Krithika Shankarraman25 May 2025

Growth tactics from OpenAI and Stripe’s first marketer

5Frameworks

15Insights

Episode overview

Lenny Rachitsky speaks with Krithika Shankarraman about lessons from leading early marketing at OpenAI, Stripe, Retool, Dropbox, and Google. She argues that effective marketing begins with deep customer and product understanding, then introduces a four-step diagnostic approach: diagnose the problem, analyze competitors, take a differentiated path, and experiment. The conversation also covers brand consistency, marketing-team processes, pricing experiments, AI-assisted work, career development, and lessons from the failed Stripe Relay launch.

Key ideas

Marketing should diagnose the company’s actual constraint instead of reflexively applying another company’s playbook.
The four-step DATE framework is to diagnose the problem, analyze competitors, take a different path, and experiment before scaling.
Product management and product marketing should work together from the beginning of product development, using customer insights and language.
Marketing performance should be measured through signups, qualified pipeline, revenue, and engagement rather than clicks, views, or impressions alone.
Consistent brand expectations and lightweight 20%-and-80% reviews can help teams move faster while maintaining quality.
Modern marketing leaders should become adaptable, comb-shaped operators who combine analytical, creative, product, and go-to-market capabilities.
AI can augment analysis and creativity, but durable differentiation still depends on taste, product understanding, and customer understanding.
Pricing and packaging decisions should be tested through experiments that reveal how customers perceive and obtain value.

Transcript available · source text is retained privately and is not published

Frameworks in this episode

Marketing3 steps

Support-Rotation Content Backlog

Mine support tickets for recurring confusion, then rank-order self-serve landing pages that answer them

Read →

Communication3 steps

The 20/80 Marketing Review Checkpoints

Two aligned checkpoints — strategy at 20%, execution at 80% — so process speeds you up instead of slowing you down

Read →

Leadership3 steps

The Comb-Shaped (Chameleon) Marketer

Move from T-shaped to comb-shaped — multiple deep skills you deploy on-demand via diagnosis

Read →

Marketing4 steps

The DATE Framework for Marketing Strategy

Diagnose, Analyze, Take a different path, Experiment — an engineer's anti-playbook for marketing

Read →

Communication3 steps

The Expectations-Reality Delta

Unhappiness is the gap between expectation and reality — and expectations are far easier to change

Read →

People & resources mentioned

Attributed to the moment in the episode. Timestamps are approximate.

People · 14

Krithika ShankarramanMentionsCritica was the first marketing hire and VP of marketing at OpenAI, the first marketing hire at Stripe
“Critica was the first marketing hire and VP of marketing at OpenAI, the first marketing hire at Stripe”
▶ Watch at ~1:00approx.
Kelly SimsMentionsmentioned
“A big thank you to Kevin Garcia and Kelly Sims for suggesting questions and stories to get into.”
▶ Watch at ~1:30approx.
Sam AltmanMentionsthere's everybody has uh and had and continues to have like such deep trust in in Sam and Greg and our leadership team
“I remember there's a video of Sam Alman being asked, "How do you make money with something like this?"”
▶ Watch at ~8:30approx.
Patrick CollisonMentions“there's this tweet by Patrick hollison that I've brought up a couple times on this podcast”
“John and Patrick were the first marketers at Stripe because they were developers themselves.”
▶ Watch at ~23:00approx.
John CollisonMentionsJohn and Patrick were the first marketers at Stripe because they were developers themselves.
“John and Patrick were the first marketers at Stripe because they were developers themselves.”
▶ Watch at ~23:00approx.
Cristina CordovaMentionsour head of partnerships at the time, Christina Cordova
“our head of partnerships at the time, Christina Cordova, handed me a an a hackpad”
▶ Watch at ~24:30approx.
Raaz HerzbergMentionsRos is Chief marketing officer and VP of product strategy at whiz
“there's so much of what you share that reminds me of Ros from Whiz”
▶ Watch at ~27:30approx.
Lenny RachitskyCoinedtoday we've got another very special compilation episode something I've been pulling on more and more with the podcast and the newsletter
“I've uh very viscerally learned the power of brand doing my newsletter.”
▶ Watch at ~38:00approx.
Brian CheskyMentionsBrian chesky at Airbnb there was a whole article where we unpacked this moment in their product development cycle
“there's a reason Brian at Airbnb merged marketing and or product marketing and product management”
▶ Watch at ~42:30approx.
Kevin GarciaMentionsOne of your former colleagues, Kevin Garcia, wanted me to ask you something.
“One of your former colleagues, Kevin Garcia, wanted me to ask you something.”
▶ Watch at ~55:00approx.
Tobi LütkeMentionsif you look at Toby from Shopify, his scores aren't that good, but that company's doing really, really well.
“That's something that stuck with me from my chat with Toby Lucky from Shopify.”
▶ Watch at ~59:00approx.
Kevin WeilMentionsWe got connected through Kevin Wheel, who is former CPO at OpenAI, now head of science at OpenAI.
“This reminds me of when Kevin Wheel was on the podcast talking about Libra”
▶ Watch at ~1:07:00approx.
April DunfordRecommendsApril Dunford was on this podcast and she has this book that she put out recently called the sales pitch
“one book that I recommend to most people is April Dunford's book on positioning called Obviously Awesome.”
▶ Watch at ~1:08:30approx.
Madeline MillerRecommendsmentioned
“one of the best reads in the last couple of years has been Meline Miller's Cersei”
▶ Watch at ~1:09:00approx.

