12: How to build an experimentation culture (even with low traffic) w/ Lucia Van Den Brink

Lucia Van Den Brink has run nearly 1,000 experiments in 14 years.She's worked with 100M+ brands, founded The Initial to help companies build experimentation in-house, co-founded Women in Experimentation, and teaches at CXL.I invited her on Love at First Try to talk about something most SaaS teams get wrong: how to actually make data-driven decisions without overcomplicating it.This episode is for you if you've ever thought "we don't have enough traffic to test" or "A/B testing is too expensive for us." Lucia challenges both of those assumptions.🧠 What you'll learn in this episode:0:00 - Intro: what this podcast is about and who it's for0:25 - Who is Lucia and why experimentation matters for smaller SaaS teams2:59 - The real difference between random testing and building an experimentation culture6:36 - Why experimentation is actually about scaling leadership (insight from Booking.com)9:15 - What counts as an "experiment" beyond simple A/B tests11:32 - How a news website validated a 6-month feature before building it12:57 - Why starting with removing elements is one of the biggest growth levers16:22 - How to validate your data before trusting any experiment19:05 - How to prioritize what to test (and where to start)21:34 - Why you shouldn't segment your tests when you're just starting25:41 - How to measure the right KPIs (including delayed metrics)31:28 - Why you should never measure just one metric36:45 - Real example: reducing churn in the first two weeks with a get started page43:10 - Why "obvious UX improvements" still need testing (the humbling 20-30% win rate)48:09 - The biggest mindset shift from junior to senior in experimentation💡 Actionable insights from Lucia:1. Start by removing, not addingOne of the biggest growth levers is removing elements from your pages. Most tools let you hide elements without code. Try removing a field, a section, or a step in your funnel. You'd be surprised how often less friction means more conversions.2. Run an AA test before any real experimentBefore you trust your data, run an empty test (control vs. control). This tells you if your tracking actually works. Skip this and you might be making decisions on broken data.3. Measure multiple KPIs, not just onePick a main metric, but always track 2-3 supporting metrics. If you're testing onboarding changes, measure signups AND activation AND delayed metrics like paid conversions. One number never tells the full story.4. Consider negative testingInstead of building a big new feature to test, try removing the opposite. Want to know if onboarding calls help? Test what happens when you remove them. You learn faster and cheaper.5. Calculate the business case for "losing" metricsSometimes a test hurts one metric but helps another. If showing a phone number increases support calls but also increases conversions, do the math. The revenue might cover the cost.