Theory of Constraints + Socratic-interview Claude skill — bottleneck identification framework
Why this is in the vault
The X post promotes a free Claude skill that does a 3-question Socratic interview based on Goldratt’s Theory of Constraints to identify the single biggest revenue bottleneck. Founder shared 2026-04-30 17:33 ET — same day RDCO’s bet-architecture playbook (2026-04-30-rdco-bet-architecture-playbook) was crystallized.
Filing because:
- The framework itself (Theory of Constraints) is canonical operator-discipline content (Bezos + MrBeast both require senior reports to read The Goal)
- The vocabulary is concrete and stealable for Sanity Check editorial use
- It’s structural confirmation that the bet-architecture playbook RDCO built today IS the right shape — we have a richer version, but the X post’s tool addresses the same operator pain point at smaller scope
The framework (3-question Socratic interview)
1. The Chain Question
Map your business as a chain of links: e.g., traffic → leads → sales → delivery. Describe what each link looks like in your specific business.
2. The Pile-Up Question
Hunt for the slowest link — the one that, if it 2x’d tomorrow, would 2x the entire business. Everything before it piles up waiting; everything after it starves. That’s your real bottleneck.
3. The Lock-In Test
Pressure-test your answer with real numbers (revenue, hours, conversion rates) before acting. So you don’t fix the wrong thing.
Output: one verdict (the single bottleneck blocking growth) + a customized playbook for breaking it.
Cadence: run quarterly. After you break this bottleneck, a new one surfaces. That’s the next mission.
Mapping against Ray Data Co — RDCO’s playbook is the richer version
The X-post tool addresses one bet at a time using Theory of Constraints. RDCO’s 2026-04-30-rdco-bet-architecture-playbook is the more sophisticated framework:
| Dimension | X-post tool | RDCO playbook |
|---|---|---|
| Scope | Single business / single chain | Portfolio of bets, each with its own chain |
| Layer count | 1 (the chain) | 4 (targeting / instrumentation / tools / feedback) — chain decomposed by function |
| Recursion | None | P&L meta-layer over sub-process targeting systems |
| Bottleneck identification | Single slowest link in chain | Load-bearing gap per layer + cross-bet modular-components mapping |
| Cross-bet payoff awareness | None | Built in (Squarely + MAC + SC + RDCO ops share modular components) |
| Lock-in test | Pressure-test with numbers | P&L meta-layer veto (vertical-farming worked example: yield gain via 2x water gets vetoed by utility-bill economics) |
| Cadence | Quarterly | Weekly review (planned) + on-demand |
The X post is correct that the bottleneck-identification operating system matters. RDCO’s version is structurally richer — but the underlying truth is the same.
Vocabulary worth stealing for SC editorial
The X post’s metaphors are concrete in a way our abstract “load-bearing gap” language isn’t. For SC editorial use:
- “The chain” — links of a business in sequence
- “The slowest link” — the binding constraint
- “Pile-up vs starvation” — what happens upstream + downstream of the bottleneck (visceral)
- “2x the link → 2x the business” — concrete sizing test
- “Lock-in test” — pressure-test with real numbers before committing
- “After you break it, a new one appears” — the never-ending bottleneck game
Combined with Bush #7 (“identify single biggest bottleneck → full attention → remove → next”) and the RDCO bet-architecture playbook, this gives a strong vocabulary stack for the operating-discipline content arc.
NOT a Sanity Check derivative pitch
Per ~/.claude/projects/-Users-ray/memory/feedback_no_derivative_sanity_check_pieces.md — don’t pitch an SC piece that just restates Theory of Constraints. The non-derivative re-frame, if any:
“Theory of Constraints assumes one business; RDCO’s portfolio of small bets generalizes the framework to a multi-bet structure where modular capabilities reduce the cost of fixing each bet’s bottleneck. The slowest link in Squarely (Amazon ads sensors) is the same SHAPE as the slowest link in Sanity Check (engagement dashboard). Build the dashboard primitive once, apply to both.”
