RevOps is a game of competing priorities. Anthony Goldbloom, co-founder of Kaggle and CEO of Sumble, thinks one job comes before all of them. Know, with precision, which accounts you should be selling to and who the right people are inside them. Account and people scoring aren't new. What's new is that you can now build them precisely enough that sales, marketing and RevOps will plan against the same list, instead of each keeping a shadow list of their own. Anthony will build an account score live, pull it apart piece by piece, and walk through the data science principles that decide whether your reps trust a score or override it. He'll show why six data steps at 80% accuracy leave you with a signal that's right 26% of the time, why closed-won is a biased ground truth, and how to backtest a score before you carve territories on it. You'll leave with a checklist for building or auditing your own account score, and a clear view of what changes once your whole GTM team targets the same accounts: territory planning, lead routing, rep prioritisation, even who gets invited to the dinner. Key takeaways - Why account and people scoring, not stack consolidation, is the highest-leverage investment a RevOps team can make - How error compounds across a data chain, and why "our data is pretty good" isn't an answer - What LLMs now let you see inside an account: exact job function, the teams and technologies behind each person, real persona counts