A framework for rapid experimentation: generate ideas, score confidence, prioritize by ROI, and build quick MVPs instead of betting on one big project per quarter. Ramp's sales tooling came from dozens of small experiments on outreach signals and timing before locking in the UI.
what I prioritize across each of those teams and in my own work is high-velocity experimentation. Like we, we run a framework where we consistently have ideas, we log them, we score our confidence on them, and then we always continue to take the, the top, the ones that have the potentially highest ROI as the work that we go work on. So sometimes you've launches and you need to pause on certain tasks or the highest priority tasks may take a backseat, but we very much focus on what can we do from building out MVPs and larger projects that would start to deliver results and importantly, learnings over time. And why we do it that way is we just learn faster than anyone else. And like, sure, you could commit to one big project in a quarter. If that project doesn't actually like hit the mark, you could have done all the right work, but it's not going to generate the results. I'd much rather take 10 shots on goal. 5 of them hit, 2 of them are failings that you learn more about your audience or the market on, and then you develop from there. So kind of giving you like a concrete example, you might have seen in the last week Ramp released like their internal sales tooling, and that's something that they've been working on for years. And like how they actually developed that was again taking many, many, many shots on goal so that before they ever committed to What are the 10 buttons we put on this screen in the UI? They had like experiments known of like what signals work for outreach, who are the personas they go after, like what's the time of day. So like there were so many experiments that had to go in to generate learnings so that they could actually codify and build infrastructure around it. So that's— I'm at a much— we're at a much earlier stage than RAP, obviously, but we're in the high-velocity experimentation phase. Where we need to adopt those, we need to capture the learnings and put them into practice before we can start to build massive infrastructure tooling. So that across the team, and there are some infrastructure tasks which you just need to do right all the time. That's how we try and prioritize everyone's work.
Roll Claude out from one laptop to the whole GTM org | Sam Gong & Zach Vidibor
