SUMMARY
In 30 seconds
Batchu says most growth projects he has seen did not succeed and argues for reaching conclusions quickly.
- 1High failure rateBatchu describes his experience.
- 2About 30%He estimates his historical success rate.
- 3Decisive learningHe emphasizes failing quickly and conclusively.
Takeaways
- The estimate reflects his experience.
- Many experiments may fail.
- Reach conclusions quickly.
FULL TEXT
Read the edited text by chapter
01
Experiment outcomes
Ramp growth leader Sri Batchu says that growth projects have had a high failure rate across the companies where he has worked.
Ramp growth leader Sri Batchu says that growth projects have had a high failure rate across the companies where he has worked.
In his personal experience, roughly 30% of these projects succeeded, leaving about two-thirds unsuccessful. This is his historical estimate, not a universal benchmark.
02
Learn quickly
Batchu argues that the key is to learn quickly and determine conclusively when an experiment has failed.
Batchu argues that the key is to learn quickly and determine conclusively when an experiment has failed.
More from this channel
1:00:17EN
20VC with Harry Stebbings· Oct 29, 2026Podcast · 3 speakersHarry Stebbings, Jason Lemkin and Rory O’Driscoll on AI financing and venture betsGuest Jason Lemkin, Rory O’Driscoll · Host Harry StebbingsThe investors debate AI lab economics, spending on infrastructure, open-model risks and how to evaluate fast-moving private investments.
23:48EN20VC with Harry Stebbings· Oct 26, 2026Podcast · 2 speakersSatya Nadella on Zero Trust for AI and Why Every Company Should Own Its IntelligenceGuest Satya Nadella · Host Harry StebbingsThe Microsoft CEO argues that capable agents need security boundaries even as alignment research remains unsettled, and explains why firms should retain the value created by their own AI use.
1:13:26EN20VC with Harry Stebbings· Oct 24, 2026Podcast · 2 speakersNaveen Rao on Energy, AI Hardware, Verification and the Future of RoboticsGuest Naveen Rao · Host Harry StebbingsThe Unconventional AI founder argues that energy efficiency will shape AI hardware, while the growing volume of model output makes verification more valuable.
59:41EN20VC with Harry Stebbings· Oct 22, 2026Podcast · 3 speakers20VC: Venture Fund Size, AI Valuations and the Changing Software MarketGuest Jason Lemkin, Rory O’Driscoll · Host Harry StebbingsHarry Stebbings, Jason Lemkin and Rory O'Driscoll debate venture concentration, infrastructure risk, code review and the assumptions behind fast-growing AI companies.
1:19:50EN20VC with Harry Stebbings· Oct 19, 2026Podcast · 2 speakersSameer Dholakia on AI Investing, Bessemer's Growth Strategy and Company CultureGuest Sameer Dholakia · Host Harry StebbingsThe Bessemer partner explains his approach to pricing fast-growing AI companies, backing founders and building teams, while discussing the risks behind his optimistic forecasts.
1:03:05EN20VC with Harry Stebbings· Oct 17, 2026Podcast · 2 speakersHarrison Chase on Building Reliable Agents, Model Routing and MemoryGuest Harrison Chase · Host Harry StebbingsThe LangChain co-founder explains why agents remain difficult to control, where domain expertise matters, and how his team manages coding-agent costs.