Resources · 33

OpenAIMentionscompany
“You were the first marketing hire at OpenAI.”
▶ Watch at ~0:00approx.
AndroidMentionssoftware · Google
“She also did marketing for Android at Google.”
▶ Watch at ~1:00approx.
Thrive CapitalMentionscompany
“Currently, she's executive in residence at Thrive Capital where she supports their portfolio and founders on all things marketing”
▶ Watch at ~1:00approx.
GoogleMentionscompany · Google
“She also did marketing for Android at Google.”
▶ Watch at ~1:00approx.
DropboxMentionscompany · Drew Houston and Arash Ferdowsi
“She was also an early marketing leader at Retool and at Dropbox.”
▶ Watch at ~1:00approx.
Lenny's NewsletterCoinednewsletter · Lenny Rachitsky
“if you become an annual subscriber of my newsletter, you get a year free of a bunch of worldclass products”
▶ Watch at ~1:30approx.
ChatGPT EnterpriseMentionsproduct · OpenAI
“When we turned on the contact sales form for ChatGpt Enterprise, which was one of my first launches at the company”
▶ Watch at ~7:30approx.
PythonUsessoftware · Guido van Rossum
“I sat down with ChachiBT and I coded up a Python script”
▶ Watch at ~8:00approx.
ChatGPTUsessoftware · OpenAI
“I sat down with ChachiBT and I coded up a Python script”
▶ Watch at ~8:00approx.
RetoolMentionssoftware · Retool
“one that comes to mind is definitely Retool.”
▶ Watch at ~17:00approx.
NetflixMentionscompany · Netflix
“Oh, Netflix is using retool. Oh, maybe I should check it out.”
▶ Watch at ~20:00approx.
AirbnbMentionscompany · Brian Chesky, Joe Gebbia, and Nathan Blecharczyk
“our product Stripe Connect which was made for marketplaces like Uber and Airbnb”
▶ Watch at ~21:00approx.
UberMentionscompany · Uber
“our product Stripe Connect which was made for marketplaces like Uber and Airbnb”
▶ Watch at ~21:00approx.
Stripe ConnectMentionsproduct · Stripe
“our product Stripe Connect which was made for marketplaces like Uber and Airbnb”
▶ Watch at ~21:00approx.
StripeMentionscompany · Patrick Collison and John Collison
“You were also, I believe, the only marketing person for three years at Stripe.”
▶ Watch at ~22:30approx.
HackpadUsessoftware
“handed me a an a hackpad at the time, which is like a a notion.”
▶ Watch at ~24:30approx.
Dropbox PaperMentionssoftware · Dropbox, Inc.
“I remember Hackpad that turned into Dropbox paper.”
▶ Watch at ~24:30approx.
WizMentionscompany
“there's so much of what you share that reminds me of Ros from Whiz”
▶ Watch at ~27:30approx.
LinkedIn AdsMentionsproduct · LinkedIn
“When you're ready to reach the right professionals, use LinkedIn ads.”
▶ Watch at ~31:30approx.
SlackUsessoftware · Slack Technologies
“it can be a slack channel where people are posting things async”
▶ Watch at ~34:30approx.
ClaudeMentionssoftware · Anthropic
“I wanted actually to ask so an interesting thing is happening with JGBT versus Claude.”
▶ Watch at ~43:00approx.
FigmaUsessoftware · Figma
“Sometimes it means looking at a Figma file for a landing page that's going out the next day”
▶ Watch at ~50:00approx.
ShopifyMentionscompany · Shopify
“That's something that stuck with me from my chat with Toby Lucky from Shopify.”
▶ Watch at ~59:00approx.
Stripe RelayMentionsproduct · Stripe
“I'll talk about a fantastic success, which is called Stripe Relay, which you probably I'm just kidding because nobody remembers it.”
▶ Watch at ~1:05:30approx.
LibraMentionsproduct · Facebook
“This reminds me of when Kevin Wheel was on the podcast talking about Libra”
▶ Watch at ~1:07:00approx.
FacebookMentionscompany · Meta
“this cryptocurrency project that Facebook ran”
▶ Watch at ~1:07:00approx.
Obviously AwesomeRecommendsbook · April Dunford
“one book that I recommend to most people is April Dunford's book on positioning called Obviously Awesome.”
▶ Watch at ~1:08:30approx.
CirceRecommendsbook · Madeline Miller
“one of the best reads in the last couple of years has been Meline Miller's Cersei”
▶ Watch at ~1:09:00approx.
SeveranceRecommendstv show · Dan Erickson
“I'm finally catching up on Severance, so no spoilers, but I'm about halfway through the first season.”
▶ Watch at ~1:09:30approx.
GranolaRecommendssoftware · Granola
“Granola for meeting notes.”
▶ Watch at ~1:10:00approx.
Lenny's PodcastCoinedpodcast · Lenny Rachitsky
“You can find all past episodes or learn more about the show at lennispodcast.com.”
▶ Watch at ~1:13:30approx.
Apple PodcastsMentionssoftware · Apple Inc.
“you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.”
▶ Watch at ~1:13:30approx.
SpotifyMentionssoftware · Spotify
“you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app.”
▶ Watch at ~1:13:30approx.

