ProfileDenise Persson marks a decade at SnowflakeFrom floppy disks to data agentsOne pipeline, one plan, six months to earn trustProfileDenise Persson marks a decade at SnowflakeFrom floppy disks to data agentsOne pipeline, one plan, six months to earn trust

The Operator's Desk / Marketing

Denise Persson Built Snowflake's Marketing Engine on a Six-Month Bet

Before the dashboards, Denise Persson carried leads to sales on a floppy disk. Three decades later, the Snowflake CMO still believes the decisive technology in a growth company is trust.

The first data pipeline Denise Persson managed could fit in a jacket pocket. A lead arrived by coupon, fax, or business card. She typed it into a spreadsheet, saved the file to a floppy disk, then walked it over to sales. The process was slow, manual, and wonderfully unforgiving. Someone on the other side either found the lead useful or did not.

There were no attribution models to soften the answer. No dashboard could turn activity into achievement. Marketing met sales at a literal handoff, and Persson learned the lesson that would follow her through four CMO jobs: the work has to move the business.

That lesson sounds obvious. Inside a growing company, it can be radical. Departments develop their own vocabulary, their own scorecards, and their own reasons the number did not arrive. Persson's operating instinct is to collapse that distance. At Snowflake, the cloud software company she joined in 2016, it took about six months before sales trusted marketing's leads. She did not try to talk the clock faster. She let delivery do it.

01 / The long handoff

A career that crossed the whole digital arc

Persson grew up in Stockholm. Her sister worked in advertising and introduced her to campaigns and communications. A friend of that sister then persuaded her to join Commodore in the early 1990s. Technology was not the destination Persson had mapped for herself, but it became the terrain on which she learned to build categories.

In 1996 she joined Genesys, a French collaboration-software startup. Over 12 years she moved from Sweden to France and eventually the United States, rising to executive vice president of global marketing. Genesys grew from roughly $20 million to more than $200 million in revenue during her tenure, built a customer base that included more than 200 of the Fortune 500, went public, and was later acquired.

Four countries, four CMO roles, and one enduring question: did the work create movement?

ON24 recruited her as CMO in 2008. Apigee followed, and Persson helped the API company through its 2015 public offering before Google acquired it in 2016. Each stop sharpened the same craft: name a category, build a global team, and connect a technical product to a business problem a customer can recognize.

“Marketing has to move beyond vanity metrics and hold itself accountable to revenue impact.”Denise Persson
02 / The trust test

Six months without a shortcut

Snowflake was still small when Persson arrived. It had about 100 people and roughly $3 million in revenue. Chris Degnan, its founding chief revenue officer, had a sales organization accustomed to running its own demand generation. A new marketing chief could not expect automatic faith from a team that had learned to feed itself.

Persson's opening position was unusually plain: marketing serves sales. The job was to create demand, build pipeline, equip reps, and help close business. Awareness mattered. So did thought leadership and brand. But the test was downstream. Did a rep have a credible reason to call? Did the customer story travel? Did pipeline become revenue?

~100Employees when Persson joined Snowflake in 2016
6 mo.The trust-building window she describes
~700People in the marketing organization she built

For half a year, that answer accumulated lead by lead. Quality improved. Pipeline strengthened. Skepticism receded because the inputs changed. The interesting part is not that trust appeared. It is that Persson treats trust as an operating result, produced by repeated evidence, rather than a cultural mood produced by a workshop.

Her partnership with Degnan eventually became one of the longest-running CMO-CRO pairings in enterprise software. They handled disagreements directly. They resisted public blame. Marketing did not declare victory by pointing at a pile of MQLs, and sales did not keep a private definition of success. The shared scoreboard was new business.

A thousandfold revenue comparison is less a campaign result than a reminder that systems have to survive their own success.
03 / Category before campaign

Build the hill, then own it

Persson compares positioning to a house's foundation: build it badly and the walls crack. The metaphor is useful because it puts order back into marketing. A team can buy reach before it knows what it means. It can publish constantly while avoiding the hard choice of which category it wants to own.

