Snowflake's CMO Cut Marketing Cost 30% by Ditching Dashboards

Snowflake's marketing team delivered a 30% reduction in cost per opportunity over six months, CMO Denise Persson said at SaaStr AI 2026. The improvement came after she stopped logging into a dashboard each morning and started asking her data plain-English questions instead.

Persson oversees a 700-person marketing organization. She said she no longer sends Slack messages asking "why did pipeline move in US West?" because she can now interrogate the data directly. The shift killed the sales-marketing data war. Both sides see a single source of truth on deal sourcing and attribution. The arguments over whose dashboard is right simply disappear.

She pulled fragmented media channels into one place and let the system recommend daily optimizations. The campaign no longer waits until it ends to learn what failed. That change drove the 30% cost-per-opportunity improvement. The daily brief now covers pipeline, org health, new hires, departures, and even travel expenses she used to review manually. Intelligence that once lived only with finance is now a question she asks before her first meeting.

At Snowflake's scale, you cannot let agents run unchecked. A wrong email to a customer is a brand impression that lasts. The solution was a centralized AI engineering team sitting above everything. Any skill used by more than a few people must be certified before it ships. The company-wide GTM agent, Raven, is used across sales and marketing, and every skill inside it is centrally certified. The dual job of that team: make sure agents behave correctly, and stop the company from building the same agent five times.

Persson was candid about the mistakes. A leaderboard ranked on token consumption rewarded activity over outcomes. High token usage meant high cost, not necessarily high results. She now caveats the leaderboard every month, telling the team that business outcomes matter more than token count. If you have to verbally correct your own metric every time you show it, the metric is sending the wrong signal.

AI spend sits at the company level, and marketing gets effectively unlimited access right now. The CEO did not want anyone's departmental budget to throttle experimentation. Persson conceded that usage is going through the roof and the bill is real. She expects to pull back in 2026. Unlimited access buys adoption speed and a bill you eventually have to reckon with.

The activation layer is still half-built. Analysis is automated: which use case to promote to which account, a workflow that used to eat enormous time. What they have not cracked is full activation. The campaign still cannot fully launch itself. That gap drives demand for GTM engineers, the most resource-constrained role in the building, Persson said.

Events are surging. Demand for in-person experiences is, in her words, going off the roof. Enablement moved under marketing to avoid content duplication. Snowflake builds content once, generates every derivative asset for every segment, and ships self-service enablement agents so sellers get training on demand. Roleplay agents let reps practice a pitch against an agent loaded with company intelligence instead of cornering their manager.

Persson said authenticity is becoming high value precisely because so much is now synthetic. Trust in a brand is where humans spend their time now, on the uniqueness the agents cannot manufacture. Her closing point: nobody can paint a clear picture of what marketing looks like in three years. The choice she puts to her team is to shape it or opt out.