OpenKT — knowledge synthesis across 5 teams
memories → MemMachine facts → synthesized KBs (conflicts · forks · relations) · connected knowledge graph · 2026-05-30
the team building this product · 30 memories from 3 people → 7 KBs · 4 conflicts · 2 forks
Knowledge graph — click any node to read it. ● memory · ▦ knowledge base · ◆ person · ━ relation
People — who is contributing what
Knowledge bases
Memory Architecture: MemMachine vs OpenKT Roles Claude, Pratham
MemMachine is the memory engine: ingest, extract episodic + semantic signals, recall via /search endpoint. OpenKT is the knowledge layer: synthesize MM's signals into team knowledge bases and skills. OpenKT does NOT recall; it builds FROM MemMachine's signals. Episodic memory in Neo4j; semantic features in Postgres. Recall never reimplemented in OpenKT.
Claude: Recall returns both MemMachine raw facts + OpenKT KB briefs in meta.knowledge_bases
Pratham: Recall must stay MemMachine-only; KB is built FROM MM signals, not a second recall path
Knowledge Base Synthesis & Clustering Claude, Pratham
Retired cosine similarity threshold (0.75) — Titan similarities for related memories only 0.15–0.32, never clustered. Moved to LLM-based synthesis: project-level stage clusters topics, writes current-state summaries, detects supersedes vs open forks, forms relations. Simple v3 prompt outperformed over-prescriptive v2; kept lean. Forks require contributor-attributed sources (semantic features strip contributor).
Embedding & Vector Strategy Claude, Pratham
Single canonical embedding model: Amazon Bedrock Titan v2. Chosen for cost, IAM-native auth, and unified system architecture. BGE entirely eradicated. Interim: OpenKT calls Bedrock directly; post-DB-collapse: OpenKT will reuse MemMachine's stored Titan vectors to eliminate double-embedding.
Database & Infrastructure Consolidation Claude, Pratham, server-agent
One shared Postgres (15432 locally) for both MemMachine and OpenKT. MemMachine coexists in public schema after Alembic migration workaround. Neo4j (17687) stores episodic memory. RabbitMQ (15673) for async jobs. Region: us-east-1 (Bedrock Titan v2 availability constraint; Singapore ap-southeast-1 unavailable). Terraform state drift must be managed with -target flag.
Pipeline Architecture & Job Tracking Claude, server-agent
Per-memory pipeline: preprocess → embed (Titan) → triage → classify_kb, with mm-sync running in parallel on 5-minute cron. Every stage tracked in agentic_jobs with idempotency keys, retries, dead-letter queue. Fallback: when MemMachine unhealthy, OpenKT reads features directly from Postgres via Drizzle so recall never hard-fails. End-to-end shakeout passed on local docker-compose.
Authentication & Access Control Claude, Pratham
Email + password only — no magic links, no social login. Dual auth layer: Supabase JWT for dashboard; Clerk webhook mirrors users and orgs into RDS. One dispatcher resolves either token into ActorContext.
Pratham: Email + password only; no magic links or social login
Claude: Dual auth: Supabase JWT + Clerk webhook for flexibility
Product Vision & Deployment Pratham, Claude
OpenKT is the context cloud for AI-native teams — knowledge bases prevent agents from repeating work already done. Knowledge bases convert into skills via Create Skills plugin. OpenKT ships as Claude Code plugin with prompts, skills, MCPs, required scripts. Ship fast: correct + minimal + shipped beats complete + perfect + siloed. Month-long build cycle; ready to go live.
B2B sales team · 36 memories from 6 people → 9 KBs · 2 conflicts · 1 forks
Knowledge graph — click any node to read it. ● memory · ▦ knowledge base · ◆ person · ━ relation
People — who is contributing what
Knowledge bases
Pricing Strategy & Deal Economics Dana, Nadia, Marcus, Greg
The company transitioned from per-seat ($50/user/month) to usage-based pricing (per 1,000 API calls) with committed-usage tiers and overage rates, effective next quarter. Legacy seat deals are grandfathered through renewal. AEs can approve discounts up to 15% without VP approval; deeper discounts route to Dana. The team will not offer deep discounts (30%+) as they lead to low usage and churn; customer fit is prioritized over winning poor-fit logos. Usage-based pricing is expected to shorten CAC payback (currently 14 months blended) by accelerating expansion revenue.
