Manu's Vault

Hypothesis-Pull

4 min readStatus:Seedling

Hypothesis Pull

Hypothesis pull is the principle that product development work should be triggered by the hypothesis that needs testing — not by a roadmap, a backlog, or a good idea for a feature.

Pull vs. Push in Manufacturing

The concept originates in lean manufacturing. In a push system, production is scheduled in advance based on forecasts: factories build inventory ahead of demand and “push” it toward customers. In a pull system, pioneered by Taiichi Ohno at Toyota, production is triggered by actual customer demand — a downstream request creates a signal (kanban) that flows upstream, authorizing only the work needed right now.

Pull systems eliminate a critical form of waste: work-in-progress (WIP) inventory. WIP sitting on a factory floor represents capital tied up in goods that haven’t yet delivered value. Large WIP amplifies problems: defects spread through a batch before being caught, and any single delay cascades across the whole inventory pile.

The Startup Application

Eric Ries adapts the pull insight to product development. In the Build-Measure-Learn-Loop, the pull signal is the hypothesis. You don’t build a feature because it appeared on a roadmap or because it seemed valuable. You build because you have a specific, testable belief — for example, “users will complete onboarding in under five minutes if we reduce the form to three fields” — and you need the minimum build required to test it.

This reframes what constitutes waste. Startup WIP is unvalidated assumption embedded in code and plans. Every feature shipped without a clear hypothesis and measurement plan is invisible inventory: effort that may never deliver learning. Like factory WIP, it accumulates silently — in unmaintained code, untested assumptions, and roadmaps divorced from evidence.

Hypothesis pull directly shapes what gets built. The question shifts from “what should we build next?” to “what do we most need to learn, and what is the minimum experiment to learn it?” The Value-Hypothesis defines whether the product creates value; the Small-Batch-Production principle ensures the experiment is small enough to complete quickly.

Contrast with Agile Push

Agile sprint backlogs are often push systems in disguise. Product managers fill backlogs with features based on assumptions, and development teams pull items from the queue — but the upstream trigger is still the idea, not the hypothesis. Hypothesis pull requires a more fundamental discipline: before any development starts, articulate the assumption, define what data would validate or invalidate it, and build only enough to generate that data.

Sources

  • Ries, Eric (2011). The Lean Startup: How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses. Crown Publishing. ISBN: 978-0-307-88791-7.

    • Chapter 9 (Batch): pull system and WIP in product development

  • Ohno, Taiichi (1988). Toyota Production System: Beyond Large-Scale Production. Productivity Press. ISBN: 978-0-915299-14-0.

    • Original articulation of pull production and the kanban mechanism; foundational for all downstream pull-system thinking

  • Anderson, David J. (2010). Kanban: Successful Evolutionary Change for Your Technology Business. Blue Hole Press. ISBN: 978-0-984521-40-4.

    • Applies pull-system thinking to software development; distinguishes pull from push in knowledge-work contexts

  • Poppendieck, Mary and Tom Poppendieck (2003). Lean Software Development: An Agile Toolkit. Addison-Wesley. ISBN: 978-0-321-15078-3.

    • Chapter on eliminating waste and managing WIP in software; bridges Toyota manufacturing insight to software delivery

  • Goldratt, Eliyahu M. and Jeff Cox (1984). The Goal: A Process of Ongoing Improvement. North River Press. ISBN: 978-0-88427-061-9.

    • Theory of Constraints: identifying the system’s bottleneck determines where to pull; informs why WIP reduction accelerates throughput

Note

This content was drafted with assistance from AI tools for research, organization, and initial content generation. All final content has been reviewed, fact-checked, and edited by the author to ensure accuracy and alignment with the author’s intentions and perspective.

Graph View

Backlinks