Snowflake's language evolved with the company, from a data warehouse built for the cloud toward the Data Cloud and then the AI Data Cloud. The labels widened as products, partners, and customer use cases widened. Yet Persson's preferred narrator remained the customer. In enterprise technology, a peer explaining a solved problem carries a kind of credibility the vendor cannot manufacture alone.

That explains the mix of rigor and theater in her work. Snowflake can tell a technical story through Olympic sport, where bobsled performance turns milliseconds and sensor readings into a plot anyone understands. The company can convene executives for peer learning instead of giving them another vendor lecture. The spectacle opens the door; the customer's evidence keeps the conversation in the room.

1 / Position first

Choose the category before multiplying campaigns.

2 / Share the number

Measure marketing where pipeline becomes business.

3 / Let customers speak

Use peer stories for both education and proof.

4 / Build to scale

Automate repetitive work before it becomes an organization chart.

04 / The cold water

The promotion that arrived before confidence

Persson's defining career moment came at 27. She had been at Genesys for three or four years when the CEO told her the marketing leader was moving on. Persson would take the job the following week. Fifty senior executives were flying in from around the world, and she would present the marketing plan.

She felt nowhere near ready. She was frightened. She said yes.

The story matters because Persson now sees the other side of that exchange. One of the pleasures of leadership, she says, is recognizing potential and giving people opportunities that can alter their lives. Readiness is not always a credential that arrives before responsibility. Sometimes another person sees the next version of you first.

“Effective leadership is about clarity, accountability, and trust.”Denise Persson

Her description of her own style is direct: hire capable people, set a high expectation, and give them ownership. She has little patience for an unfinished strategy presented as finished thinking, but pairs that standard with support for people executing with purpose. She also values truth-tellers. Persson has described her husband as the person who grounds her and offers the feedback that can hurt, precisely because it is honest.

This makes trust more demanding than friendliness. A trusted colleague is allowed to say the plan is weak. A trusted leader can hear it without punishing the messenger. The loop takes years to build and seconds to break.

05 / After the dashboard

AI as an operating practice

Persson's career began before digital marketing and now sits inside the rush toward agentic software. Her position is enthusiastic but conditional: an AI strategy begins with a data strategy. If the data is fragmented, poorly governed, or untrusted, a faster interface only delivers a faster argument.

She has described starting her day by asking questions of data in natural language instead of opening a static dashboard. The change is subtle but consequential. A dashboard tells a leader what happened inside a frame someone designed earlier. A conversation can keep moving: why did it happen, where did the pattern change, and what should the team inspect next?

Inside Snowflake marketing, Persson established an AI council made up of curious people from different functions. The group experiments, then shares what works and what fails during quarterly AI marketing days. It is a practical answer to a common organizational problem. Curiosity is plentiful but uneven. A council turns scattered enthusiasm into a learning system.

The human obligation grows with the tool. Brands need to explain clearly how customer data is used, where it goes, and how someone can opt out. Trust remains the asset underneath the new interface. The technology changed again. The handoff did not.

06 / Writing the operating manual

A decade becomes a field guide

In 2025, Persson and Degnan published Make It Snow: From Zero to Billions. The title plays with the brand, but its structure is operational: startup, build, scale. The pair wrote about product-market fit, partners, culture, hiring, global expansion, and the work required to keep sales and marketing aligned after the company becomes too large for alignment to travel by instinct.

The useful idea is that harmony is not the goal. A growth company needs arguments about customers, positioning, territories, and priorities. What Persson rejects is the wasteful version, where two functions defend separate scoreboards and ask the CEO to referee. Direct conversation keeps the argument attached to the work.

By 2026 she had spent ten years at Snowflake. Forbes placed her at No. 20 on its annual list of influential CMOs. She was also looking outward, through board and advisory work with technology companies and Patricia Industries, and forward, toward a Snowflake that could move from the top 20 software companies into the top 10.

The ambition is large. The method remains almost stubbornly concrete. Pick the category. Listen more than you talk. Put a useful lead in a rep's hands. Share the number. Tell each other the truth. Then repeat the loop at a scale where no one can carry the data across the office anymore.