Pricing Packaging Tracker Agent Skill

Pricing & Packaging Tracker

Purpose

Track competitor pricing and packaging as a diffable time series, not a screenshot: **capture (or

diff) → structure changes → price changes → signals → next-step options.** The stable capture schema

is the point — run N diffs against run N-1, so the report is what *moved*, with evidence. And the

skill watches structure as closely as numbers, because packaging changes (gates, limits, tier

restructures) usually signal strategy earlier than price changes do: a tier that disappears telegraphs

a move upmarket months before the price list admits it.

Input

Works best with: the competitor set to track (or the snapshot they came from), and the prior

pricing capture if one exists — that's what turns this run into a delta report.

Also useful: your own pricing context, if the comparison should include you.

Input supplied inline with the invocation — text after the skill name, a pasted context dump, or an

appended ARGUMENTS: line — counts as answers already given. Use it against the question budget;

don't re-ask.

Arriving empty-handed? That works too. With no prior capture the skill runs baseline mode:

full capture per competitor, and the delta value starts next run. With no competitor list, it asks —

that's one of its two budget questions.

Example invocation: `Pricing tracker on [Competitor A], [Competitor B], [Competitor C] — prior

capture from April pasted below.`

Key Concepts

• Governing protocol: honors the autonomous-investigation

contract — question budget of 2, Fact/Inference/Assumption labels, Just Enough Mode, stable schema,

4-option Final Step. Discipline: SIGINT (pricing pages, site diffs) with FININT corroboration — see

intelligence-collection-disciplines.

• Time series, not screenshot. A single capture is trivia; the sequence is intelligence. The

schema never drifts between runs, because diffability is the entire value.

• Structure over numbers. Record tiers, billing units, feature gates, usage limits, free-tier

boundaries, and enterprise-floor signals — not just prices. The gate moves first; the price

follows.

• "Contact sales" is data, honestly labeled. Where pricing is opaque, track what *is* public

(editions, gates, published floors) and mark the rest. Infer only with labels; never invent

prices, tiers, limits, discounts, or negotiated figures.

• When NOT to use: you need pricing *strategy* advice →

finance-based-pricing-advisor (this tracks the

market; it doesn't set your price); pricing is fully opaque across the whole set → expect thin

results, and consider voice-of-customer-miner or win/loss

evidence instead.

Application

• Determine mode. Prior capture provided → delta mode (report changes only). None → baseline

mode (full capture, then stop).

• Credit inline context, then ask only the unanswered questions (max 2):
• Which competitors' pricing should I track?
• Prior capture to diff against, or is this the baseline?
• Capture from live pricing pages — plan-comparison pages, published rate cards, credible

pricing-change coverage — with URL and as-of date per competitor.

• Emit the schema below exactly — diffability is the point.

Output schema (do not reorder)

~~~markdown

Pricing Capture / Delta Report

1. Run Header

Competitor set: | Prior capture date: [or "baseline run"] | This run date:

2. Pricing Capture (per competitor)

[Competitor] — [pricing page URL, as-of date]

• Tiers: [name: price / unit / billing terms, one bullet each]
• Key gates: [which capabilities gate which tier, 2-4 bullets]
• Usage limits: [the limits that matter, 1-2 bullets]
• Free tier / trial: [boundary, 1 bullet]
• Enterprise signals: [published floors, "contact sales" scope]

(Baseline mode: full capture per competitor. Delta mode: this section

only for competitors with changes.)

3. Changes Since Last Capture (Delta Mode)

[Competitor] — [4 to 8 word change summary]

• Then / Now: [old -> new, labeled Fact]
• Evidence: [URL, date]
• Reading: [Inference — repositioning, monetization push, response to whom?]

If nothing changed: "No pricing or packaging changes this cycle."

4. Signals

• [Cross-competitor patterns: direction of the market's pricing, labeled]
• [Implications for your pricing or battle cards: 2-3 bullets]

Assumptions to Validate

• [Assumption 1] / [Assumption 2] / [Assumption 3]

~~~

A copy/paste fill-in version of this schema, with quality checks, lives in template.md.

Final Step (offer exactly 4 options)

• Update battle card pricing sections from these changes (battle-card-builder)
• Deep-dive one competitor's packaging logic
• Compare your pricing against this capture
• Set the cadence and watchlist for the next run

Accept 1, 2, 3, 4, 1 and 2, Verbose Mode, or a custom path.

Examples

A delta entry reading structure, not just price (fictional):

> ### [Competitor B] — API access moved from Pro to new Scale tier

> - Then / Now: API access included in Pro ($79/seat) → gated behind new "Scale" tier

> ($149/seat, 10-seat minimum) — Fact

> - Evidence: [pricing page vs. archived capture, URL, dates]

> - Reading: the price didn't change; the *gate* did. Developer-led accounts are being pushed

> into a sales conversation — Inference: monetization push on their fastest-growing usage

> pattern, and a window for us with API-first buyers while their community absorbs it.

The cross-competitor signal: three of four tracked competitors added usage-based add-ons this

quarter while holding seat prices flat — Inference: the market is testing consumption pricing

without headline price increases. Your pricing debate is no longer hypothetical; the Signals section

names the evidence.

See examples/sample.md for a complete worked delta run (fictional

FSM-software market) where both material moves are structural — a new tier and a relocated gate —

and no headline price changed at all. examples/sample-industrial.md

shows the fully opaque case: zero published prices, and structure as the only public signal.

Common Pitfalls

• Screenshot thinking. Capturing once and calling it competitive pricing intel. The value

compounds only on a cadence — baseline, then diff, forever.

• Number fixation. Reporting only price changes while a tier quietly gained a 5-seat minimum.

Structure is the leading indicator; watch the gates.

• Inventing the opaque. Filling "contact sales" cells with plausible figures. Publish the

honest boundary and label any inference — a battle card built on invented enterprise pricing

detonates in a live deal.

• Schema improvisation. Reordering or "improving" the capture format between runs breaks every

future diff. The schema is a contract with your own future runs.

• Tracking without a consumer. If no battle card, pricing review, or watch report consumes the

output, the cadence dies quietly. Wire the Final Step's first option into a real artifact.

References

• autonomous-investigation (Workflow) — the governing protocol
• intelligence-collection-disciplines (Component) — SIGINT sources: Wayback diffs, pricing-page monitoring
• competitive-intel-watch (Workflow) — the broader watch this tracker feeds
• battle-card-builder (Workflow) — consumes the pricing snapshot sections
• finance-based-pricing-advisor (Interactive) — sets *your* price; this skill informs it with market evidence
• Adapted from market-intelligence/pricing-packaging-tracker-prompt.md in the

https://github.com/deanpeters/product-manager-prompts repo.