organvm — Public Process
← Back to dissertation overview

Chapter 4: Results

Precision Over Volume Chapter 4 of 7 @4444J99 March 04, 2026
3.8k words 15 min
mcdacareer-pipelinemathematical-proofscompetitive-analysisprecision-hiring

CHAPTER 4 | RESULTS

4.1 Introduction

This chapter presents the results of the three analytical methods described in Chapter 3: (1) formal mathematical proofs of optimality (Section 4.2), (2) systematic competitive analysis (Section 4.3), (3) empirical pipeline analysis with era-separated cohort comparison (Section 4.4), and an additional analysis of the Aristotelian rhetorical framework embedded in the system’s architecture (Section 4.5). Results are organized by research question to facilitate direct comparison with the hypotheses stated in Chapter 1.

4.2 Mathematical Proofs of Optimality

4.2.1 Theorem 1: WSM Boundedness and Consistency

Theorem 1. For any pipeline entry a with dimension scores s_i in {1, 2, …, 10} for i = 1, …, 9, and any weight configuration w = (w_1, …, w_9) satisfying w_i > 0 and Sum(w_i) = 1.0, the composite score V(a) = Sum(w_i * s_i) is bounded on [1.0, 10.0].

Proof. Since each s_i >= 1 and each w_i > 0:

V(a) = Sum(w_i * s_i) >= Sum(w_i * 1) = Sum(w_i) = 1.0

Since each s_i <= 10:

V(a) = Sum(w_i * s_i) <= Sum(w_i * 10) = 10 * Sum(w_i) = 10.0

Therefore 1.0 <= V(a) <= 10.0 for all entries and all valid weight configurations. QED.

Corollary 1.1. The 9.0 threshold corresponds to the 89th percentile of the feasible range: (9.0 - 1.0) / (10.0 - 1.0) = 8.0/9.0 = 0.889.

Corollary 1.2. The composite score is scale-independent: changing the scoring scale from [1, 10] to any [a, b] with a < b produces a linearly transformed composite V’ = a + (b-a) * (V-1)/9 that preserves the ordinal ranking of all entries.

4.2.2 Theorem 2: WSM Optimality Under MPI

Theorem 2 (Keeney & Raiffa, 1976). If the 9 scoring dimensions satisfy mutual preferential independence (MPI), then the Weighted Sum Model V(a) = Sum(w_i * s_i) is the unique value function (up to positive affine transformation) consistent with the decision-maker’s preferences.

Verification for the career application domain. As argued in Chapter 3 (Section 3.4.4), the 9 dimensions satisfy MPI because:

(a) Mission_alignment can be assessed independently of portal_friction. (b) Network_proximity can be assessed independently of deadline_feasibility. (c) Evidence_match can be assessed independently of financial_alignment.

More generally, each dimension measures a genuinely distinct aspect of the application opportunity. The preference for higher scores on any dimension does not depend on the levels of the remaining dimensions. This is verified by examining all 36 pairwise independence conditions (9 choose 2 = 36 pairs).

Implication. Under MPI, alternative aggregation methods (TOPSIS, ELECTRE, PROMETHEE) do not provide a better representation of the decision-maker’s preferences than WSM. WSM is not merely adequate but optimal for this domain.

4.2.3 Theorem 3: Cold Application Impossibility (Job Track)

Theorem 3. Under the WEIGHTS_JOB configuration, no pipeline entry with network_proximity = 1 (cold) can achieve a composite score >= 9.0, regardless of scores on all other dimensions.

Proof. The maximum composite score for a cold-network job-track entry is achieved when all non-network dimensions are at their maximum (10):

V_max_cold = Sum_{i != net}(w_i * 10) + w_net * 1

= (0.25 + 0.20 + 0.15 + 0.10 + 0.05 + 0.03 + 0.01 + 0.01) * 10 + 0.20 * 1

= 0.80 * 10 + 0.20

= 8.0 + 0.2

= 8.2

Since 8.2 < 9.0, it is impossible for a cold-network job-track entry to reach the qualification threshold. QED.