Spot an error or want something removed? Request a correction or removal.

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 3

Hot Take10:00

Why Copying Another Company's Playbook Doesn't Work

Founders constantly ask her to reveal 'how Stripe did it' or 'how OpenAI did it,' and she resists. Outcomes are the product of context, competitive landscape, and timing. Copying the outputs and tactics ignores the inputs and variables that actually produced the strategy — so the same moves may not apply to you at all.

Founders want the playbook; she refuses to hand over tactics
Success depends on context, competition, and zeitgeist that won't repeat
Copying outputs ignores the inputs that drove the strategy
Better to become a diagnostician than to imitate someone else's result

“if you're just kind of kind of copying the outcomes or the outputs of the strategy and trying to follow in the footsteps of the…”
Krithika Shankarraman · 10:30

#marketing#strategy#playbooks

Hot Take32:30

Good Process Speeds a Company Up — And the 80% Review Rule

Contrary to move-fast startup instinct, Krithika argues sufficient process actually accelerates a company and keeps the brand consistent. She recommends a transparent 'marketing review' forum plus two checkpoints: a 20% strategy review and an 80% review. Reviewing at 99% is worthless because there's no slack left to make substantive changes.

Sufficient process speeds a company up rather than slowing it down
Run a transparent marketing-review forum (meeting, Slack, or email alias)
20% review = strategy; 80% review = work done but still changeable
A 99% review is a worthless rubber stamp with no room to change anything

“good process or sufficient process is actually something that speeds up a company rather than slow it down.”
Krithika Shankarraman · 32:30

“if you come in at the 99% mark and you're just looking for a rubber stamp of approval and you don't really have the slack…”
Krithika Shankarraman · 35:30

#process#marketing-ops#brand#scaling

Hot Take43:00

Why ChatGPT Owns Mindshare Even When Claude Is Better at Some Things

Asked why ChatGPT dominates AI mindshare despite Claude being arguably better at coding and writing, Krithika points past eval one-upmanship. Loyalty builds when there's a small delta between a user's expectations and their reality, and it compounds when expectations are exceeded. She argues these companies must think long-term about AI as a force for humanity, not just win the best-chatbot race.

