model-ticket-cost

name	model-ticket-cost
description	Use when modeling annual ticket-administration cost for an existing client or a prospect — OSINT T-shirt sizing, analog selection, and Low/Base/High cost scenarios from measured effort priors.
argument-hint	<client-or-prospect-name> [--analog <existing-client>] [--volume <tickets/mo>]

Model Ticket Cost

Size a client or prospect via OSINT and model their annual ticket-administration cost from measured effort priors. Executes Phases 3–4 of the effort-analysis methodology.

The Iron Law

NO COST FIGURE WITHOUT SCENARIOS AND STATED EXCLUSIONS FIRST

Every cost model presents Low/Base/High scenarios and states — verbatim — that figures exclude executing the work. A single point estimate without exclusions gets pasted into a pricing deck as "the cost" and is wrong twice.

Announce at Start

Before any other action, say verbatim:

I am using the model-ticket-cost skill to size the target and model annual ticket-administration cost with Low/Base/High scenarios.

Red Flags

Thought	Reality
"The data broker says $218M revenue, good enough"	Data-broker revenue for private subsidiaries is often wrong by orders of magnitude. Headcount is the reliable signal; verify load-bearing numbers.
"Big client, so lots of tickets"	Size does NOT predict volume — monitored footprint and tuning maturity do. Use an analog client's actual volume.
"I'll research the name directly"	Disambiguate via the Assets/CMDB record first. Generic names researched blind poison the model.
"One annual figure is cleaner"	Iron Law violation. Low/Base/High, always.
"Use the billable rate — it's more impressive"	Lead with burdened cost; opportunity rate is upside, valid only if analysts are utilization-constrained.
"Steady-state volume from day one"	Onboarding runs 2–4× steady-state for the first months. State it in risk notes.
"Rates can go in the deliverable committed to the repo"	Rates, client names, and volumes stay in local artifacts and the local priors file. Never committed.
"No effort priors? I'll estimate minutes per ticket"	Run /analyze-ticket-effort first, or use its priors file. Invented effort numbers are not measurements.

Prerequisites

• artifacts/metrics/effort-priors.yaml with measured effort_priors and rates populated (run /analyze-ticket-effort first if missing)
• Perplexity MCP for OSINT sizing (falls back to WebSearch)
• jr CLI for analog-client volume queries

Workflow

Stage 1: Disambiguate and size (Phase 3)

1. Pull the target's internal record (jr --output json assets view <KEY>) if it exists — location/sector context before any external query.
2. Dispatch the osint-researcher agent (Harper) for organizational sizing: revenue, headcount, sector scale metric, with citations and trust tiers.
3. Bucket on organizational scale (headcount primary, revenue secondary): Large / Medium / Small per the method doc definitions.

Stage 2: Pick the analog (Phase 4)

1. For a prospect: choose the closest existing client — same sector + scale — as the volume/type-mix analog. State the choice and why.
2. Query the analog's actual volume and type mix via jr (or reuse a fresh Phase-1 baseline).
3. Use the size-bucket average as the Low anchor; --volume overrides when the caller has better information.

Stage 3: Model (Phase 4)

volume (tickets/mo) × type mix × per-type effort (priors, with overhead)
= hours/yr × rate = annual ticket-administration cost

1. Compute Low/Base/High volumes → hours/yr → burdened cost (leads) and opportunity cost (upside).
2. Attach the standard risk notes: onboarding spike (2–4×), double-entry/mirroring, priors-measured-on-one-client caveat.
3. If asked about severity-gated pricing tiers, follow the method doc §4.4: quote the math, flag vendor-label severity honestly, and offer the better levers (gate ticketing not triage, digest tickets, tuning-first, automation of templated types).

Stage 4: Deliverable

Write artifacts/metrics/cost-model-<target>-<date>.md: scenario table, sizing findings with citations, analog rationale, risk notes, and the mandatory exclusion statement. Present the summary inline. Never put rates or client identifiers in anything that leaves artifacts/.

References

• ${CLAUDE_PLUGIN_ROOT}/data/effort-analysis-method.md — canonical method (Phases 3–4)
• ${CLAUDE_PLUGIN_ROOT}/templates/effort-priors-tmpl.yaml — rates and priors live here, locally
• /analyze-ticket-effort — produces the effort priors this model consumes