That’s a real angle. Could stack into the existing 3 SC research-brief candidates if founder green-lights.
Goldratt’s The Goal — bookshelf candidate
The book is canonical operator-discipline content:
- Jeff Bezos: requires all senior Amazon execs to read it
- MrBeast: had his first 250 employees read it
- ISBN: 9780884271956 (3rd ed; multiple editions exist)
- Originally published 1984, still the canonical TOC text
- Format: business novel (fictional manufacturing-plant manager applies Theory of Constraints) — readable, not academic
Decision for founder: add to bookshelf? Founder may already own it. If so, OCR via the Wheeler-pipeline (VitalSource → screencap+tesseract). If not, ~$15 Kindle / ~$20 paperback — a clean Link-test purchase candidate (assuming Kindle DRM, the OCR pipeline is proven).
If we add it, the bookshelf gains canonical TOC content, which directly grounds the bet-architecture playbook’s bottleneck-identification process. Strong cross-bet payoff because TOC applies to every bet.
Author identified — Ole Lehmann (@olelehmann1337)
Founder shared the GitHub link 2026-04-30 17:33 ET: https://github.com/olelehmann1337/bottleneck-skill
Author is Ole Lehmann — operator-discipline X poster working in the Theory of Constraints / coaching space. Worth promoting to tracked-author / X-follow-forward candidate.
What’s in the actual skill (worth knowing)
Fetched the SKILL.md. 11 specific patterns Lehmann figured out beyond the 3-question public framing:
- Premise-stating opening — anchors with “the goal of every business is to make money, now and in the future” before asking anything. Otherwise the conversation drifts.
- Business-type detection in Q1 — hardcoded mental model of where bottlenecks typically live for SaaS / e-commerce / agency / services / content / course / marketplace / manufacturer / brick-and-mortar. Each gets a typical bottleneck location. Hypothesis, not verdict.
- High-risk-step probe — targeted follow-up about the step where this business type’s bottleneck most commonly hides (e.g., SaaS → “what does signup-to-first-value look like in days?”)
- Missing-step detection — “people often describe their chain in 4-6 steps but skip the bottleneck step. If they say ‘we run ads, leads come in, we close them, we deliver,’ ask what happens between ‘leads come in’ and ‘we close them.’ That gap is where pile-ups hide.” Sharp move.
- Refuse lists in Q2 — “Theory of Constraints only works if we commit to one. Naming two means we fix neither.” Stay on it until they pick one.
- Live-arithmetic cross-check — uses numbers the user gave in Q1 to push back on misidentified bottlenecks. “You said 200 leads/mo at 2% conversion. Doubling leads only takes you from 4 customers to 8. The bigger leak looks like conversion.”
- False-bottleneck patterns list — explicit symptoms-disguised-as-constraints to watch for:
- “We need more leads” → usually conversion or offer
- “We need to hire” → usually founder hasn’t built a system someone else can run
- “We need better tools” → almost never; comfort move
- “We need more time” → usually priority confusion or decision-queuing on the founder
- “Marketing is the bottleneck” → often retention, churn, or LTV
- “Our team is too small” → often the work flowing in is unfocused, not undersized
- “We need more capital” → almost never; throughput rarely capital-constrained at small scale
- Q3 forced specificity — “If that step had infinite capacity starting tomorrow, would monthly revenue move within 90 days? By how much?” Symptom vs real-bottleneck filter.
- Diagnosis card output format — single screenshot-friendly block with 5 focusing steps + 7-day experiment.
- 5 focusing steps language — Identify / Exploit / Subordinate / Elevate / Repeat (Goldratt-canon).
- 7-day falsifiable experiment — “If we got this right, doing X for 7 days should move metric Y by at least Z%. If it doesn’t, the bottleneck is somewhere else.”
Adaptation plan for RDCO
Per founder framing 2026-04-30 17:33: “Looks more like something we would adapt for ourselves than outright lift and shift.”
