Conversation
Highlights

How GitLab’s Communications Team Workshopped One CNBC Analogy for Months Before Their IPO

The quiet period before an IPO compresses your entire company story into a single day. Everything about your business is already documented in the S-1, and when the embargo lifts, you get one shot to control the narrative. For Natasha Woods, who ran PR at GitLab through their public offering and five subsequent earnings cycles, the constraint wasn’t information—it was translation.

In a recent episode of The Narrative, Natasha, now Senior Director of Communications at The Linux Foundation, broke down how a single analogy—”GitLab is the iPhone for DevOps”—required multiple workshop sessions between CEO Sid Sijbrandij, the CLO, and the communications team before it was camera-ready for CNBC.

The Broadcast Translation Problem

CNBC segments run 60 seconds to three minutes. You get three questions maximum. No room for technical architecture explanations or feature walkthroughs. The constraint forces you to find language that creates instant recognition—not just comprehension, but the visceral “oh, I get it” moment that reads on camera.

“If you’re going to go on broadcast and you have to explain what a DevOps platform is to a broad, mainstream audience, it’s not the easiest thing in the world,” Natasha said.

She reframed the challenge by asking Sid: “Your parents aren’t in tech. They’re from the Netherlands. They’re not original English speakers. You’re sitting around the dinner table and you’re with your siblings and you’re talking about what a DevOps platform was. What would you use analogy of to explain it to them?”

The question forced abstraction away from product features toward functional understanding—how do non-technical people already think about consolidation?

Multiple Iterations Before Air-Ready

The brainstorming involved Sid, the CLO, and Natasha testing different frameworks. They weren’t optimizing for accuracy—they were optimizing for delivery under pressure, on camera, with no second takes.

“He goes, well, it’s kind of like the iPhone,” Natasha recalled. “And I was like, let’s go down that path. Why is it an iPhone? And he goes, well, before you’d have like a calculator, a Walkman, a phone. You’d have all these different things and you’d have all these different tools that you used. And now every single thing is in an iPhone. He goes, that’s basically what a DevOps platform is. You have all these different tools that developers would use, but we’ve brought them all into one single place, one single platform.”

The analogy mapped technical consolidation to consumer experience everyone already internalized. But Natasha didn’t stop at concept validation. She pushed: “So GitLab is the iPhone for DevOps?” When Sid confirmed, she asked the delivery question: “Can you say that on air?”

Testing whether the executive can actually deliver the analogy naturally—not just understand it intellectually—is where most prep falls apart.

Live Validation on CNBC

When Sid delivered the analogy during GitLab’s first post-listing CNBC interview, the prep work validated immediately. “You see this recognition in the host’s eyes and they immediately know what he’s talking to. They immediately connect with him. They immediately smile,” Natasha said.

She was monitoring from NASDAQ’s studio. “I was like, yes, this is great. And I’m with the NASDAQ comms people and they’re like, they’re looking at me and they’re like, this is amazing. I was like, I know, I’m so excited this landed.”

That facial recognition moment—the host processing and accepting the analogy in real time—is what months of workshop sessions bought. It’s non-linear ROI that’s impossible to measure in traditional attribution models but fundamentally shapes how your story propagates.

Customer Narrative Extraction Over Founder Vision

Natasha’s narrative development methodology inverts the typical founder-first approach. When she joined Interana, a behavioral analytics startup founded by two ex-Facebook engineers who built the behavioral analytics backend for Facebook, she started with: “Who are your customers?”

The answer: Tinder. This was 2014—Tinder was relatively new, dating app behavioral analytics wasn’t an obvious category, and the use case was non-obvious.

“I was like, I’m sorry, Tinder uses you for like matches and all these backends?” Natasha asked. Instead of pitching the founders’ Facebook pedigree or technical capabilities, she called Tinder’s communications team to extract why they chose this specific technology.

Then she pitched a reporter she knew who had just moved to Fortune, framing it around how Tinder’s matching system actually worked behind the scenes—using the customer story as the narrative vehicle.

The story took nine months to develop. “I literally went on maternity leave and came back. That’s how long it took to put the story together. But we ended up with this really cool piece in Fortune around how Tinder utilizes this behavioral analytics technology that nobody has heard of, but you should.”

The methodology: “Every founder is like, I want to be in Fortune. I want to be in Forbes. I want to be in the New York Times. And it really came about because we went in that customer route.”

When Founder Ego Kills Coverage

Natasha shared a failure mode: launching a big data company founded by an ex-VMware R&D executive. She secured a call with a prominent tech journalist and prepped the founder on expected questions, including competitors.

When the journalist asked about competitive landscape, the founder responded: “Well, you obviously didn’t do your homework because why are you asking me who my competitors are? You should know this if you were a great journalist.”

“I was mortified. I was absolutely mortified,” Natasha said. She intervened to end the call early, then separately apologized to the journalist and called the founder to debrief what happened.

The journalist wrote the story but significantly truncated it. “It was very short, direct to the point. It wasn’t what I was hoping for originally when I had pitched it, but I knew that was probably the result if he wrote it at all.”

The failure mode: treating journalist questions as knowledge tests rather than opportunities to shape quotable narrative. Journalists ask questions to extract your perspective and get on-record statements, not because they lack information.

Trading Mixed Coverage for CEO Transition Credibility

When managing Andela’s CEO transition—Jeremy Johnson stepping down, Carol Chang (from Uber) taking over—Natasha inverted the typical “only positive news” approach with Forbes reporter Alex Conrad.

