The Crosswalk: Mapping the 8421 Decision Framework
How do you know your evaluation framework isn't a sophisticated vibe check? You validate it against every serious published source you can find.
When considering how to build a quantitative venture studio, staring into the chasm of endless possibilities available to venture creators was a daunting prospect. How could we possibly go about designing evaluation criteria in such a chaotic and unstructured environment? We started upside down, by inductively analyzing the structural characteristics of nascent and growing ventures. We worked to establish a set of independent indicators which could constitute a handful of relatable dimensions, stripping away statistically redundant indicators and optimizing for dimensional orthogonality. We ended up with 33 proprietary indicators that mutually exclude each other in their constitution of 6 domains: Demand, Feasibility, Economics, Go-to-Market, Defensibility, Leadership
But, it's not enough for scoring to be rigorous or methodology sophisticated. What matters is whether the categories themselves capture the uncertainties that actually matter. So, we set about validating our dimensions against every serious published source we could find.
The Eight Sources
The crosswalk validates against a deliberately heterogeneous set of references, and the heterogeneity is the point. These eight sources are not organized around a common structure. They are organized around different structures entirely.
Sequoia Capital's business plan guide defines ten sections of a pitch. Y Combinator's application is an operational checklist: specific questions about product, users, revenue, and founder capability. Thiel's seven questions from Zero to One are exactly that: questions (engineering, timing, monopoly, people, distribution, durability, secret). Kaplan and Stromberg (2004, Journal of Finance) derived three risk categories (internal, external, execution) from 67 VC investment memoranda. Gompers, Gornall, Kaplan, and Strebulaev (2020, Journal of Financial Economics) surveyed 885 VCs to identify the factors that actually drive investment decisions. IDEO's DFV model uses three design lenses: desirability, feasibility, viability. SRI International's NABC defines four evaluation criteria (Need, Approach, Benefits/Cost, Competition). Sahlman (1997, Harvard Business Review) frames venture evaluation around four interdependent factors: people, opportunity, context, risk/reward.
Pitch sections. Application questions. Risk categories. Decision factors. Design lenses. Evaluation criteria. Interdependent factors. Eight sources, eight different organizing logics, developed independently by different communities for different purposes. All of them ostensibly focused on imposing some structure onto the chaos of venture creation.
The crosswalk maps all of them onto a common structure: six dimensions of uncertainty. When sources organized around fundamentally different logics converge on the same underlying categories, that convergence is more meaningful than agreement among sources that started from the same premises.
| Dimension | Sequoia | YC | K&S 2004 | GKGS 2020 | IDEO DFV | Thiel | NABC | Sahlman |
|---|---|---|---|---|---|---|---|---|
| D1 Demand | Full | Full | Full | Full | Full | Full | Full | Full |
| D2 Feasibility | Full | Partial | Full | Full | Full | Full | Full | Partial |
| D3 Go-to-Market | Absent | Full | Partial | Absent | Absent | Full | Partial | Partial |
| D4 Economics | Full | Full | Full | Full | Full | Partial | Full | Full |
| D5 Defensibility | Partial | Partial | Full | Partial | Absent | Full | Partial | Absent |
| D6 Leadership | Full | Full | Full | Full | Absent | Full | Absent | Full |
Where Everything Agrees
Demand. Every single source assesses it. Full coverage across all eight. If eight independent models, studies, and practitioner guides all treat demand as a first-class evaluation category, that is because demand is a load-bearing uncertainty in venture creation. The question of whether a real problem exists, whether someone will pay to solve it, and whether the market is large enough to justify the investment is so fundamental that no serious evaluation can omit it.
Feasibility receives similarly broad coverage: full treatment in six of eight sources, partial in two (YC and Sahlman). Economics is fully covered in seven of eight, with Thiel as the partial outlier. These dimensions are not novel. They are validated.
The 8421 framework's value with these dimensions is not in discovering categories that others missed. It is in ensuring that each indicator within the dimension has exactly one home, that coverage is complete, and that scoring is structurally decomposed rather than holistically assessed. Six demand indicators, each scored independently through separate evaluation passes, preventing the halo effects that contaminate holistic assessments. The same problem, dissected more precisely.