Corollary 3.1. The minimum network proximity score required for a job-track entry to potentially reach 9.0 (with all other dimensions at 10) is:

s_net_min = (9.0 - 8.0) / 0.20 = 1.0 / 0.20 = 5.0

Since acquaintance = 4 < 5.0 and warm = 7 > 5.0, the minimum relationship level for job-track qualification is warm (assuming perfect scores on all other dimensions).

Corollary 3.2. For typical (non-perfect) dimension scores, the minimum required network proximity is higher. If the average non-network dimension score is 8 rather than 10:

V_other = 0.80 * 8 = 6.4

s_net_min = (9.0 - 6.4) / 0.20 = 2.6 / 0.20 = 13.0

Since 13.0 > 10 (maximum possible), a cold-network job-track entry with average dimension scores of 8 cannot reach the threshold at any network level. This entry requires improvement on non-network dimensions before network cultivation becomes sufficient.

4.2.4 Theorem 4: Kelly Criterion for Application Strategy

Theorem 4. Under observed market parameters, cold applications (p = 0.02, b = 5) have a negative Kelly fraction f* = -0.176, indicating negative expected geometric growth rate. Network-supported applications (p = 0.15, b = 10) have a positive Kelly fraction f* = +0.065.

Proof. The Kelly criterion gives the optimal fraction of capital (effort) to wager on a binary bet:

f* = (b*p - q) / b

where p is win probability, q = 1-p, and b is the payoff ratio.

Cold applications:

The negative Kelly fraction means: do not make this bet. The expected geometric growth rate is negative; each cold application destroys expected effort capital over time.

Warm referral applications:

The positive Kelly fraction means: allocate 6.5% of effort capital to each warm referral application.

Strong referral applications:

A substantial 17.5% allocation is optimal, supporting the pipeline’s strategy of concentrating effort on high-probability opportunities.

Interpretation. The Kelly analysis reveals a sharp discontinuity in optimal strategy: below a conversion rate threshold of approximately p* = q/b = 0.98/5 = 0.196 (for b = 5), the Kelly fraction is negative and the optimal allocation is zero. Cold applications fall well below this threshold. The precision pipeline’s structural enforcement — making it mathematically impossible for cold-network job-track entries to qualify — is therefore not merely a heuristic preference but a Kelly-optimal constraint.

4.2.5 Theorem 5: Reservation Score Optimality

Theorem 5. Under McCall’s (1970) job search model, the optimal strategy is a threshold strategy: pursue opportunities with composite scores >= w-bar (the reservation score) and decline all others. The 9.0/10 threshold is a rational reservation score under the current market conditions.

Proof sketch. McCall’s reservation wage theorem establishes that the optimal policy for a sequential decision problem with i.i.d. offers is a threshold strategy. The proof follows from the contraction mapping theorem applied to the Bellman operator:

(Tv)(w) = max{w/(1-beta), c + beta * E[v(w’)]}

The operator T is a contraction on the space of bounded continuous functions with the sup norm. By the Banach fixed-point theorem, T has a unique fixed point V* = TV, and the associated policy is the unique optimal policy. Since V(w) is non-decreasing and the accept payoff w/(1-beta) is strictly increasing while the reject payoff c + betaE[V(w’)] is constant in w, there exists a unique threshold w-bar at which the two are equal. Above this threshold, accepting dominates; below, continuing dominates.

Mapping to the pipeline. The composite score V(a) functions as the “offer” in McCall’s model. The threshold 9.0 functions as the reservation wage. The comparative statics (Chapter 2, Section 2.6.2) justify the threshold level:

4.2.6 Theorem 6: Absorbing Markov Chain Properties

Theorem 6 (Kemeny & Snell, 1960). The pipeline state machine, modeled as an absorbing Markov chain with 8 transient states and 2 absorbing states, has the following properties:

(a) The fundamental matrix N = (I - Q)^{-1} exists and all entries are finite and non-negative. (b) The expected number of steps before absorption from any transient state is finite. (c) The absorption probability matrix B = NR satisfies Sum_j(B_{ij}) = 1 for all transient states i.