She asked: “What are the numbers you need me to get for you to make sure that you and your editor are comfortable covering this?”

Conrad specified: revenue growth metrics, actual valuation, forward projections. The constraint: Andela’s revenue hadn’t grown as projected in Conrad’s previous coverage of the company.

Natasha went back to Jeremy and Carol with the ask. “I was like, but I advise I think this is a good idea. I think it’s very authentic for the company to get out there, and it also builds that trust during a CEO transition. And they said, yep, we agree with you. Let’s get you those numbers.”

Conrad’s article noted valuation and stagnant sales growth explicitly but framed the CEO transition as positive inflection. “Overall, it was a win,” Natasha said. “And I think what happens is a lot of people just want these overly positive articles, and it’s like, no, neutral, transparent, authentic is the way brands should go.”

The trade: accepting mixed coverage to build institutional credibility during leadership transition. Credibility compounds; puff pieces don’t.

Conference Timing Is Relationship Infrastructure, Not Venue Selection

Founders default to “announce at the conference” without evaluating embargo management capability or relationship infrastructure. Natasha’s KubeCon approach: send complete news packages to media three weeks pre-event so reporters can file stories in advance.

“A lot of them had their stories filed to publish day one of the conference when the embargo lifted,” she explained.

The decision matrix for conference announcements: competitive timing, reporter relationship depth, whether your news fits keynote narrative (for roundup inclusion), and your team’s embargo coordination capability.

“I always make sure that the reporters get my news several days before the conference,” Natasha said. The strategy isn’t presence at the event—it’s pre-briefing the right reporters with sufficient lead time for quality story development.

If you lack the relationships to pre-brief effectively, conference timing doesn’t buy you anything except ambient noise.

Why CRO-Owned Marketing Destroys Brand Equity

Natasha sees accelerating consolidation of marketing under Chief Revenue Officers—a structural shift she believes fundamentally undermines brand building despite short-term efficiency gains.

“When you make comms or that kind of top of funnel, corporate marketing team deliver the results from a sales perspective, you lose a lot of that narrative. From a storytelling perspective, you lose a lot of that creativity, you lose a lot of that connection points,” she explained. “When you start reporting your comms metrics to a CRO, there’s a huge shift.”

The failure mode: comms metrics get evaluated through pipeline contribution, forcing storytelling into demand-gen attribution frameworks. This kills the creative oxygen required for top-of-funnel brand work.

“I think they’re making the shift from like an efficiency, financial, business perspective. And I don’t think they’re really looking at what is this going to do to my brand long term.”

Her position: maintain tight sales-marketing partnership while keeping brand work structurally separate from revenue accountability. “If you get a CRO in there that really understands marketing and will decouple that top of funnel, it could potentially work. I just, I haven’t really seen it yet.”

Brand lovability drives funnel performance, but measuring it by direct pipeline contribution destroys the space it requires to function.

Earnings Operations Are Approval Chain Orchestration

Running earnings at a public company isn’t content creation—it’s coordinating legal, finance, and executive sign-offs under time compression with last-minute changes guaranteed.

Natasha’s operational focus during GitLab’s five earnings cycles: ensure financial website content accuracy for investor access, and shepherd the press release through multiple mandatory review layers.

“There’s so many different reviews that need to happen internally before you can release something. And then you’re always going to get something last minute, have to make sure it’s changed. You need to know who needs to review it. Having everyone’s phone numbers on hand is really helpful during that,” she said.

The second component: executive interview prep. GitLab CFO Brian Robbins (now at Snowflake) would practice on weekends in front of his wife and mirror before Natasha’s sessions. “This really came through in all of his interviews and you could just see the progression over time that he spent the hours that he needed to make sure that he showed up well in those interviews and he was constantly invited back.”

The practice optimized for bridging mechanics when asked unexpected questions, body language alignment with verbal messaging, and facial expression management on camera—not answer memorization.

Press Release Structural Changes for LLM Parsing

Traditional press release format (headline, subhead, body copy) fails LLM extraction. The structural fix: insert 3-5 bullet points under the subhead, before body text, that summarize key information.

“That’s what’s getting picked up by AI. That’s what’s going to get picked up on like Perplexity and like other AI search engines,” Natasha explained.

Earnings releases now follow this pattern universally. If your press releases still use traditional structure, you’re not getting indexed by how content actually gets discovered in 2025.

“Press releases are great if you optimize them for AI, if it doesn’t have a lot of crazy jargon in it, they do not need to be five pages and you can maybe link to a survey that you’re doing or for more information.”

The format shift: front-load the summary in bullet structure for LLM parsing, minimize jargon, link to supporting materials rather than embedding everything in a five-page document.

The “You Didn’t Do Your Homework” Failure Mode

One more cautionary example: Natasha was launching a big data startup, ex-VMware R&D founder, stealth mode exit. She secured Jonathan Vanian (prominent tech journalist) and prepped the founder extensively—including the inevitable competitor question since the website wasn’t live yet and they were pre-launch.

On the call, when Jonathan asked about competitors, the founder said: “Well, you obviously didn’t do your homework because why are you asking me who my competitors are? You should know this if you were a great journalist.”

Natasha had to salvage with an apology call to Jonathan and debrief the founder on what went wrong. The article ran but significantly shorter than Jonathan’s typical depth—a visible outcome of damaged relationship.

Journalists ask questions you could answer yourself to get your specific framing and quotable material. Treating it as a knowledge test rather than narrative opportunity is how you convert potential features into minimally viable coverage.