AMA: Coinbase Head of Product Marketing, Base, Sarah Khogyani Wolf on Product Marketing KPI's

I always start by understanding the structure and maturity of the business and product. If we’re early-stage or pre product-market fit, I embed PMMs directly within product pods. This ensures tight alignment with product and eng, so marketing can shape narrative, validate positioning, and deeply understand user needs as the product evolves. As the org matures and we’ve validated growth segments, I prefer to organize by audience. This enables PMMs to think holistically across multiple products an ...Read More

The biggest challenge is isolating metrics that marketing can truly own, and that matter. Too often, PMMs get judged on metrics they influence but don’t control, like revenue or DAUs, without a clear line to how their work contributes. Many times, the right metric simply doesn’t exist yet. You have to build the plumbing - define the funnel, implement tracking, clarify the conversion path - before you can set a KPI. I’ve tackled this by partnering closely with data science and product to co-creat ...Read More

This is a common challenge, especially in B2B or hybrid funnels. A few tactics I’ve seen work: Lead-stage tagging: Tag leads based on the assets they interacted with (e.g., downloaded a one-pager, attended an event). Over time, you can correlate these touchpoints with downstream conversions. Attribution scoring: Build a multi-touch attribution model that weighs influence across content types. Usage surveys: Post-conversion, ask “What content influenced your decision?” Time-to-close impact: Measu ...Read More

For adoption, I focus on top-of-funnel metrics like acquisition or activation tied to specific campaigns. Campaign attribution models (even simple UTMs) can help measure this impact clearly: did a product launch or announcement move the needle on usage? For retention, I partner with data science to look at post-adoption engagement: are users sticking around, using the product repeatedly, and growing in value over time? For example, we might track 7-day and 30-day retention, feature stickiness, o ...Read More