Delightful products win on the same fundamentals, not just eval scores
Loyalty grows when the gap between expectations and reality is small
Exceeding expectations is accretive to brand and loyalty
The real game is long-term change management, not chatbot one-upmanship

“there's there's a sense of loyalty that builds up over time when there is a shorter and and smaller delta between your expectations and your…”
Krithika Shankarraman · 44:00

#ai#chatgpt#claude#brand-loyalty

Explainer· 2

Explainer05:00

What Marketing Actually Did at OpenAI: Create the Use-Case Epiphany

ChatGPT never had an awareness problem; everyone had heard of it. One zoom-level deeper, the real gap was that people didn't know what to use it for or what it replaced. So marketing's job became engineering moments where a user thinks 'I had no idea ChatGPT could do that' rather than running generic top-of-funnel tactics.

Awareness was never OpenAI's problem — usage confusion was
Users didn't know what ChatGPT replaced (search? something else?)
Marketing's real job was creating use-case epiphanies, not clicks
Be diagnostic about what marketing should do rather than defaulting to funnel playbooks

“the work of marketing ended up becoming creating the sort of use case epiphany where people could say I had no idea chatt could do…”
Krithika Shankarraman · 05:30

#openai#marketing#chatgpt#positioning

Explainer52:30

The Chameleon CMO: Go From T-Shaped to Comb-Shaped

Conventional wisdom says a CMO should be T-shaped — deep in one pillar (brand, demand, or product marketing). Krithika argues modern marketing leaders must be analytical, best friends with data science, creative in service of revenue, and unable to operate in a silo. The move is from T-shaped to 'comb-shaped,' going deep across multiple domains as the diagnostic demands.

Conventional CMO advice is to be T-shaped in one pillar
Modern leaders must be analytical and partner with data science
Creativity has to serve the buyer journey and revenue goals
The upgrade is comb-shaped — depth in several domains, chosen by diagnosis

“the ability to diversify your interest maybe going from T-shaped to comb shaped is probably the right”
Krithika Shankarraman · 53:30

#career#cmo#marketing-leadership

Story· 6

Story07:30

ChatGPT Enterprise Leads 40x'd Overnight — So She Coded Her Own Lead Scorer

One of Krithika's first launches at OpenAI was turning on the contact-sales form for ChatGPT Enterprise. Lead volume 40x'd overnight, far beyond expectations. With no demand-gen problem to solve, she instead sat down with ChatGPT and coded a Python script that became OpenAI's first lead-qualification and scoring model, used in production far longer than she'd admit.

Turning on the ChatGPT Enterprise contact-sales form 40x'd lead volume overnight
The demand problem was inverted — too many leads, not too few
She used ChatGPT to write a Python lead-scoring script
That script ran as the production lead-qualification model for a long time

“our lead volume 40xed overnight.”
Krithika Shankarraman · 07:30

“Like I sat down with ChachiBT and I coded up a Python script that ended up functioning as our first lead qualification, lead scoring model…”
Krithika Shankarraman · 08:00

#openai#enterprise#demand-gen#lead-scoring

Story19:00

How Retool Zigged: Customer Storytelling Instead of Copycat Content

Retool's competitors leaned on content marketing and events, and Retool could have followed. Instead they doubled down on customer marketing and storytelling because their real differentiator was true enterprise traction — customers who paid, believed, and expanded. No copycat could replicate that logo bench, so they let those customers tell the story on Retool's behalf.