Proof. Properties (a)–(c) follow from the standard theory of absorbing Markov chains (Kemeny & Snell, 1960, Chapter 3). An absorbing Markov chain has the property that Q^n -> 0 as n -> infinity (the probability of remaining in transient states converges to zero). This implies that (I - Q) is non-singular and N = (I - Q)^{-1} = Sum_{k=0}^{infinity}(Q^k) converges. Property (c) follows from NR = (I-Q)^{-1}R and the row-stochastic property of P: since each row of P sums to 1, each row of B = NR must also sum to 1.

Interpretation. These properties guarantee that every pipeline entry eventually reaches a terminal state (outcome or withdrawal) — no entry is trapped in an infinite transient loop. This is a basic sanity property of the pipeline design: the state machine does not have absorbing non-terminal cycles.

4.3 Competitive Analysis Results

4.3.1 Survey Scope

The competitive analysis surveyed 60+ products across 5 categories (Chapter 2, Section 2.10.1): applicant tracking systems (employer-side), job search aggregators, resume optimization tools, AI-powered application assistants, and career strategy platforms. Additionally, 15 academic decision support prototypes documented in the MCDA and career management literature were evaluated.

4.3.2 Capability Matrix Results

Table 4.4: Competitive System Feature Gap Analysis

Capability Dimension Precision Pipeline Best Competitor Gap
1. Multi-criteria scoring Full (9-dim WSM) Partial (Teal: basic scoring) 0.5
2. Network proximity analysis Full (5-level, 6-signal) Partial (LinkedIn: mutual connections) 0.5
3. Time-decayed network signals Full (4-tier step function) None 1.0
4. Portfolio diversification Full (9 tracks) Partial (Huntr: multiple boards) 0.5
5. Reachability analysis Full (per-entry gap analysis) None 1.0
6. Bayesian outcome learning Full (70/30 blend) None 1.0
7. Absorbing Markov chain modeling Full (10-state FSM) Partial (some stage tracking) 0.5
8. Mode switching Full (3 modes) None 1.0
9. Block-based content composition Full (4-tier depth) Partial (Teal: content templates) 0.5
10. Identity position framework Full (5 positions) None 1.0
11. Relationship cultivation workflow Full (score-integrated) Partial (CRM tools: basic tracking) 0.5
12. Market intelligence integration Full (112 sources) Partial (LinkedIn: salary data) 0.5
Total 12.0/12 3.0/12 (Teal, best) 9.0

4.3.3 Gap Analysis

The three capability dimensions with zero implementation across all surveyed systems represent the precision pipeline’s most distinctive contributions:

Time-decayed network signals (Dimension 3). No existing system applies time decay to relationship quality scores. CRM tools (Salesforce, HubSpot) track contact recency but do not integrate it into a scoring framework for career decision-making.

Reachability analysis (Dimension 5). No existing system computes per-entry sensitivity analysis showing the minimum relationship level needed to cross a qualification threshold. This capability is unique to the precision pipeline.

Bayesian outcome learning (Dimension 6). No existing system adjusts scoring weights based on observed conversion outcomes. All surveyed systems treat configuration as static, requiring manual recalibration.

Identity position framework (Dimension 10). While some career coaching services recommend tailoring materials to specific audiences, no existing system formalizes this as a structured framework with pre-composed credential mappings, block selections, and resume templates for each position.

Mode switching (Dimension 8). No existing system implements market-condition-aware governance that adjusts thresholds and constraints based on the applicant’s circumstances and market conditions.

4.3.4 Hypothesis H2 Assessment

H2a: No existing system achieves more than 50% (6/12) of the capability dimensions. Confirmed. The best competitor (Teal) achieves 3/12 (25%), well below the 6/12 threshold. No other surveyed system exceeds 3/12.

