Krithika Shankarraman on Marketing Without a Playbook
Krithika Shankarraman in conversation with Lenny Rachitsky on Lenny’s Podcast (~2025). Krithika was VP of Marketing and first marketing hire at OpenAI, first marketing hire at Stripe (sole marketer for three years), early marketing leader at Retool and Dropbox, and Google Android. Currently Executive in Residence at Thrive Capital.
Key ideas
- The DATE framework over blind playbooks. Krithika resists handing founders the “Stripe playbook” or “OpenAI playbook” because context — competitive landscape, market timing, company values — is what generated the strategy, not the tactics. Her four-step substitute: Diagnose the actual funnel problem, Analyse competitors to find gaps, Take a different path (differentiate deliberately), and Experimentvalidate before scaling.
- Use case epiphany as marketing’s core job at ChatGPT. Awareness was never OpenAI’s problem; the gap was that users did not know what to do with the product. Marketing’s job became creating moments where people realised “I had no idea ChatGPT can do that” — a reminder that the right marketing motion depends entirely on where the funnel is leaking.
- Vanity metrics vs. pipeline metrics. Clicks, views, and impressions are “bullshit numbers.” What matters is signups (PLG) or sales-qualified pipeline and revenue (B2B). At Retool, paid social looked active but was driving no pipeline; customer storytelling — leveraging enterprise logos no competitor could replicate — turned out to be the highest-signal channel.
- Brand is a cross-functional, whole-company asset. Brand is not the marketing team’s artefacts; it is the product experience, customer support quality, candidate experience, and every other touchpoint combined. Velocity and brand quality are not in tension — clarity about brand actually accelerates execution by giving teams shared guardrails.
- The Chameleon CMO and comb-shaped marketers. Modern marketing leaders must be analytically rigorous, commercially accountable (sharing goals with sales and product), and creatively capable. The T-shape of one deep specialism is no longer sufficient; a comb-shape — multiple deep areas of competence — is the target, and AI tools make that breadth more reachable.