Competitors did content marketing and events; Retool took a different path
Retool's edge was genuine enterprise customers who expanded usage
Customer storytelling and marquee logos were the differentiator
No copycat competitor could replicate that customer bench

“the thing that differentiated Retool from a lot of the copycat competitors in the market was that we had terrific traction with true enterprises who…”
Krithika Shankarraman · 19:00

“no other company could replicate the kind of customers that retool had in its bench.”
Krithika Shankarraman · 19:30

#retool#differentiation#customer-marketing#enterprise

Story21:30

Stripe's Reverse-RFP Play: Ranking for 'Payfac' Without Selling It

Stripe Connect's real competition was companies deciding to become payment facilitators themselves. Consultants used legacy jargon to help them. Stripe instead wrote content ranking for those SEO terms — a 'secret playbook' for becoming a payfac that honestly admitted the process is onerous and annoying, then recommended using Stripe Connect instead. Zigging where others zagged.

Connect's competition was companies becoming payfacs themselves
Stripe targeted the SEO terms buyers were searching for
They published a genuinely useful 'how to become a payfac' guide
The guide's honest conclusion: it's onerous, use Stripe Connect instead

“here's the secret playbook. Like, here's all the things that you have to do. And by the way, if this feels ownorous or annoying, it…”
Krithika Shankarraman · 22:00

#stripe#seo#content#differentiation

Story29:00

Turning Support Tickets Into a Self-Serve Marketing Funnel at Stripe

Every Stripe new hire did a support rotation to build customer empathy. On hers, Krithika noticed recurring confusion — people didn't know Stripe handled subscriptions, recurring payments, or payouts, because Stripe never told them. That became a stack-ranked backlog of educational landing pages. For developers, the best marketing funnel often looks like self-directed education, not a sales call.

All Stripe new hires did a support rotation to build customer empathy
Recurring support themes revealed undocumented capabilities
Those gaps became a prioritized backlog of educational landing pages
For developers, the funnel is self-serve education, rarely a sales conversation

“a fantastic marketing funnel sometimes doesn't look like talking to sales. It often never looks like talking to sales. It looks like a self-directed educational…”
Krithika Shankarraman · 30:30

#stripe#developer-marketing#content#customer-research

Story1:00:00

The Controversial Retool Bet: Making Self-Hosted Self-Serve

Retool took a controversial pricing experiment — the self-hosted version, previously gated behind sales, was made self-serve to anyone. The insight: customers wanted self-hosted for many reasons (regulated industries, private data/PII), not just enterprise size. It shrank the sales team's pipeline but let them focus up-market on higher-ACV deals. The only way to make the hard trade-off was piloting to build conviction.

Self-hosted Retool was previously gated behind a sales conversation
They made it self-serve to anyone, a controversial pricing move
Demand for self-hosted came from regulated/PII needs at any company size
It cut sales pipeline volume but refocused reps on higher-ACV deals
Experimentation and piloting were the only way to build conviction

“we decided to do something quite controversial which is to take the thing that our sales team was gated on a self-hosted version of retool…”
Krithika Shankarraman · 1:00:30

#retool#pricing#self-serve#experimentation

Story1:05:30

Stripe Relay's Flop: When Timing of the Market Kills a Good Idea

Stripe Relay, launched in 2014, was meant to let e-commerce companies tap into social commerce via buy buttons. It launched to fanfare and then failed to produce the revenue expected. The lesson: even with conviction, the timing of the market — and of multiple parties needed to make a platform work — really matters. They hadn't gone deep enough on market dynamics or user research.

Stripe Relay (2014) targeted social commerce buy buttons
It launched to fanfare but failed to hit revenue expectations
Platform success needs multiple parties to align at the right time
They under-invested in market-dynamics and user research before launch

“the timing of the market really matters and the timing of multiple parties coming together to make a platform work really matters.”
Krithika Shankarraman · 1:06:30

#stripe#failure#product-timing#platforms

Takeaway· 4

Takeaway18:00

Stop Optimizing Vanity Metrics — Measure Pipeline and Revenue

Most marketing metrics — clicks, views, impressions — are vanity numbers. What actually matters is impact on signups for a self-serve/PLG product, or sales leads, pipeline, and revenue for a B2B company. At Retool, this diagnostic revealed their paid social channels were doing almost nothing, forcing investment in new engines.