Forecast Accuracy & Pipeline Management Nadia, Dana, Marcus, Priya
Forecast categories are Commit / Best Case / Pipeline with strict enforcement rules. Deals without both amount and close date are excluded ('no fields, no forecast credit'). Opportunities slip out of Commit if not advanced within 14 days. Weekly Monday forecast calls require reps to bring next-step and close-date for all Commit deals or deals are pulled. Economic Buyer identification is non-negotiable for Commit category. Pipeline coverage target is 3x quota per quarter; SDR capacity opens when coverage falls below target. The team missed Q2 forecast by 18% due to 'happy-ear' deals with no Economic Buyer engaged, driving stricter enforcement. Pipeline attribution tracks UTM parameters end-to-end from lead to opportunity by channel.
Trial & POC Management Marcus, Tomas, Dana, Nadia
The team conducts structured trials with kickoff calls where 2-3 success criteria are defined upfront and documented in deal notes. Standard trial length was extended from 14 to 30 days after field feedback indicated 14 days was insufficient for real evaluation cycles. Trial dates auto-sync to Salesforce opportunity close dates for accurate forecasting. Success criteria travel the full customer lifecycle and become the agenda for QBRs, creating continuity from trial through renewal.
Sales Presentation & Deal Framing Strategy Dana, Marcus, Tomas
First-call presentations use single-slide problem framing rather than 12-slide corporate decks, which converts significantly better. Against Acmeflow, the team leads with security/compliance depth (native SSO + audit logs) demonstrated in the first 10 minutes, positioning implementation speed (30 days vs Acmeflow's 90 days) as a quantified differentiator in every competitive deal.
Marcus: Price is the closing factor in procurement-driven deals; we win on 20% list price advantage vs Acmeflow
Tomas: Never lead with price vs Acmeflow; it commoditizes the product and trains buyers to negotiate; lead with security/compliance depth instead
Sales Operations & Technology Stack Nadia, Priya, Tomas
Salesforce is the single source of truth for the team. Outreach manages sequences; Gong records calls and tracks objections. Nightly automated sync maintains data consistency across Salesforce, Outreach, and Gong. Trial start/end dates auto-sync to opportunity close dates. SDR-sourced leads carry original campaign UTM through to opportunities for pipeline attribution by channel. Objection library is systematically built in Gong tracker to support team learning.
Sales-to-Customer Success Handoff & Renewal Management Greg, Marcus, Tomas
Sales-to-CS handoff occurs within 48 hours of close via structured handoff documentation including success criteria, champion, and risks. Success criteria from trial kickoff travel the full lifecycle and become the literal agenda for QBRs. Renewal conversations open 90 days before term end to prevent day-30 surprises that lose renewals. Net revenue retention is the CS north-star metric. CS books expansion revenue after day-1 handoff; original AE does not receive credit for post-sale expansion. Multi-threading requirement now mandates minimum 3 contacts per account to prevent single-champion dependency and account resilience risk (triggered by $120k renewal loss when champion departed).
Sales Qualification & Discovery Methodology Dana, Marcus, Priya, Tomas
The team uses MEDDICC framework for qualification with strict enforcement: opportunities cannot advance to Stage 3 without identified Economic Buyer and Champion. BANT is the minimum standard for SDR-qualified meetings (15 per month target). Discovery relies on the 'what happens if you do nothing?' question to surface cost of inaction as a deal forcing function. Cold outreach leads with customer proof points from the same vertical rather than feature lists, which has doubled reply rates.
Target Customer Profile & Segment Strategy Dana, Priya, Marcus
ICP is mid-market fintech with 200-2000 employees and a named security owner, where win rates are highest. Outbound strategy shifted from spray-and-pray SMB lists to ICP-only targeting. Empirical analysis showed SMB segments have faster sales cycles (62 days vs ICP) but churn within two quarters, making ICP the preferred segment despite longer cycles. Pipeline quality is prioritized over volume; non-closing SMB deals were eliminated from outbound lists.