H2b: At least three dimensions are entirely unique to the precision pipeline. Confirmed. Five dimensions (3, 5, 6, 8, 10) have zero implementation across all surveyed systems. This exceeds the hypothesized minimum of three.

4.4 Empirical Pipeline Analysis

4.4.1 Score Distribution Analysis

The pipeline’s 1,000+ entries exhibit the following score distribution characteristics:

Volume-era entries (pre-March 4, 2026):

Precision-era entries (post-March 4, 2026):

4.4.2 Network Proximity Impact

Table 4.6: Maximum Achievable Score by Network Level and Track

Network Level Score Max Composite (Job, others=10) Max Composite (Creative, others=10)
Cold 1 8.2 8.92
Acquaintance 4 8.8 9.28
Warm 7 9.4 9.64
Strong 9 9.8 9.88
Internal 10 10.0 10.0

This table demonstrates the structural role of network proximity in the qualification decision:

4.4.3 Research Pool Analysis

The research pool contains approximately 948 auto-sourced entries, primarily job-track entries identified through ATS API scanning (source_jobs.py). At the 9.0 threshold, zero of these entries auto-qualify. The primary reason is the network_proximity dimension: auto-sourced entries have network_proximity = 1 (cold) by default, as no relationship exists at the time of sourcing.

This result validates the precision pipeline’s design: the research pool serves as a discovery mechanism, not a qualification mechanism. Entries in the research pool require relationship cultivation before they can qualify — exactly the workflow that cultivate.py and the reachability analysis are designed to support.

4.4.4 Staged Entry Triage

The 54 entries that reached staged status under volume-era rules were triaged using the v2 scoring system:

The triage revealed that the majority of staged entries had been advanced under volume-era rules without adequate network cultivation, resulting in scores below the precision-era threshold. This empirical finding validates the v1-to-v2 transition: the volume strategy was advancing entries that the precision system correctly identifies as under-qualified.

4.4.5 Pipeline Velocity Comparison

Pipeline velocity (time in each status) differs between eras:

Volume era:

Precision era:

The precision era produces a longer pipeline but with higher-quality submissions at each stage. The extended time in early stages (research, qualified) is invested in the relationship cultivation that the Kelly criterion analysis identifies as the prerequisite for positive expected value.

4.5 The Aristotelian Framework: Ethos, Logos, Pathos

4.5.1 Rhetoric Embedded in System Architecture

The precision pipeline’s architecture instantiates the Aristotelian rhetorical triad at every level:

Table 4.7: Aristotelian Dimensions in System Architecture

Rhetorical Mode System Component Implementation
Ethos (Credibility) Identity positions 5 canonical credential frameworks mapping applicant experience to reviewer expectations
  Block library (identity/) Pre-composed credibility statements at 4 depth tiers
  Scoring dimensions track_record_fit and evidence_match capture ethos signals
  Portfolio evidence System metrics (“103 repos, 2,349 tests, 810K words”) as existence proof
Logos (Logic) Scoring engine 9-dimension WSM with formal mathematical justification
  Storefront content Metrics-first, scannable, quantified evidence
  Reachability analysis Logical framework for cultivation priority
  Conversion analytics Data-driven strategy adjustment
Pathos (Emotion) Cathedral content Deep narrative, mission resonance, emotional connection
  Cultivation workflow Reciprocity-based relationship building (Cialdini, 2006)
  Block library (framings/) Identity-position narratives connecting applicant to mission
  Follow-up protocol Sustained engagement demonstrating genuine interest

4.5.2 The Storefront/Cathedral Architecture as Dual-Channel Rhetoric

The Storefront/Cathedral content architecture maps directly to two distinct persuasion channels:

Storefront (Logos channel). Optimized for the 6-second screening window (TheLadders, 2018). Content design principles:

The Storefront layer maximizes Shannon’s signal-to-noise ratio for the initial screening channel, where the bandwidth is severely constrained (6–7.4 seconds of reviewer attention).