Where the Framework Adds Something New
D3, Go-to-Market, as a standalone dimension is the framework's strongest novelty claim.
Most of these sources subsume distribution into demand or feasibility. Sequoia does not treat it independently. Gompers et al. do not treat it independently. IDEO DFV does not treat it independently. The implicit assumption is that if you have identified demand and can build the product, the go-to-market question will resolve itself.
This assumption is empirically wrong. The distinction between "people want this" and "we can reach and sell to those people" is structural. A venture can face genuine demand, possess a feasible product, and still fail because the cost of acquiring customers exceeds the lifetime value they generate. Or because the sales cycle is too long for the venture's capital structure. Or because the distribution channels required to reach the buyer are controlled by incumbents.
Only Thiel and YC treat distribution as independent from demand or product. Thiel asks explicitly: "Do you have a way to not just create but deliver your product?" YC asks about growth metrics and distribution channels as standalone evaluation criteria. The 8421 framework elevates this to a full dimension with five dedicated indicators, because ventures evaluated at the idea stage face distribution risk that is structurally different from demand risk and must be assessed independently.
D6, Leadership, contains the most distinctive structural inversion. Most sources that assess leadership evaluate the quality of the existing team. Sequoia, YC, Kaplan and Stromberg, Gompers et al., Thiel, and Sahlman all include some form of team quality assessment.
The 8421 framework does not assess team quality. It assesses the difficulty of the leadership problem.
At the point of initial evaluation, the venture has no CEO. There is no team to assess. The relevant question is not "is this team good enough?" but "how hard will this company be to lead?" What domain expertise is required? What regulatory complexity must be navigated? What management challenges are inherent in the market structure? The answer to these questions determines the leadership capability profile that will be required at spin-out, without presupposing who will fill it.
Two of the eight sources (IDEO DFV and NABC) omit leadership entirely. The remaining six assess team quality. The 8421 framework assesses leadership difficulty. This is a structural difference with direct operational implications for how operators are eventually matched to ventures.
MECE Coverage
The crosswalk reveals something about the framework's architecture that is difficult to achieve by intuition and important to verify through validation: the 33 indicators are mutually exclusive and collectively exhaustive.
Mutually exclusive means that each uncertainty has exactly one home. Willingness to Pay belongs to D1 (Demand), not to D4 (Economics), even though both relate to revenue. Customer Acquisition Cost belongs to D3 (Go-to-Market), not to D4, even though it directly affects margin. The placement decisions follow a structural logic: each indicator lives in the dimension that governs the type of uncertainty it resolves, not the financial output it feeds. This structural placement is what allows Resolution Value to rank which uncertainties are worth resolving next.
Collectively exhaustive means that no meaningful venture uncertainty lacks a home. The crosswalk is the evidence for this claim. If any published source assesses a category that the 8421 framework does not cover, that gap would indicate a failure of exhaustiveness. No such gap exists in the current crosswalk. Every category assessed by any of the eight sources maps to at least one dimension in the 8421 framework.
This is a structural requirement, not a feature. Multi-attribute evaluation depends on preferential independence between dimensions, meaning that the score on one dimension can be assessed without reference to scores on other dimensions. If dimensions overlap, scoring one contaminates another. If gaps exist, uncertainties go unmanaged. The MECE property ensures that the decomposition is clean enough for the scoring methodology to function as designed.
The Structural Finding
The convergence table reveals a pattern worth stating explicitly. Where every source agrees (demand, feasibility, economics), the 8421 framework agrees too. Where sources diverge or leave gaps (go-to-market, defensibility, leadership), the 8421 framework has a specific structural position supported by the inductive derivation.
The crosswalk does not prove the framework is correct. We should be clear, that is not the assertion here. What it does is establish that the framework's categories are not arbitrary, that its coverage is validated against four decades of independent thinking, and that its distinctive contributions (GTM independence, leadership difficulty inversion, unique indicator decomposition) occupy gaps that the existing literature has identified but not addressed.
The framework's value is not in discovering new categories of venture uncertainty. It is in establishing structured, validated coverage where each categorized uncertainty maps to one home in the framework, with no gaps and no overlaps.