Interview, Fireside Chat
Instagram Founders Reveal Their New App "Artifact"
- The core product thesis prioritizes "core recommendations" driven by machine learning and data analysis to serve individual user interests rather than maximizing company-side metrics like ads or sales.
- The recommendation engine operates on the premise that users fall into ideological and interest-based clusters, allowing for accurate initial suggestions (e.g., linking San Francisco residents to tech news) even if the user perceives their tastes as unique.
- Founders aim to bootstrap high-quality recommendations by leveraging early press coverage to acquire hundreds of thousands of signups on the first day, creating an initial data set necessary to model niche interests like "hi-fi audio systems" or "Japanese architecture."
- Current retention metrics for the non-social component focus on daily engagement with fresh content and the ability to discover new stories aligned with specific user interests, distinguishing the product from early Instagram which relied primarily on novelty filters rather than a dynamic network.
- Internal beta metrics (1,000–2,000 users over one year) have been surpassed by public usage regarding the number of stories consumed per day, validating interest beyond the "four degrees of separation" from the founders.
- Social retention benchmarks currently track specific engagement signals in the closed beta, including the ratio of posts with reactions versus posts with no engagement ("speaking into the void") and the volume of answered versus unanswered direct messages.
- Product development strategy adheres to the "Jobs to be Done" theory, where features are only added if they directly support the core user need of a "digital assistant" keeping users connected to deep interests without causing "product dissonance."
- The current roadmap emphasizes a simple "vanilla" experience, with complex features, social networking capabilities, and engagement tools deferred until the core recommendation engine is robust enough to justify the expanded scope.
- The team acknowledges the difficulty of scaling data to support hyper-niche interest groups but intends to expand the product's purview over time as the user base grows and more behavioral data becomes available.