Evidence-Based Validation: The Minimum Proof Before Building

Key Takeaways:

• The Problem: Most studios either over-build before they have evidence or skip validation entirely, burning capital on ideas that should have been killed at $50K, not $500K.

• The 9point8 View: Validation is a studio's core competitive advantage. The goal is better shots on goal, not more shots on goal. A disciplined venture studio validation process converts abstract theses into quantifiable evidence before committing build resources.

• The Outcome: A four-step playbook for building an evidence-based validation system: philosophy, evidence thresholds, stage-gates, and industrialized customer discovery.

The second you launch anything, you have created an anchor point. A brand. A URL. Customer expectations. Technical debt. Emotional attachment. Once anchored, the cost of killing the idea escalates by orders of magnitude. That is why the studios with the best portfolio outcomes are not the ones that build the fastest. They are the ones that validate the most ruthlessly before building anything.

The venture studio validation process is what separates studios from accelerators, incubators, and traditional VC. Accelerators compress the growth timeline for ideas that already exist. Studios own the creation process, which means they own validation. The studio decides what gets built, with what evidence, and under what constraints. When designed well, this process becomes the single most capital-efficient mechanism in the venture creation stack.

Here is how to build one.

Before designing your validation system, three things must be in place:

• A defined thesis. Validation tests a hypothesis. Without a thesis, you are running experiments without a control. Your thesis defines the problem space, target market, and the type of ventures you will build. Validation confirms (or kills) specific ventures within that frame.
• A capital allocation framework. You need a defined budget per venture concept for the validation stage. This forces discipline. Idealab's $50,000 limit per concept (documented by Bill Gross and Howard Morgan across multiple Venture Studio Forum presentations) is the canonical example: enough to test technical viability, not enough to fall in love with a bad idea.
• A defined kill authority. Someone must have the explicit power to terminate concepts. If killing a venture requires consensus from a committee of twelve, nothing will ever die. Define who can pull the trigger, and under what conditions, before your first concept enters the pipeline.

Your thesis determines your validation approach, and your choice shapes everything from team composition to budget to timeline. Three models from operating studios illustrate the spectrum.

The "Fail Fast, Fail Cheap" model. Bill Gross and Howard Morgan's Idealab pioneered this approach: initiate projects with a strict $50,000 cap. Test technical viability in weeks, not months. If unit economics do not signal breakout potential, kill the project before it creates operational drag. This model optimizes for equity efficiency through extreme capital constraints and works best for studios with a high volume of software-centric concepts.

The "Foresight" model. Creative Dock invests heavily in upfront scenario modeling. Their team of 500 specialists, including 25 to 30 data scientists (per Creative Dock's published company data), builds "future-back" scenarios to de-risk ventures before a line of code is written. The result: a reported 78 to 80% success rate for ventures that reach market. This model demands more upfront capital and time but dramatically reduces failure rates. It fits studios building in complex, regulated, or capital-intensive sectors.

The "Scientific Teaming" model. Carbon13 vets approximately 750 candidates per cohort, selects roughly 80, and runs a 10-week teaming sprint where scientists and engineers iterate around a climate problem before formalizing co-founding agreements. The validation is the team itself: can these people work together, and can their combined expertise produce a viable approach? This model works for deep-tech or science-based studios where team composition is the primary variable.

Model	Budget per Concept	Timeline	Kill Signal	Best Fit
Fail Fast (Idealab)	$50K cap	Weeks	Unit economics fail	High-volume software
Foresight (Creative Dock)	Higher upfront load	Months	Scenario model fails	Regulated, complex sectors
Scientific Teaming (Carbon13)	Team formation cost	10 weeks	Team cannot coalesce	Deep-tech, science-based

Common mistake: Picking the "cool" model instead of the one that matches your thesis. A climate deep-tech studio running $50K sprints will produce surface-level prototypes that miss the science. A B2B SaaS studio running 10-week teaming sprints will burn time that should be spent talking to customers.

Before a concept earns a single dollar of build capital, define what counts as evidence. Opinions, excitement, and slide decks are not evidence. Customer conversations, signed LOIs, and pre-orders are.

The 100-conversation threshold is a benchmark adopted by multiple operating studios as a qualitative minimum: conduct at least 100 structured customer discovery conversations with potential customers who match your target profile, recorded and analyzed for patterns. Not 10 casual chats. Not a survey. One hundred recorded, structured interactions. For a two-person validation team, 100 conversations takes roughly four to six weeks of focused outreach. It is a significant but bounded commitment.

What passes the evidence bar:

• Letters of Intent (LOIs): A non-binding written commitment from a potential customer to purchase the product when it exists. Include pricing terms and volume expectations.
• Pre-orders or deposits: Cash on the table. The strongest signal available at the validation stage.
• Signed pilot agreements: Particularly valuable for B2B. Whatnot's 6-Step Method (presented at the Venture Studio Forum) uses a signed pilot or partner contract as the explicit pass gate for market entry.
• Technical feasibility proof: For deep-tech, a working proof of concept that demonstrates the core mechanism functions. Not a prototype. Not a demo. A functioning mechanism. One signal experienced studio operators watch closely during validation: behavioral transparency from founders. Across VSF practitioner discussions, willingness to be voluntarily transparent with data, setbacks, and discovery findings is consistently cited as a leading indicator of coachability and responsiveness. Founders who obscure results or avoid hard questions are telling you something about how they will operate post-funding.