Keep verbatim from Lehmann
- 3-question Socratic shape (chain → pile-up → lock-in)
- 5 focusing steps language (Goldratt-canon)
- “Refuse lists” rule
- Live-arithmetic cross-check discipline
- Lock-in test’s forced specificity
- False-bottleneck patterns list (the 7 symptom patterns above)
- Diagnosis card output format
RDCO-fy
- Premise opening: replace “goal of every business is to make money” with the four-layer + recursive P&L framework. The TOC question becomes “which layer is binding this bet’s targeting system most?”
- Business-type detection: swap for bet-type detection (Squarely / MAC / Sanity Check / RDCO ops / new-bet-evaluation). Each has known bottleneck locations from today’s audits.
- Q1 chain question: ask through 4 layers (targeting → instrumentation → tools → feedback) instead of a flat chain. Our chain has structure built in.
- Q3 lock-in test: layered. Real-numbers test PLUS “does this survive the P&L meta-layer veto?” (vertical-farming pattern)
- Diagnosis card: writes back to the bet-stack YAML at
~/rdco-vault/07-bet-stacks/<slug>.yaml— auto-updatescritical_componentfield + appends adecision_tracesentry via/log-bet-decision. - Cross-bet payoff column: when a bottleneck is identified, surface whether it shares shape with another bet’s load-bearing gap (modular-component opportunity flag).
Add that Lehmann doesn’t have
- Persistent state via the bet-stack YAML. Lehmann’s skill is one-shot per session; ours remembers + cross-references prior diagnoses.
- Cross-bet pattern detection: when a bottleneck is identified, automatically check whether the same shape exists in another bet. Concrete example: Squarely’s Amazon-ads-sensor gap and SC’s engagement-dashboard gap ARE the same shape — Lehmann’s skill can’t make that finding because it’s single-business-scoped.
- Auto-logging: invokes
/log-bet-decisionto update the bet’s stack-view YAML. - HQ integration: the diagnosis card surfaces in the per-bet HQ view (depends on PR #3 merge + Phase B).
Skill scaffold (when built)
~/.claude/skills/identify-bottleneck/SKILL.md — invoked as /identify-bottleneck <bet-slug> or /identify-bottleneck (interactive bet-pick). Reads the bet-stack YAML for prior context, runs the Socratic interview, writes diagnosis card + auto-updates YAML.
Build now or queue?
Today’s audits (Squarely 2026-04-30 morning, MAC + SC 2026-04-30 afternoon) already produced the same shape of artifacts the formalized skill would produce. Don’t need the skill urgently.
Worth building when:
- A 4th bet is being evaluated (need the formalized intake)
- Cross-bet capability investment needs the discipline (e.g., “should we build the engagement-dashboard primitive now?”)
- The HQ Phase B work is unblocked (PR #3 merged) — then the diagnosis card has a render surface
Until then, the playbook page + the 3 audit notes carry the same content. Queue as a Notion task; build when a triggering need surfaces.
Cheap immediate win
The false-bottleneck patterns list (the 7 symptom patterns above) is high-leverage AND cheap to fold in NOW. Add as a discipline reminder section to the existing 06-reference/2026-04-30-rdco-bet-architecture-playbook.md. Founder reviewing his own bet’s audit can self-check against the list. ~5 min edit, no new skill needed.
Related
- 2026-04-30-rdco-bet-architecture-playbook — RDCO’s richer version of the same operating discipline
- 2026-04-30-rdco-thesis-targeting-systems-feedback-loops — canonical thesis; bottleneck-identification is the prioritization filter at work
- 2026-04-30-dickie-bush-how-to-think-like-a-billionaire-without-ever-meeting-one — Bush #7 (single biggest bottleneck → full attention → remove → next) is the same operating system at the individual-operator level
- 2026-04-28-mrbeast-production-playbook — MrBeast’s Critical Component field discipline is the bet-level analog
~/.claude/projects/-Users-ray/memory/feedback_targeting_system_prioritization_filter.md— Ray’s behavioral filter for capability questions~/.claude/skills/log-bet-decision/SKILL.md— RDCO’s auto-logging skill (analog to the X-post tool’s “playbook” output)