Clicks, tweet views, and impressions are vanity metrics
Track signups (PLG) or sales leads, pipeline, and revenue (B2B)
A channel can look good on the surface but drive no real pipeline
Retool's diagnostic found paid social was underperforming

“A lot of marketing metrics again tend to be vanity metrics. They tend to be about the number of clicks that you got, number of…”
Krithika Shankarraman · 18:00

#metrics#growth#retool#b2b

Takeaway36:30

A Brand Is an Expectation You Create — Not Just Marketing Artifacts

Krithika reframes brand as the expectation you set in your audience's mind, spanning every touchpoint — product experience, how support resolves tickets, even the candidate experience. Companies that prize velocity often treat brand and speed as separate silos, but consistency in the brand actually empowers the org to move faster.

A brand is an expectation you create within your audience
Brand spans product, support, and recruiting — not just marketing artifacts
Velocity-obsessed companies wrongly silo brand and speed
A strong, trusted brand makes launching new products dramatically easier

“a brand is an expectation that you create within your audience.”
Krithika Shankarraman · 38:30

“And the brand is not just marketing artifacts. It is your product experience.”
Krithika Shankarraman · 37:00

#brand#trust#consistency

Takeaway55:00

Taste Becomes the Differentiator in the AI Era — Build It Through Exposure Hours

As AI makes drivel generatable by anyone, Krithika argues taste becomes a distinguishing factor: the companies that stand out show their craft and connect deep product and customer understanding. To build taste, she echoes Vercell's Guillermo on 'exposure hours' — deliberately increasing your exposure to great work, which at Thrive takes the form of a share channel for things that resonated.

AI makes generic output free, so taste becomes the differentiator
Standout companies show craft and connect product to customer understanding
AI should augment human craft, not subsume it
Build taste by increasing 'exposure hours' to great work

“the the companies that are going to distinguish themselves are the ones that show their craft, that they show their true understanding of the product,…”
Krithika Shankarraman · 56:00

#ai#taste#craft#creativity

Takeaway1:10:30

Her Life Motto: The Delta Between Expectations and Reality Is the Function for Unhappiness

Krithika's most-repeated principle: unhappiness is a function of the gap between expectations and reality, and it's far easier to change expectations than to change reality. She spends disproportionate energy setting expectations — externally with customers through marketing, and internally with stakeholders and teams about trade-offs and decisions.

Unhappiness is the delta between expectations and reality
It's much easier to change expectations than to change reality
She invests heavily in setting expectations externally and internally
Applies to customer marketing and to internal stakeholder alignment

“it is much easier to change expectations”
Krithika Shankarraman · 1:10:30

#philosophy#leadership#expectations

Key ideas

Frameworks in this episode

Support-Rotation Content Backlog

The 20/80 Marketing Review Checkpoints

The Comb-Shaped (Chameleon) Marketer

The DATE Framework for Marketing Strategy

The Expectations-Reality Delta

People & resources mentioned

People · 14

Resources · 33

Insights & moments

◆Hot Take· 3

Why Copying Another Company's Playbook Doesn't Work

Good Process Speeds a Company Up — And the 80% Review Rule

Why ChatGPT Owns Mindshare Even When Claude Is Better at Some Things

✶Explainer· 2

What Marketing Actually Did at OpenAI: Create the Use-Case Epiphany

The Chameleon CMO: Go From T-Shaped to Comb-Shaped

❝Story· 6

ChatGPT Enterprise Leads 40x'd Overnight — So She Coded Her Own Lead Scorer

How Retool Zigged: Customer Storytelling Instead of Copycat Content

Stripe's Reverse-RFP Play: Ranking for 'Payfac' Without Selling It

Turning Support Tickets Into a Self-Serve Marketing Funnel at Stripe

The Controversial Retool Bet: Making Self-Hosted Self-Serve

Stripe Relay's Flop: When Timing of the Market Kills a Good Idea

▲Takeaway· 4

Stop Optimizing Vanity Metrics — Measure Pipeline and Revenue

A Brand Is an Expectation You Create — Not Just Marketing Artifacts

Taste Becomes the Differentiator in the AI Era — Build It Through Exposure Hours

Her Life Motto: The Delta Between Expectations and Reality Is the Function for Unhappiness

Hot Take· 3

Explainer· 2

Story· 6

Takeaway· 4