# Making Data-Driven Product Decisions — WP Product Talk (Zach Katz & Katie Keith, guest Derek Ashauer) ## 🗂️ Triage — annotated highlights surfaced > [!note]- Cross Reference > - **Analytics-only customers churn from overwhelm, not price** — reframes retention as an onboarding/clarity problem, not a pricing one. #xref > - **Annotate personal life events on business metric timelines** — without it, an AI (or human) defaults to blaming the nearest business event for a dip. #xref #aphorisms > - **A "bad" metric can flip to good from a second angle** — the 61%-single-site figure reread through search-intent behavior; the episode's core methodological claim. #xref #aphorisms ## 🧭 Bib capsule Derek Ashauer (ConversionBridge WP), guest on *WP Product Talk* hosted by Zach Katz (GravityKit) and Katie Keith (Barn2 Plugins), 2026. https://www.youtube.com/watch?v=DG5Ih0WzhPw ## 🧠 Author's thesis (Adler Q1+Q2) "Data-driven" for a small WordPress product team is less about acquiring more data and more about three disciplined habits: track conversions and usage deliberately (not everything), validate any given number against a second independent source before acting on it, and actively check for confounders — market-wide trends, personal life events, a metric's hidden segment structure — before concluding. AI/MCP tooling (Claude connected to GA4, BigQuery, Help Scout, ClickUp, GitHub) has genuinely lowered the cost of the analysis step and even displaced a paid pricing-analysis agency, but it hasn't lowered the cost of getting the interpretation right. ## 🧩 Key arguments - Track conversions, not just visits — pageviews are the bare minimum; the real signal is every intent-bearing action (signups, purchases, form submits, cancellation surveys). - Validate, don't trust: cross-check multiple analytics platforms against actual order-export CSVs (99.9% agreement) and corroborate quantitative patterns with direct customer conversations. - Context beats the metric: a dip can be email cadence, a market downturn, or a divorce — annotate confounders and reread "bad" numbers from a second angle before acting. ## 🗒️ Notes — Literature atoms ### Analytics-only customers churn from overwhelm, not price > [!quote] > "people install analytics and then they never look at it because they feel overwhelmed and don't know what to do with it" - **Where**: ~00:18:30, Derek Ashauer, on ConversionBridge WP churn analysis - **Why**: Overwhelm — not cost or missing features — drove churn for a segment; retention becomes an onboarding/clarity problem. ### Annotate personal life events on business metric timelines > [!quote] > "I think those life events are just as important as those site redesigns... to consider in what's affecting where your business is at" - **Where**: ~00:34:57–00:36:00, annotating his sales chart with his divorce-finalization date - **Why**: Without the annotation, an AI or analyst defaults to attributing a dip to the nearest business event. ### A "bad" metric can flip to good from a second angle > [!quote] > "I think it's to make sure that you look at the data from all angles because data can actually trick you." - **Where**: ~00:48:00, closing "best advice," illustrated by the 61%-single-site-license reframe - **Why**: The episode's core methodological claim; the other atoms are instances of it. ## 🔖 Permanent candidates - **Validate a metric against an independent source before acting on it** — Before changing a product, price, or marketing decision on a single data point, check it against at least one independently-collected second source (a different analytics platform, a direct customer conversation, an order export) rather than acting on the first read. ## Links - [[Product & Startup Instincts MOC]] — home MOC; deliberate instrumentation, triangulation, and confounder-checking are core product-judgment instincts for small teams. - [[The Importance of Customer Obsession and Data-Driven Insights in DevRel at Vercel]] — same "validate quantitative signal against direct customer contact" discipline, in a DevRel context rather than a plugin-business one. - [[A Metric Is Vanity Only Against Your Own Outcome Data]] — extends the episode's second-angle rule into a test: a metric earns meaning only when checked against an independently observed outcome. - [[3. Resources/MOCs/Book Review Synthesis MOC]] — applies-in: the MOC lists this note under Existing atoms and literature. ## ✅ Processing checklist - [x] Claim atoms carried from deep-read analysis - [ ] Permanent candidates reviewed for promotion - [ ] Version History Diff check