Marketplace Tactics: Metrics, Growth vs. Focus, Business Strategy | Field Notes by Connie Chan

Ancestry's AI Strategy and Data Expansion

• AI Efficiency: AI reduced the digitization time for the 1950 US Census from nine months (manual) to nine days.
• Data Scale: Ancestry holds 40 billion records, 75% of which are exclusive.
• Growth Targets: 5 billion pieces of content were digitized in the past year; 15 billion records are projected for 2023 via AI integration.
• Future Vision: The company aims to automate family storytelling by transcribing oral histories, digitizing personal letters/photos, and generating family books without manual user input.

Marketplace Strategy and Lessons from Meta

• Core Philosophy: Successful marketplaces must identify a "right to win" in a narrow category before expanding to concentric circles of adjacent categories.
• Differentiation vs. Scale: Marketplaces like Etsy or Airbnb succeed by focusing on specific niches (e.g., handmade goods or local rentals) rather than attempting immediate breadth, though long-term goals often involve recreating broad platforms.
• Risk of Over-Expansion: Expanding too quickly into non-core categories (e.g., 3D printed goods or mass-manufactured items) risks damaging brand authenticity and alienating core buyers.
• Trust Architecture: Reputation systems can become double-edged swords; eBay's original model allowed retaliatory negative feedback from sellers, suppressing buyer complaints and degrading trust.
• Policy Fixes: Trust models must prevent retaliation, such as allowing simultaneous feedback release (Airbnb) or prohibiting sellers from giving negative buyer reviews (eBay's later policy).
• Incentive Design: Platforms must gamify behavior to encourage quality service and competitive pricing, rather than merely preventing bad behavior.
• Fraud Prevention: Founders must architect systems to prevent seller, buyer, and collusion fraud from day one.

Metrics, Growth, and Retention Dynamics

• Metric Shift: Growth stage marketplaces should prioritize retention curves (30, 90, 180-day) and frequency of visits over raw GMV.
• Transaction Frequency: Chasing high-value transactions (e.g., mobile phones) can depress frequency; focusing on low-value, high-frequency items (e.g., toothpaste) drives deeper engagement and ecosystem lock-in (e.g., Prime).
• Leaky Bucket Problem: Even with high acquisition spend, marketplaces fail if they cannot retain users, as seen historically with eBay's churn issues.
• Browsing Incentives: For browse-heavy marketplaces, UI must feel dynamic (e.g., reshuffled "Top Picks") to maintain daily traffic even when new inventory is low.
• Contextual Personalization: Algorithms must distinguish between a user's "home routine" behavior (repetitive purchases) and "discovery" behavior (new locations/categories).

Career Advice and Executive Leadership

• Hiring Challenges: Deb Liu joined Ancestry via Zoom without meeting the team, noting the importance of building trust remotely; the Ancestry team also conducted "blind" interviews where executives met her before knowing her identity.
• Advocacy Principle: Employees must explicitly state their career goals to management; silence is often misinterpreted as a lack of ambition.
• Rejection Management: Founders and leaders should view rejection as a numbers game; the goal is to secure approval from the "right one" rather than everyone, requiring resilience to hearing "no."
• Narrative Arc: Liu's career spans connecting people through commerce (Facebook) and connecting people to their past (Ancestry), unified by the mission of using technology to bring people closer together.