By Teresa Torres · producttalk.org · @ttorres on X · LinkedIn
Teresa Torres and Petra Wille examine the build vs. buy decision, which AI is reshaping. They argue the classic rule of thumb, to avoid building non-core functions, has limits, since vendors may be better positioned even for core-adjacent needs like payments. Torres contrasts moving her blog to Ghost rather than building a platform with building her own task management system, where personal workflows and data ownership tipped the balance. The conversation covers how cheaper prototyping and vibe coding lower building costs, why data portability matters for vendor lock-in, and why teams should treat the decision as a discovery problem by testing feasibility and trialing vendors. They close by warning that being able to build something does not mean one should own it.
01Key takeaways
- Ask whether the thing is core to your value, not just whether you can build it.
- Even core functions like payments or infrastructure may be better served by specialized vendors.
- Evaluate data portability beyond simple exports before committing to a tool.
- AI and vibe coding lower prototyping costs, making small targeted tools more viable.
- Test build and buy options like discovery assumptions: prototype, trial vendors, and validate.
- Account for long-term maintenance and hidden complexity, not only the initial build cost.
02Key sections
- Why build vs. buy keeps recurring
- The decision appears in every product org at every stage, and AI makes building feel easier without removing the underlying tradeoffs.
- The core value rule and its limits
- The traditional advice is to avoid building anything not core to your value stream, but even core areas like payments may be better handled by vendors.
- Two contrasting personal examples
- Torres moved her blog to Ghost rather than building one, yet built her own task manager because her workflows were highly personal.
- Data ownership and vendor lock-in
- When underlying data becomes the real product, owning and controlling it changes the calculation, and exports alone may not guarantee portability.
- Treating the decision as discovery
- Options should be treated as assumptions to test through feasibility prototypes and real vendor trials rather than trusting marketing claims or gut feel.
03From the post
“Listen to this episode on: Spotify | Apple Podcasts Build vs. buy is one of those product decisions that never goes away — and AI is changing how teams think about it (but maybe not in the way you expect). In this episode of All Things Product, Petra Wille and Teresa Torres”
04Frameworks mentioned
Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.