Technical Spike
A time-boxed investigation designed to answer a specific technical question or reduce uncertainty before committing to a development approach. Spikes produce knowledge and recommendations rather than shippable product increments.
Technical spikes are essential when the team faces significant uncertainty about feasibility, performance, or implementation approach. By allocating a fixed amount of time, typically one to three days, to investigate a question, teams avoid both analysis paralysis and the risk of committing to an approach that will not work. The output of a spike is a written recommendation with supporting evidence.
AI product development generates more technical spikes than traditional software because so many questions can only be answered empirically. Can the model distinguish between these ten categories with acceptable accuracy? How much does latency increase when we add retrieval augmented generation? Will the embedding model fit within our memory budget? Growth teams benefit from spikes that answer questions like: Can we personalize onboarding flows using existing behavioral data? What is the minimum amount of user data needed for recommendations to outperform a popularity-based baseline? These spikes turn unknown unknowns into actionable information that de-risks both product and growth investments.
Related Terms
Product-Market Fit
The degree to which a product satisfies strong market demand. Achieving product-market fit means customers are actively seeking, using, and recommending your product because it solves a real and pressing problem for them.
Jobs to Be Done
A framework that defines customer needs as functional, emotional, and social jobs people hire products to accomplish. It shifts focus from demographic segments to the underlying progress customers are trying to make in specific circumstances.
Minimum Viable Product
The simplest version of a product that can be released to test a core hypothesis with real users. An MVP delivers just enough functionality to gather validated learning while minimizing development time and cost.
Minimum Lovable Product
An evolution of the MVP concept that emphasizes delivering enough quality and delight that early users genuinely love the product. It balances speed-to-market with the emotional engagement needed to drive organic word-of-mouth growth.
Design Sprint
A five-day structured process for rapidly prototyping and testing ideas with real users. Developed at Google Ventures, it compresses months of debate into a focused week of mapping, sketching, deciding, prototyping, and testing.
Lean Startup
A methodology for developing businesses and products through validated learning, rapid experimentation, and iterative releases. It emphasizes reducing waste by testing assumptions before building fully-featured solutions.