Territory Model & Compensation Structure Dana, Nadia
Territories are split by region, not named accounts. SDRs and AEs share the same book and commission on team-sourced pipeline, creating alignment on shared pipeline quality.
clinical care team · 37 memories from 6 people → 12 KBs · 4 conflicts · 1 forks
Knowledge graph — click any node to read it. ● memory · ▦ knowledge base · ◆ person · ━ relation
People — who is contributing what
Knowledge bases
Antibiotic Stewardship & De-escalation Dr. Patel, Sam Okoro, Dr. Liang, Dr. Ekwueme
The team balances rapid empiric coverage with systematic de-escalation. Empiric HAP coverage recently shifted from piperacillin-tazobactam to cefepime plus vancomycin based on 2026 antibiogram showing rising pip-tazo resistance in Pseudomonas; vancomycin retained for MRSA. All broad-spectrum starts trigger automatic 48h EHR stewardship review prompts mandating de-escalation. At 48-72h post-culture, narrow to organism-specific therapy and discontinue broad-spectrum. Every antibiotic order documents indication and planned stop date at time of ordering to prevent unnecessary prolongation. The team recognizes operational tension between rapid sepsis bundle initiation and stewardship discipline—both are non-negotiable.
Discharge Planning & Readmission Prevention Thandi, Sam Okoro
No discharge occurs without (1) pharmacy medication reconciliation completed and (2) patient teach-back documented. High-readmission-risk patients (LACE≥10) receive follow-up call within 48h and appointment booked before discharge. For CHF specifically, documented weight-monitoring plan and diuretic regimen are required before discharge (protocol implemented after 6-day readmission due to missing scale/plan). SNF placement averages 2.3 days to arrange; referral initiation begins on admission for likely SNF patients, not at discharge. Pharmacy owns final medication list at discharge, including DOAC vs warfarin and AUC-vanc decisions.
Inpatient Glucose Management Dr. Ekwueme, Dr. Liang, Sam Okoro, Rosa
The team implements permissive glycemic control targeting 140–180 mg/dL on the medical floor, avoiding tighter targets due to hypoglycemia risk. Evidence base: NICE-SUGAR trial demonstrated harm from tight control. Point-of-care glucose checks occur before meals and at bedtime; any reading <70 mg/dL triggers hypoglycemia protocol immediately. Basal-bolus insulin is standard; sliding-scale-only regimens are rejected for type 1 diabetes and steroid-induced hyperglycemia as they chase highs reactively rather than prevent them. Nocturnal hypoglycemia is recognized as an independent fall risk factor.
Dr. Liang: Tight control 110–140 mg/dL to improve infection outcomes
Dr. Ekwueme: Permissive 140–180 mg/dL to minimize hypoglycemia risk (NICE-SUGAR evidence)
Sepsis Recognition & Initial Management Rosa, Dr. Ekwueme, Sam Okoro, Dr. Liang
The team implements a rapid-response sepsis protocol with two entry points: (1) ED triage uses qSOFA quick-screen (RR≥22, altered mentation, SBP≤100); any 2 criteria trigger immediate rapid-response activation. (2) Once sepsis suspected, the 1-hour bundle executes: lactate draw, blood cultures before antibiotics, broad-spectrum antibiotics, 30 mL/kg crystalloid for hypotension or lactate ≥4. Door-to-antibiotic target is <60 minutes for septic shock. Serial lactate trends (repeat at 2-4h if elevated) guide ongoing resuscitation more than single values.
Code Status & End-of-Life Care Dr. Ekwueme, Dr. Liang, Thandi
Code status is addressed and documented within 24h of admission for every patient over 75 or with advanced illness. The team clearly distinguishes DNR (do-not-resuscitate) from DNT (do-not-treat): DNR means no CPR but does not preclude escalation of care short of CPR. Palliative consult is offered early for metastatic cancer and end-stage organ failure, not reserved for final 48 hours of life.
Postoperative Pain Management & Opioid Stewardship Dr. Ekwueme, Sam Okoro, Rosa
Post-operative pain management is multimodal-first: scheduled acetaminophen + NSAID with opioids reserved for breakthrough pain only. The team has moved away from opioid-first ordering. Pain must be reassessed within 60 minutes of any opioid dose and documented; un-reassessed doses trigger chart-audit flags. At discharge, every opioid prescription includes naloxone co-prescription and is limited to 5-day supply for acute pain. This systematic approach prioritizes minimizing opioid exposure while maintaining effective analgesia.
Atrial Fibrillation & Anticoagulation Dr. Ekwueme, Sam Okoro, Rosa
For new non-valvular AFib, the team defaults to DOAC (apixaban) and has discontinued routine warfarin bridging except for mechanical valves and severe renal disease. Apixaban requires dose reduction to 2.5mg BID if two of: age≥80, weight≤60kg, creatinine≥1.5; pharmacy verifies every order. Perioperative enoxaparin management: hold 12h before procedures and document last-dose time on board for anesthesia verification. Pharmacy performs final medication reconciliation at discharge, serving as arbiter of DOAC vs warfarin decisions.
Infection Control & Contact Precautions Dr. Patel, Rosa
Contact precautions are implemented for MRSA and C. difficile patients. For C. difficile specifically, soap-and-water hand hygiene is mandatory because alcohol-based gel does not kill spores. Central-line dressing is changed q7d or when soiled; daily chlorhexidine (CHG) bathing is performed for all line patients to reduce CLABSI risk.
Fall Risk Assessment & Prevention Rosa, Dr. Liang
The team implements structured fall risk screening using Morse Fall Scale on admission and every shift. Scores ≥45 trigger yellow band, bed alarm, and hourly rounding. Pressure injury prevention uses Braden Scale daily for high-risk patients (score ≤14); immobile patients are repositioned q2h and heels are floated for Braden ≤14. Nocturnal hypoglycemia is recognized as an independent fall risk factor; a documented case of a patient falling after 3am hypoglycemic episode reinforced the link between glucose management and fall prevention.
Vancomycin Dosing & Monitoring Dr. Patel, Sam Okoro
Vancomycin is now dosed by AUC-guided monitoring using Bayesian pharmacokinetic calculations run by pharmacy on every patient, replacing trough-only monitoring. Pharmacy owns the calculator and dosing decisions. Clinicians no longer order reflexive daily vancomycin troughs; this practice caused unnecessary venipunctures and targeted the wrong parameter under AUC-based dosing.
Antibiogram & Resistance Surveillance Dr. Patel, Sam Okoro
Institutional resistance surveillance data (antibiogram) directly feeds empiric antibiotic protocol updates. The team views antibiotic stewardship and antibiogram as a closed feedback loop: resistance patterns inform empiric choices, which are then de-escalated based on culture sensitivities. The 2026 HAP protocol shift from pip-tazo to cefepime exemplifies this data-driven approach.
VTE Prophylaxis Dr. Liang, Dr. Ekwueme
VTE prophylaxis is prescribed on every hospital admission unless contraindicated. Default is enoxaparin 40mg SC daily; for renal impairment, switch to heparin 5000u q8h. Systematic protocol ensures no admission is missed.
corporate legal team · 36 memories from 6 people → 17 KBs · 1 conflicts · 1 forks
Knowledge graph — click any node to read it. ● memory · ▦ knowledge base · ◆ person · ━ relation
People — who is contributing what
Knowledge bases
Liability Cap Strategy & Structure Derek, Helena, Raffi
The company has transitioned from a flat $1M aggregate liability cap to a proportional 12-months-fees model, effective for all new deals. This change addresses asymmetric outcomes on small vs. large deals. Legacy contracts retain the $1M cap until renewal. The indemnity cap is carved separately from general liability and defaults to 12-months-fees unless negotiated higher. Uncapped indirect/consequential damages are a non-negotiable walk-away term requiring escalation to Helena. When customers resist, Derek uses mutual limitation of liability as a fallback, accepting symmetry to maintain favorable cap levels.
Contract Lifecycle Management & Metadata Raffi, Derek
Every executed agreement is filed in the CLM with metadata: counterparty, value, renewal date, auto-renewal flag, and assigned attorney. Auto-renewal clauses are monitored with 90-day pre-renewal alerts to prevent unintended vendor lock-in. Governing law and venue are tracked for quick forum exposure assessment.
Employment Law Compliance & Terminations Carla
The company tracks FTC non-compete rule changes and state bans, adapting employment agreements to use trade-secret and non-solicit protections. For any RIF or termination involving protected class thresholds, an adverse-impact analysis is conducted and severance is offered in exchange for a release under OWBPA. Contractor misclassification is a live risk; the ABC test is applied in California to assess 1099 vs. W-2 reclassification exposure, focusing on control and independence prongs.
EU Data Transfer Compliance & GDPR Aisha, Derek, Raffi
Cross-border EU data transfers rely on the 2021 EU Standard Contractual Clauses with Transfer Impact Assessments; Privacy Shield language has been retired. Every vendor handling personal data must sign a DPA before go-live—Raffi enforces this gate in procurement with no exceptions. GDPR breach notification to supervisory authorities must occur within 72 hours of awareness; the team maintains a pre-drafted notice template for rapid compliance. DPA liability provisions are tied to MSA caps, so the 12-months-fees liability change directly affects data-protection exposure. Vendor security review and DPA signing are combined into a single gate with joint sign-off from privacy and security teams.
Option Pool & Equity Dilution Strategy Ben, Helena
Standard new-hire option vesting is 4-year with a 1-year cliff; early exercise is allowed for the first 90 days. The team disagrees on option pool refresh timing before Series B.
Ben: Refresh the pool to 15% BEFORE the Series B term sheet so dilution lands on existing investors rather than founders alone
Helena: Do NOT pre-emptively top up the pool; size it in the round to avoid diluting founders ahead of a valuation we haven't locked
IP Indemnification & Third-Party Claims Helena, Derek, Raffi
The company has evolved from unilateral (customer-only) IP indemnification to mutual indemnification for third-party IP claims arising from each party's materials. Indemnity is scoped to third-party claims specifically, not all claims. Conditions include prompt notice, sole defense control by the indemnifying party, and prohibition on customer settlement without consent. The indemnity cap is carved out from general liability and defaults to 12-months-fees unless negotiated higher.
Equity Grant & Cap Table Management Ben, Raffi
83(b) elections must be filed within 30 days of grant for early-exercised shares with no IRS extensions available; Ben sends reminders at signing. Cap table is maintained in Carta with all grants, exercises, and SAFE conversions reconciled before board meetings.
IP Assignment & Employment Agreements Carla, Ben
All employees and contractors sign a PIIA (Proprietary Information & Invention Assignment) with IP assignment and confidentiality provisions before day one—no exceptions. The PIIA assignment chain must be clean for patent value preservation; broken chains compromise patent enforceability in M&A and funding diligence. Non-competes have been dropped entirely from offer letters following the FTC rule and state bans; the company now relies on trade-secret and non-solicit protections instead.
Vendor Procurement & Security Gate Aisha, Raffi
Vendor security review and DPA signing are combined into a single gate; privacy and security teams sign off together before a data processor goes live. Raffi enforces DPA compliance during vendor procurement onboarding with no exceptions.
Dispute Resolution & Governing Law Helena, Derek, Raffi
Dispute resolution has been updated from courts-of-NY to binding AAA arbitration with a class-action waiver, seated in Delaware. Delaware is the preferred governing law for commercial deals; the company resists customer attempts to flip to their home state unless the deal is strategic. Every signed contract's governing law and venue are tracked in the CLM for quick forum exposure assessment.
Board Approval & Delegation of Authority Helena, Raffi, Derek
Board approval is required for contracts over $500k TCV, any debt, and any equity issuance. Raffi flags board-level items to the approval queue. The company maintains a tiered delegation-of-authority matrix allowing commercial counsel to sign contracts up to defined thresholds without General Counsel escalation, balancing operational efficiency with governance.
Trade Secret Enforcement & Exit Procedures Carla
The company has practical experience enforcing trade secret claims against departing employees who misappropriate source code. Thorough exit interview documentation is critical for claim viability. A well-papered PIIA combined with detailed exit interviews has proven effective in trade secret litigation.
Trademark Filing & Pre-Launch IP Clearance Ben
Trademarks must be filed in the US, EU, and UK at minimum on any new product name before public launch, as first-to-file jurisdictions require early filing. Trademark filing and clearance in key jurisdictions are mandatory before any new product name goes public.
Litigation Hold & E-Discovery Readiness Helena, Raffi
Litigation holds are issued the moment litigation is reasonably anticipated. IT preserves mailboxes and systems before anyone accesses them, ensuring e-discovery readiness. The team understands the reasonable anticipation trigger and implements preservation proactively.
Patent Strategy & Provisional Applications Ben, Carla
Provisional patent applications are used to secure priority dates; the full filing decision is made within the 12-month statutory window. Clean PIIA assignment chains are essential for patent value preservation in M&A and funding diligence.
US Privacy Law & Product Compliance Aisha
CCPA/CPRA 'sale/share' opt-out requirements are handled at the product layer (product team owns the toggle); legal owns the privacy policy disclosure. This split responsibility ensures technical implementation and legal transparency are coordinated.
Compliance Training & SOC 2 Controls Aisha
Annual security and privacy training is mandatory and tracked as an audited SOC 2 Type II control. Completion is attested to in compliance reports.
growth marketing team · 37 memories from 6 people → 11 KBs · 1 conflicts · 1 forks
Knowledge graph — click any node to read it. ● memory · ▦ knowledge base · ◆ person · ━ relation
People — who is contributing what
Knowledge bases
Lead Nurture & Lifecycle Email Omar, Pavel
Paid leads route into 5-email nurture sequence before sales contact; raw leads skipping nurture convert at half the rate (f30-f31). Program migrated from batch-and-blast to triggered behavioral emails keyed to product activation events (f64-f66). Onboarding sequence prioritizes single activation milestone (first project created) over all secondary metrics in first 7 days (f68-f70). Sunday 6pm email sends outperform Tuesday morning for B2B audience—counterintuitive but validated across 3 months of testing (f72-f74). Multi-month A/B tests required before committing to optimization; short-term results not sufficient (f75). Email deliverability managed with seed-list validation and warmup checks before list imports; learned from Gmail throttling incident (f76-f80). Post-event follow-up implemented as tailored lifecycle track tied to session content attended, distinct from generic nurture (f105-f106).
Attribution Modeling & Measurement Pavel, Jamal, Sofia
Team has migrated from last-touch attribution (deprecated) to a three-source triangulation framework combining multi-touch attribution (MTA), media-mix modeling (MMM), and self-reported attribution. Last-touch systematically over-credited branded search and under-counted top-of-funnel channels, distorting budget allocation. Current approach: no single attribution source is trusted in isolation. Sofia's self-reported attribution ('How did you hear about us?' on demo form) serves as tiebreaker when MTA and MMM conflict. Pavel validates MTA against quarterly MMM cross-checks. This resolved the attribution conflict that previously drove budget misallocation.
Brand vs. Performance Marketing Budget Allocation Jamal, Sofia, Yuki, Pavel
OPEN FORK: Jamal advocates ~10% brand budget cap, prioritizing performance spend where pipeline-per-dollar is demonstrable (f32). Sofia counters with 30% brand budget floor, arguing performance harvests demand that brand creates; reducing brand raises CAC across all channels within two quarters (f36-f37). Yuki supports Sofia with empirical proof: unaided brand awareness lifted from 8% to 14% over one year when brand budget was maintained—a lift performance cannot manufacture (f39-f41). Pavel provides causal evidence: measurable 2-3 week lag between brand-search volume and blended CAC reduction (f43-f45), validating Sofia's systems-thinking model (f38, f42). Jamal's position rests on attribution opacity of brand; Sofia's rests on systems causality and empirical CAC correlation. Both hold live positions with supporting data.
Jamal: Cap brand/awareness at ~10% of budget; reallocate to performance where pipeline-per-dollar is measurable
Sofia: Protect brand at 30% minimum; performance harvests demand brand creates; starving brand raises CAC across all channels within two quarters
Positioning & Messaging Strategy Sofia, Yuki, Lena
Repositioned category from 'project tool' to 'team knowledge layer' (f87). Messaging house framework cascades from one core line into broader structure (f88). Content asset structure leads with customer problem in their words, then category, then product—never product-first (f90). Win/loss interviews show buyers differentiate on knowledge-synthesis angle, not feature parity; positioning leans into differentiation (f97-f99). New positioning implemented across homepage, BOFU comparison pages, and ad hooks for cross-channel consistency (f93-f96). One consistent message prioritized across all surfaces (f95).
Content Strategy & SEO Lena
Migrated from keyword-volume chasing to intent-based topic clusters with pillar pages plus supporting posts per search intent (f46-f49). Every post mapped to TOFU/MOFU/BOFU funnel stage and primary intent before writing, not after (f50). BOFU comparison/competitor pages convert ~5x blog average and feed sales directly; prioritized accordingly (f51-f53). Abandoned thin AI-generated content at scale after experiencing traffic loss from Google core update; learned quality-over-quantity principle (f54-f59). Topical authority damaged by low-quality content; now prioritizes quality-first approach (f55, f58). Content categorized by function: BOFU pages classified as harvest layer (converting existing demand), not demand creation (f61).
Analytics Rigor & Testing Discipline Pavel, Jamal
No campaign ships without tracking plan: UTMs, named primary KPI, and defined minimum detectable effect (f82). A/B tests held to 95% significance with pre-registered sample sizes; no early stopping when results look promising (f91-f92). CTR treated as vanity metric; only meaningful scorecard is pipeline volume and CAC payback period (f83). Empirical testing preferred over conventional wisdom; counterintuitive results (e.g., Sunday 6pm email sends) validated before adoption (f73).
Performance Marketing Operations & Channel Management Jamal, Pavel
Blended CAC target is $400 with 3-month payback threshold (f18-f19). Channels underperforming targets are paused and reworked rather than scaled (f20). Scaling requires 50+ conversions minimum; below that, CPA variance is treated as statistical noise (f25). LinkedIn drives lowest-CAC enterprise leads; Meta cheaper per lead but skews SMB/self-serve (f26-f28). Google Performance Max capped at 30% of paid budget due to lack of transparency and branded-search cannibalization (f22-f23). Platform CAC tracked by segment (enterprise vs SMB) to optimize channel mix (f27-f28).
Marketing Technology Stack & Data Architecture Pavel, Omar
Full-stack: HubSpot (marketing source of truth) syncs to Salesforce; GA4 plus BigQuery warehouse backs MMM (f112). Duplicate lead records across HubSpot and Salesforce managed with nightly automated dedupe rules keyed on email ownership (f115-f116). Consent and suppression lists synchronized across all email send platforms; unsubscribes honored everywhere, not just where they occurred (f117). Email deliverability managed with seed-list validation and warmup checks before list imports (f77-f79).
Field Marketing & Account-Based Marketing Sofia, Jamal
Shifted field budget from trade-show booths to executive dinners; dinners show higher pipeline-per-dollar for enterprise (f100-f101). ABM targets named 200-account list; ads, content syndication, and SDR outreach align to same list rather than open market (f102-f104).
Marketing Funnel Architecture & Channel Classification Jamal, Lena
Two-layer funnel model: creation layer (brand + TOFU content) generates demand; harvest layer (organic BOFU pages + paid retargeting) converts it (f60-f62). Channels categorized by function (creation vs harvest) rather than by paid/organic format (f63). This systems view aligns with Sofia's brand-performance causality model: brand and TOFU create harvestable demand; performance and BOFU harvest it.
Brand Asset Management & Creative Consistency Yuki
Brand guidelines and asset library centralized in one place; off-brand one-off creatives undermine consistency (f118-f119). Quality-over-quantity principle applied to creative assets; centralization essential for maintaining brand coherence.
restaurant ops team · 37 memories from 6 people → 8 KBs · 1 conflicts · 1 forks
Knowledge graph — click any node to read it. ● memory · ▦ knowledge base · ◆ person · ━ relation
People — who is contributing what
Knowledge bases
Front-of-House Service Standards & Recovery Amira, Dani
Service greeting standard is within 60 seconds of seating; water and bread are delivered quickly. Service pace is intentional and table-read (not rushed). Service recovery follows LAST model (Listen, Apologize, Solve, Thank). Every comped or remade plate requires manager approval and witness before leaving the kitchen. Allergy and VIP notes from Resy are reviewed at pre-shift briefing with entire floor staff before doors open. Comps are tracked nightly by reason code; spikes in 'long ticket time' comps indicate kitchen staffing issues rather than service errors.
Food Cost Management & Menu Engineering Kwame, Dani, Marco
The team targets 28% food cost (tightened from 32% due to sustained protein price increases) and achieves this through menu engineering, portion re-specification, and whole-animal yield tracking. Menu items are evaluated quarterly using a four-quadrant framework (stars, plowhorses, puzzles, dogs); high-margin low-sales items ('puzzles') are repositioned to featured menu spots rather than cut. Pricing strategy uses 3x plate cost multiplier. Whole-animal and whole-fish breakdown with zero-trim-waste strategy (bones and trim converted to stock and specials) is core to hitting the 28% target. Plate costs are re-specified tactically (portion sizes, garnish swaps) and loaded into POS recipe cards.
Inventory Management & Vendor Strategy Kwame, Rosa, Marco
Full physical inventory is conducted every Sunday night. Ordering is done to par based on forecasted covers, not intuition. Vendor diversification policy maintains primary-plus-backup vendor on every critical product; three-bid rule is used for high-spend new items to prevent single-outage 86s (learned from a single-vendor seafood outage that 86'd half the specials one Saturday). The 86 board and order guide are linked with automatic flagging for next order when items run out. Nightly high-value spirit counts are reconciled against POS; inventory variance is investigated same week and not carried into the following week.
Reservation & Seating Capacity Management Amira, Dani
The restaurant operates a reservation-led seating model via Resy with mandatory credit card holds and same-day confirmation texts. This reduced no-shows from 12% to 4%. Dining room is reserved; walk-ins fill the bar and any gaps. Table turn limits are enforced: 90 minutes for two-tops, 2 hours for larger parties; hosts communicate limits at booking. Reservation volume (approximately 200 covers on peak Saturdays) directly drives kitchen prep scope and labor scheduling.
Beverage Cost Control & Program Curation Theo, Kwame
Pour cost targets are 20% on spirits and 28% on wine by the glass. Cocktails are re-costed whenever base spirit price moves. Jigger discipline is enforced on every pour (no free-pouring); half-ounce drift per drink compounds significantly on high-volume nights. The signature cocktail is the highest-margin item in the building and is featured as an upsell on every check. Reserve wine list was consolidated from 40 bottles to 18 bottles (by-the-glass-forward model) to optimize working capital liquidity and eliminate slow-moving SKUs that tied up cash. Bar par levels and kitchen order guide share one vendor sheet to hit minimums and consolidate deliveries.
Kitchen Food Safety & Allergy Protocol Rosa, Marco, Dani, Amira
The kitchen maintains rigorous allergy workflow: fresh gloves, clean board, dedicated pan, and expo double-check before pass. Sesame and nut garnishes are stored in sealed bins off the main line to prevent cross-contact. Temperature danger zone protocol is enforced (cold ≤40°F, hot ≥140°F, 4-hour max in 40-140°F zone before disposal). FIFO rotation is enforced on all shelves and walk-in; all prep items are labeled and dated. Walk-in cooler temperature is logged twice per shift. Allergen matrix is posted at pass and on server tablets and updated same day as recipe changes. Pre-shift line check tastes every component and confirms par before service; nothing goes on the line untasted.
Operational Procedures & Accountability Systems Dani, Rosa
Opening and closing shifts run off printed checklists for consistent procedure execution. Closing manager signs temp log, cash drop, and line-down sheet nightly for accountability. Health-code criticals (handwashing, temperatures, labeling) are self-audited weekly so official inspections are routine non-events. Bar close reconciles register against POS nightly and counts high-value spirits as part of closing procedure.
Labor Cost Management & Scheduling Dani, Amira, Marco
Target labor cost is 30% of sales. Labor schedule is built to forecasted covers rather than prior-week flat templates. Staffing is adjusted tactically based on anticipated volume. Cut priority is floor staff before line; kitchen is kept fully staffed to maintain ticket times and guest experience. Staff are cut from the back of shift first as covers taper; the closer's station is protected to ensure sidework completion. Kitchen staffing directly impacts ticket times, which drives guest experience quality.
Amira: Pull a server and run larger sections; FOH can absorb it and guests barely notice on a slow night.
Marco: Never cut the line before the floor; a short kitchen blows ticket times and that's what actually kills the guest experience.