a 10x+ difference measured by the after hour pages to oncall engineers - silent night, sweet dreams! It is not just for goals at work. If you can barely do one pull up at a time right now, how about a goal to do 10 times in a roll in 2023. Trust me, I did that myself. You can do it too :-)