06-reference/research

mammoth client field study feasibility

2026-05-21·research-brief·source: deep-research
macsanity-checkmammoth-growthanonymizationfield-studycontent

Anonymizing a Mammoth Growth Engagement Into a Serialized MAC Field Study

The question

Is there an existing Mammoth Growth client engagement (Progress, Nutrafol) that could anonymize into a serialized field study? The case-study version of Mechanic 5 is more credible than the synthetic version. (Source: project, High priority. Derivative of [[2026-05-11-practitioner-newsletter-acquisition-tactics.md]].)

What we already know (from the vault)

What the web says

Convergences and contradictions

Convergence: vault and web both confirm Nutrafol is already publicly named as an MG client with public outcome metrics (60% revenue / Unilever exit). The "is it possible to name the client" question is settled by MG's own published collateral.

Convergence: the technical pattern (MAC 3x6 matrix, PRO-303 override case study) is methodology the founder developed and applied across MG engagements. Methodology is independently developed; column names, table names, BRDs are client-confidential. The publishable artifact splits at exactly that line.

Contradiction: vault's MAC landing page build spec blocks ANY mention of "Mammoth", "Progress", "Nutrafol", "gold_opp_pipeline", "silver_fct_*" via grep audit. This is currently the safest posture, and it's stricter than NDA actually requires. The contradiction: the bet-architecture playbook treats the Progress engagement as the live MAC instrumentation that anonymized publishing is the goal of, but the landing-page spec blocks it. The spec is the right default for a public MAC page; the Sanity Check serialized field study is a different surface that could thread the needle.

Contradiction: MG itself publishes Nutrafol case-study collateral and brand attribution. RDCO publishing about working on a "well-known DTC hair-supplement brand with Magento + Snowflake + Segment that scaled through a 2022 acquisition" reads as Nutrafol to anyone who reads MG's collateral. Anonymization without explicit Nutrafol-permission is fragile — the persona-class reader will identify the brand within 60 seconds. The protection is methodology-focused, not identity-focused.

Synthesis for RDCO

The "anonymize a MG engagement into a serialized field study" question splits into three feasibility paths, ordered by risk:

Path 1: Fully synthetic field study, MG-inspired but not MG-derived (LOW RISK, HIGH PROVENANCE COST). Build a fictional company (e.g., "a Series B DTC supplements brand running Magento + Snowflake + dbt") with fictional table names and a fictional PRO-303-style override pattern. The pattern is real and load-bearing; the data is made up. Pro: zero NDA risk, no MG-permission asks, publishable today. Con: the credibility floor is lower (Mechanic 5 explicitly preferred case-study over synthetic exactly because synthetic reads as marketing). Footnote-disclosure required ("pattern observed across multiple analytics-engineering engagements; details synthesized").

Path 2: Anonymized field study built from MG patterns with explicit MG approval (MEDIUM RISK, HIGH CREDIBILITY). Same shape as Path 1 but the founder confirms with MG senior leadership (Dylan Cruise or Tom Clinton) that the methodology piece is fair game and the abstracted pattern doesn't reveal client-specific facts. Vault evidence suggests this conversation would land well: Dylan publicly framed the cc-wrapped agentic harness as the founder's "brainchild." That's permission-adjacent language. Pro: the founder can write with conviction, can use realistic numbers ("a 14-existing-tests-to-95-MAC-cells gap is typical in mid-market dbt projects" is a credible claim once a 30-second Dylan-conversation confirms anonymized-pattern publishing is fine). Con: requires one outreach action to Dylan/Tom, requires waiting for the response, requires founder discipline to honor the boundary they set.

Path 3: Explicit named-client case study with client permission (HIGH RISK, HIGHEST CREDIBILITY). Reach out directly to a Nutrafol or Progress contact, ask for explicit testimonial-style permission to write a published field study referencing them by name. Pro: highest credibility, opens the door to named-client testimonials. Con: the founder is mid-transition to phData; the optics of poking active MG client relationships for personal-brand content while leaving MG is awkward and could damage the standing return-offer Tom/Jeff explicitly extended. Almost certainly the wrong move in 2026; revisit in 2027 once the MG transition is past.

Recommended posture: Path 2. The founder has the social capital with Dylan and Tom to ask one direct question ("any concern if I write a Sanity Check field study on anonymized MAC patterns I built during the MG engagement, scrubbed of all client identifiers?"). The most likely answer is yes-with-the-obvious-guardrails. That answer materially upgrades every Sanity Check piece from Path 1 (synthetic) to Path 2 (methodology with real numbers from real engagements, anonymized).

Operational implication for Sanity Check serialization. Once Path 2 is unlocked, the candidate field-study arc looks like:

  1. Episode 1: The 14-to-95 gap. Anonymized MAC application showing the typical existing-tests-to-MAC-cells gap on a real dbt project. Hooks the persona on the diagnostic value.
  2. Episode 2: The PRO-303 pattern. Why conventional schema.yml testing misses dual-path overrides, with an anonymized PRO-303-style override example. Hooks on the bug-class-by-construction value.
  3. Episode 3: Severity tiers in production. How Stop/Pause/Go let all tests run without alert fatigue. Hooks on the operational-deployability angle.
  4. Episode 4: Cross-model reconciliation. Anonymized Aggregate-scope rules (gold↔gold reconciliation). Hooks on the audit-readiness value.
  5. Episode 5: The cc-wrapped harness pattern. Dylan's on-record characterization of the harness as transformative — this episode IS the MAC delivery-vehicle thesis, told as a real-world adoption story. (Requires explicit Dylan permission to quote.)

This is the case-study-version of Mechanic 5 that the practitioner-newsletter brief named as higher-credibility than the synthetic version. It's accessible at the cost of one outreach conversation.

Action recommendation for the founder: schedule the Dylan-or-Tom permission conversation within the first 30 days at phData (so the MG transition is cleanly past, but the relationships are still fresh). Frame as: "I'd like to write a Sanity Check field-study series on the MAC methodology I built across MG engagements, with full anonymization of client identifiers, code references, and BRD content. Want to confirm there's no concern from your side before I draft." Document the response in the vault. Path 2 unlocks.

Open follow-ups

Sources

Vault:

Web: