The Leverage Map: 90‑Day Focus Template for AI‑Augmented Services, Productized Services, or PLG Micro‑SaaS

Pick one and commit. No new offers until your end date.

• Path: [PATH] (choose one: AI‑Augmented Services | Productized Services | PLG Micro‑SaaS)
• Start date: [START_DATE]
• End date (90 days later): [END_DATE]
• Why this path (1–2 sentences): [WHY_THIS_PATH]
• No‑new‑offers pledge owner: [OWNER]
• Exceptions allowed? [YES/NO]. If yes, what qualifies: [EXCEPTION_RULE]
• Success definition by Day 90: [PRIMARY_OUTCOME] (e.g., “+$[TARGET_MRR] MRR and ≥[GROSS_MARGIN]% gross margin with churn ≤[CHURN_TARGET]%”)

Cover your life and ops first so you aren’t forced to pivot mid‑sprint.

• Base cost of living/month: [COL_MONTHLY_USD]
• Flights + visas amortized/month: [FLIGHTS_VISAS_MONTHLY_USD]
• Coworking/data/SIMs/month: [WORKSPACE_DATA_MONTHLY_USD]
• Essential software (non‑COGS)/month: [SOFTWARE_OPS_MONTHLY_USD]
• Buffer (10–25%): [BUFFER_USD]

Visa Run Revenue (VRR) floor = [COL_MONTHLY_USD] + [FLIGHTS_VISAS_MONTHLY_USD] + [WORKSPACE_DATA_MONTHLY_USD] + [SOFTWARE_OPS_MONTHLY_USD] + [BUFFER_USD]

VRR floor: $[VRR_USD]

Units to clear VRR by path (pick your unit):

• Services unit = 1 client month. Needed clients = CEILING([VRR_USD] / [CLIENT_ARPU_USD]).
• Productized unit = 1 subscription. Needed subs = CEILING([VRR_USD] / [SUB_PRICE_USD]).
• PLG unit = 1 paying user/account. Needed users = CEILING([VRR_USD] / [ARPU_USD]).

Budget the week you actually have, not the week you wish you had.

• Total hours available/week: [TOTAL_HOURS_PER_WEEK]
• Non‑work hours block (travel, errands): [LIFE_HOURS]
• CEO time (finance, hiring, sales reviews): [CEO_HOURS]
• Builder time (product/ops build): [BUILDER_HOURS]
• Billable/service delivery time: [BILLABLE_HOURS]
• Admin/comms buffer (15–20%): [BUFFER_HOURS]

Check: [TOTAL_HOURS_PER_WEEK] − [LIFE_HOURS] − [CEO_HOURS] − [BUILDER_HOURS] − [BILLABLE_HOURS] − [BUFFER_HOURS] = 0

Utilization targets by path:

• AI‑Augmented Services: Billable ≥ [SERV_BILLABLE_TARGET]% of work hours; Builder ≤ [SERV_BUILDER_CAP]%.
• Productized Services: Request throughput ≥ [REQUESTS_CLOSED_PER_DAY] per weekday; Turnaround SLA ≤ [SLA_HOURS].
• PLG Micro‑SaaS: Ship 1 critical improvement/week; Support ≤ [SUPPORT_HOURS_TARGET]/wk.

Note time‑zone friction: if serving [REGION], hold live windows [WINDOW_1] and [WINDOW_2].

Price so margin survives API, vector, and contractor costs.

Common inputs:

• Path: [PATH]
• Price per unit (client/sub/user): $[PRICE_PER_UNIT_USD]
• Expected ARPU (if tiered): $[ARPU_USD]
• Refund/chargeback allowance %: [REFUND_RATE_PCT]%

COGS lines (adjust for your stack):

• LLM tokens per unit: input [TOKENS_IN] • output [TOKENS_OUT]
• LLM price per 1M tokens: input $[PRICE_IN_PER_M] • output $[PRICE_OUT_PER_M]
• Image/gen costs per unit (if any): [IMAGE_CALLS_PER_UNIT] × $[IMAGE_PRICE_PER_CALL]
• Vector DB: pods [PODS] × $[POD_PRICE_PER_HOUR]/hr × 730 = $[VECTOR_MONTHLY]
• Inference hosting (GPU/API): $[INFERENCE_MONTHLY]
• Human help per unit (QA/ops/design): [HUMAN_MINS_PER_UNIT] mins × $[CONTRACTOR_RATE_PER_HR]/hr = $[HUMAN_COGS_PER_UNIT]
• Payment fees: [PAYMENT_FEE_PCT]% + $[PAYMENT_FEE_FLAT]

Formulas:

• Token COGS per unit = ([TOKENS_IN]/1,000,000)×[PRICE_IN_PER_M] + ([TOKENS_OUT]/1,000,000)×[PRICE_OUT_PER_M]
• Unit COGS = Token COGS + Image/gen + (Vector alloc per unit) + Human help + Payment fees
• Gross margin % = 1 − (Unit COGS / [PRICE_PER_UNIT_USD])

Defaults (edit these today from vendor docs):

• Example LLM (Claude Sonnet input/output): input $[DEFAULT_SONNET_IN]=3.00 /M, output $[DEFAULT_SONNET_OUT]=15.00 /M
• Example LLM (OpenAI mini/4x mix): input $[DEFAULT_OPENAI_IN] • output $[DEFAULT_OPENAI_OUT]
• Example vector pod: $[DEFAULT_POD_PRICE]=0.111/hr

Decision rule: If Gross margin < [MIN_GROSS_MARGIN]% at your median usage, raise price, cap usage, or change model before shipping the offer.

Only fill this if you touch LLMs/vectors.

• Average prompts/calls per paid unit: [CALLS_PER_UNIT]
• Avg input tokens/call: [AVG_INPUT_TOKENS]
• Avg output tokens/call: [AVG_OUTPUT_TOKENS]
• Cache/Batch savings estimate %: [CACHE_SAVINGS_PCT]%
• Effective token price after savings: input $[EFFECTIVE_IN] • output $[EFFECTIVE_OUT]
• Vector reads/writes per unit: [VEC_R_PER_UNIT] / [VEC_W_PER_UNIT]
• Cold‑start % (first‑time heavy calls): [COLD_START_RATE_PCT]%

Formula helpers:

• Effective input tokens/unit = [CALLS_PER_UNIT]×[AVG_INPUT_TOKENS]×(1−[CACHE_SAVINGS_PCT]%)
• Effective output tokens/unit = [CALLS_PER_UNIT]×[AVG_OUTPUT_TOKENS]×(1−[CACHE_SAVINGS_PCT]%)
• Vector alloc per unit (rough) = [VECTOR_MONTHLY] / [PAID_UNITS_PER_MONTH]

Guardrails:

• Alert if monthly LLM spend > $[LLM_SPEND_GUARDRAIL]
• Alert if vector bill > $[VECTOR_SPEND_GUARDRAIL]
• Freeze freebies if monthly COGS/Revenue > [COGS_REV_RATIO_GUARDRAIL]%

Match channel to model. Pick 1–2 channels only.

AI‑Augmented Services (examples):

• Partner directories (Zapier/Make/Shopify): [YES/NO] — Weekly target: [LISTINGS_UPDATED] listings refreshed; [REVIEWS_REQUESTED] reviews requested.
• Referral flywheel: [YES/NO] — Ask [REFERRAL_ASKS/WK] past clients/week.
• Targeted outbound (ops pain): [YES/NO] — [EMAILS/WK] emails; Offer: [OFFER_PROMISE].

Productized Services:

• Word‑of‑mouth + founder brand: [YES/NO] — Post [THREADS/WK] ops threads with before/after.
• Authority page + case studies SEO: [YES/NO] — Ship [CASE_STUDIES/WK] mini‑cases; update pricing page [PRICING_UPDATES/MO]×/mo.
• Affiliate partners: [YES/NO] — Recruit [AFFILIATES/WK] partners; 20% rev‑share default.

PLG Micro‑SaaS:

• Programmatic SEO: [YES/NO] — Publish [PAGES/WK] pages with [TEMPLATE_NAME] template.
• Audience/X + demo loops: [YES/NO] — [DEMO_CLIPS/WK] clips; [LIVE_DEMOS/MO] live demos/mo.
• Integrations/marketplaces: [YES/NO] — Submit/maintain [APPS/MO] listings.

Exit criteria per channel (define pass/fail):

• Cost/start ≤ $[COST_PER_START]
• Time spent ≤ [CHANNEL_HOURS_CAP]/wk
• Leading metric lift ≥ [LEADING_LIFT_TARGET]% by Week [WEEK_NUMBER]

Turn this into three 30‑day sprints.

Sprint 1 (Days 1–30):

• One‑sentence sprint goal: [SPRINT1_GOAL]
• Non‑negotiables (max 3):
  1. [S1_NONNEG_1]
  2. [S1_NONNEG_2]
  3. [S1_NONNEG_3]
• Ships (weekly): [S1_SHIP_LIST]

Sprint 2 (Days 31–60):

• Goal: [SPRINT2_GOAL]
• Ships: [S2_SHIP_LIST]

Sprint 3 (Days 61–90):

• Goal: [SPRINT3_GOAL]
• Ships: [S3_SHIP_LIST]

No‑new‑offers pledge: I, [OWNER], will not launch [FORBIDDEN_LIST] until [END_DATE]. Signature: [SIGNATURE].

Track 12 weekly gates. Pass only if all 3 are true. Edit metrics by path.

Columns (make this a 12‑row table):

• Week #: [WEEK]
• Date range: [DATE_RANGE]
• Leading metric A: [METRIC_A_NAME] — Target: [A_TARGET] — Actual: [A_ACTUAL]
• Leading metric B: [METRIC_B_NAME] — Target: [B_TARGET] — Actual: [B_ACTUAL]
• Quality/ship metric C: [METRIC_C_NAME] — Target: [C_TARGET] — Actual: [C_ACTUAL]
• Gate result (PASS/FAIL): [RESULT]
• Notes/decision: [NOTE]

Suggested metrics by path:

• AI‑Augmented Services: Qualified leads/wk; Proposals sent/wk; Client NPS or on‑time delivery %.
• Productized Services: Net new subs; Same‑day request closes; Weekly churn %.
• PLG Micro‑SaaS: New signups; Activation % (A→HA moment); Paid conversion % or WAU.

Name the obvious travel risks and pre‑decide the fix.

Risk entries (copy 5–8 rows):

• Risk: [RISK_NAME] (e.g., “Wi‑Fi loss >24h”, “Schengen 90/180 cap”, “Timezone slip”, “GPU/API rate‑limit”, “Payment processor flag”)
• Trigger (when it’s ‘on’): [TRIGGER]
• Impact: [IMPACT]
• Mitigation (prework): [MITIGATION]
• Action on trigger (who/what): [ACTION]
• Owner: [OWNER]

Example rows:

• Wi‑Fi loss >24h → Auto‑failover to [BACKUP_ISP] and push async updates in [CHANNEL]; owner [NAME].
• Schengen 90/180 → Plan sprints around [COUNTRY_A]→[COUNTRY_B] move on [DATE]; owner [NAME].
• LLM cost spike → Pause non‑critical tasks if LLM spend > $[LLM_SPEND_GUARDRAIL]; switch to [BACKUP_MODEL] for [TASK].

A 25–30 minute Monday ritual. Keep receipts (screenshots/links) in this doc.

Agenda:

1. Open last week’s row. Mark PASS/FAIL. If FAIL, schedule a scope cut today.
2. Update pipeline: [LEADS_OPEN], [DEALS_THIS_WEEK], [RISKS_TOP3].
3. Review margin quick‑check on 3 random units: price $[PRICE], unit COGS $[COGS], margin [MARGIN%]. If <[MIN_GROSS_MARGIN]% twice, raise price or cap usage.
4. Decide 1 keep / 1 change for the coming week. Log below.
5. Confirm travel constraints for next 14 days: [FLIGHTS], [BORDER_DAYS_LEFT], [TIMEZONE_SHIFTS].

Use or edit one as your Day‑1 starting point. They’re conservative and meant to be achievable.

A) AI‑Augmented Services → $15k MRR

• Offer: [SERVICE_OFFER] (e.g., AI receptionist/ops)
• Price/client: $[CLIENT_ARPU_USD]=1,250
• Needed clients: CEILING(15,000 / [CLIENT_ARPU_USD]) = [NEEDED_CLIENTS]
• Hours cap: [BILLABLE_HOURS]/wk; Max active clients = [BILLABLE_HOURS] / [HOURS_PER_CLIENT_PER_WEEK]
• Unit COGS: LLM $[COGS_LLM], Telephony $[COGS_VOICE], Human QA $[COGS_QA] → Total $[COGS_TOTAL]
• Margin/client: 1 − ($[COGS_TOTAL]/$[CLIENT_ARPU_USD]) = [MARGIN_CLIENT%]
• Channel: Partner directory + referrals. Weekly targets: [PROPOSALS/WK], [CASE_STUDIES/MO].

B) Productized Services → $20k MRR

• Offer: [PRODUCTIZED_OFFER] (e.g., unlimited AI video edits)
• Tier mix (example): [TIER_A] $[PRICE_A] × [COUNT_A] + [TIER_B] $[PRICE_B] × [COUNT_B]
• Subs needed: [COUNT_A] + [COUNT_B] = [SUBS_TOTAL]
• Throughput gate: Close ≥[REQUESTS_CLOSED_PER_DAY] requests/day, SLA ≤[SLA_HOURS].
• Team: [CONTRACTORS_COUNT] contractors × $[RATE]/hr for [HOURS]/wk.
• Churn guardrail: weekly churn ≤ [CHURN_TARGET]%.

C) PLG Micro‑SaaS → $30k MRR

• Pricing: $[PLAN_1_PRICE], $[PLAN_2_PRICE]
• ARPU assumption: $[ARPU_USD]=18.00
• Paid users needed: CEILING(30,000 / [ARPU_USD]) = [PAID_USERS_TARGET]
• Activation goal: [ACTIVATION_EVENT] ≥ [ACTIVATION_RATE_TARGET]% of signups
• Free→paid conversion: [FREE_TO_PAID_TARGET]% by Week [WEEK_TARGET]
• SEO cadence: [PAGES/WK] programmatic pages; Integration listings: [LISTINGS/MO]
• Infra budget: LLM ≤ $[LLM_BUDGET]/mo; Vector ≤ $[VECTOR_BUDGET]/mo; GPU/API ≤ $[INFERENCE_BUDGET]/mo.

One row per decision you made to keep/change something.

Columns:

• Date: [DATE]
• Decision: [KEEP_OR_CHANGE]
• What changed: [WHAT_CHANGED]
• Why (evidence): [WHY]
• Expected impact by [DATE]: [EXPECTED_IMPACT]
• Result (measured later): [RESULT]
• Next action: [NEXT_ACTION]

Changelog (versioning your template/process):

• [DATE]: v[VERSION] — [SUMMARY]
• [DATE]: v[VERSION] — [SUMMARY]