Cathedral (Pathos channel). Optimized for deep evaluation when the application advances past the initial screen. Content design principles:

The Cathedral layer triggers narrative transportation (Green & Brock, 2000), oxytocin release (Zak, 2015), and neural coupling (Hasson et al., 2012), creating the emotional engagement that increases favorable evaluation during the deeper review stages.

4.5.3 Identity Positions as Audience-Specific Rhetoric

The five identity positions (Independent Engineer, Systems Artist, Educator, Creative Technologist, Community Practitioner) function as audience-specific rhetorical frameworks. Each position provides:

  1. A credibility template (ethos): which credentials, experiences, and portfolio elements to foreground
  2. A narrative template (pathos): which personal story arc resonates with the target audience
  3. An evidence template (logos): which metrics, case studies, and work samples demonstrate capability

The identity position system ensures that the same underlying qualifications are framed differently for different audiences — a direct application of Aristotle’s principle that effective rhetoric requires adapting the message to the audience.

4.5.4 Cialdini’s Principles in the Cultivation Workflow

The relationship cultivation workflow operationalizes multiple Cialdini influence principles:

Reciprocity: The “give before ask” protocol (research the contact’s work, share relevant insights, contribute to their projects before requesting a referral) creates reciprocity debt that facilitates the eventual referral request.

Liking: The research phase identifies common ground (shared university, shared professional interests, shared values) that increases perceived similarity and liking.

Social proof: Mutual connections and organizational density provide social proof that the applicant is embedded in the relevant professional community.

Unity: The identity position system frames the applicant as a member of the same professional community (“we” language), leveraging the unity principle.

Scarcity: The block composition system highlights unique capabilities (“the only system that unifies MCDA with network theory and portfolio optimization”) that invoke the scarcity principle.

4.6 Limitations of Results

4.6.1 Mathematical Proof Limitations

The mathematical proofs establish optimality under specific conditions (MPI for WSM, i.i.d. offers for McCall, binary outcomes for Kelly). Real-world conditions may deviate from these assumptions:

These deviations are acknowledged as limitations of the theoretical framework, not as invalidations. The proofs establish that the pipeline’s design is justified under well-defined conditions; they do not claim that these conditions hold perfectly in all cases.

4.6.2 Competitive Analysis Limitations

The competitive analysis is limited to products available for evaluation as of March 2026. The career technology landscape is rapidly evolving, and new products may enter the market that address some of the identified gaps. The analysis should be updated periodically to maintain currency.

4.6.3 Empirical Limitations

Limited post-pivot outcome data constrains the empirical analysis. The volume-era vs. precision-era comparison provides evidence of different score distributions and pipeline velocities, but outcome-based conversion comparisons require more accumulated data. The Bayesian outcome learning system will provide ongoing empirical validation as outcomes are recorded.

4.7 Summary of Key Findings

Table 4.8: Summary of Key Findings by Research Question

Research Question Hypothesis Finding Status
RQ1 (Logos): Is the scoring engine optimal? H1a: Composite bounded [1.0, 10.0] Proven (Theorem 1) Confirmed
  H1b: Cold apps have negative Kelly fractions f* = -0.176 (Theorem 4) Confirmed
  H1c: Step-function consistent with exponential Bounded error within tier width Confirmed
RQ2 (Ethos): Does the pipeline offer unique capabilities? H2a: No competitor achieves > 6/12 Best competitor: 3/12 Confirmed
  H2b: >= 3 unique dimensions 5 unique dimensions identified Confirmed
RQ3 (Pathos): Does precision reduce human cost? H3a: Fewer rejections per success Proportional to cold-to-warm ratio (est. 4–10x) Supported
  H3b: Precision workflow produces higher scores Score distribution shift observed Supported
RQ4 (Synthesis): Is v2 the “gold standard”? H4: Under joint conditions, v2 meets gold-standard criteria All conditions satisfied Confirmed

The most practically significant finding is the Kelly criterion result (Theorem 4): cold applications have mathematically negative expected value. This means that the optimal strategy for cold applications is to not submit them at all — a finding that challenges the dominant “apply to everything” advice in the career counseling industry.

Related Repositories