The real job during validation is to identify what kills the idea and evolve it. Business model, product, market: everything is on the table. If nothing changes through validation, you were not validating. You were confirming.

Common mistake: Treating validation as a checkbox. Running 100 conversations and then building exactly what you planned before you started. The evidence must reshape the venture concept, or you are ignoring your own data.

Every gate is a capital allocation decision, not a developmental milestone. The studio must maintain "drop-and-go" capability at every gate. Each stage is funded independently. Failure to clear the gate results in termination, not a second chance.

A proven four-stage structure:

• Stage 0: Gap Analysis and Talent Sourcing. Identify the market gap through thesis-driven research (not random idea generation). Some studios source through IP licensing with patent scoring systems drawing on decades of historical licensing performance, shifting intake from "founder brings idea" to "studio identifies research with commercial viability." This stage is research and recruitment. It costs almost nothing.
• Stage 1: Teaming and Iteration. The core team forms and iterates around the problem for a defined period. Scientists, engineers, and operators test whether they can work together and whether their combined approach produces viable solutions. Gate criteria: formalized co-founding agreement or kill.
• Stage 2: Validation and Evidence. Execute the customer discovery process. Run the 100 conversations. Build the minimum technical artifact needed to test the core value proposition. Gate criteria: evidence threshold met (LOIs, pre-orders, signed pilots) or kill.
• Stage 3: Investment and Scale. Transition to formal incorporation and the first external capital raise. The studio shifts from primary operator to strategic advisor and platform provider. Gate criteria: we recommend setting financial viability gates at target gross margins above 70% and CAC payback under 12 months. These thresholds should be calibrated to your sector and venture type, but the principle holds: define quantitative bars, not subjective readiness. Each gate decision touches all four studio customers: the Studio Itself (capital preservation), Entrepreneurs/Founders (clear expectations), Follow-on Capital (de-risked pipeline), and LPs/Stakeholders (portfolio discipline). A weak gate does not just let a bad venture through. It erodes trust with every constituency the studio serves.

The kill switch at every gate is a feature, not a failure. It is the studio's primary capital preservation mechanism. Studios that cannot kill ventures at each gate end up with zombie projects: ventures that consume resources and human energy without clearing meaningful milestones.

Common mistake: Making gates advisory instead of binding. If the investment committee can override a failed gate because "the team just needs more time," you do not have gates. You have suggestions.

The highest-performing studios do not treat customer discovery as a per-project activity. They treat it as a permanent intelligence layer. One European studio has recorded and transcribed over 14,000 customer discovery calls (as shared in a Venture Studio Forum member presentation), linked them to ideal customer profiles and venture concepts, and loaded the entire corpus into AI for analysis. When they evaluate a new idea, they can immediately identify gaps, resistance patterns, and best-fit customer profiles from existing data. They also record internal team conversations after calls to capture context that formal notes miss.

This intelligence layer compounds over time. A studio in its first year will not have 14,000 calls to draw from, and that is expected. The discipline is starting the system on day one so that by year three, your fifth venture benefits from discovery data gathered during your first.

This is the difference between artisanal validation and industrial validation. Artisanal means every venture starts from zero: new interviews, new market research, new customer relationships. Industrial means every venture builds on the intelligence from every previous one.

To industrialize discovery:

• Record and transcribe every customer conversation. No exceptions. Unrecorded conversations are lost data.
• Link conversations to customer profiles and venture concepts. Tag by industry, problem type, customer segment, and objection pattern.
• Build a searchable, queryable corpus. When the next concept enters the pipeline, the first question should be: "What do we already know about this problem from previous discovery?"
• Use scoring systems for repeatable evaluation. One studio built a patent scoring system drawing on 30 to 40 years of historical licensing performance, allowing them to evaluate IP-based concepts against a quantitative benchmark rather than gut feel. The goal: better shots on goal, not more shots on goal. Every dollar spent on customer discovery should compound across the portfolio, not evaporate when a single venture is killed.

A studio with a functioning evidence-based validation system shows specific signatures:

• Post-gate success rates above 70%. Creative Dock's reported 78 to 80% success rate for ventures reaching market after clearing strategic foresight gates is the benchmark. If your post-gate failure rate exceeds 30%, your gates are not selective enough.
• Kill decisions happen in weeks, not quarters. The $50K cap is not just a budget constraint. It is a forcing function for speed. Slow kills are expensive kills.
• Capital gets reallocated, not hoarded. The purpose of killing early is not austerity. It is reallocation. Every dollar saved from a killed venture is available for the next concept that clears the evidence threshold.
• Discovery data compounds across the portfolio. Each validation cycle makes the next one faster and cheaper because the intelligence layer grows. Studios that build this compounding advantage create a structural moat that cannot be replicated by simply running more cohorts. The validation process is where the venture studio model earns its performance premium. Not in the idea. Not in the brand. In the discipline of requiring minimum proof before building.

---

About 9point8 Collective:

9point8 Collective is a specialist consultancy that designs, builds, and launches venture studios. We do not build startups; we engineer the operating systems, governance, and talent pipelines that allow universities, corporations, investors, and regional organizations to build portfolios of startups at scale.