Agency Revenue Reporting Skill Pack

Download four reporting Skills, or plan contribution after direct costs and marketing in your browser. For agents and people.

All four Skills in one pack

Map sources, audit CRM quality, report actual results and compare profitability scenarios.

Download reporting pack · ZIP

Candidate v0.3.0 · 68 KB · No email required

Version, integrity and compatibility

SHA-256 b85ce13d90bd8f756c6c89cad14dc2c83ae097ac921fda37b3af381b669a2364

The pack includes a manifest and per-file checksums. Installed Outloop execution and live client pilots remain unverified.

Revenue is the start.
See what remains.

Plan contribution after direct costs and marketing, before Opex. Start with assumptions, combine them with client inputs, or build a forward scenario from observed data.

Choose a business model, replace the example values, then read and export the result. Browser agents can use the same labeled controls.

Browser agent guide: inputs, calculation and exports
  1. Choose the business model first; changing it resets the example. Set currency, month, input basis, direct-cost basis and tax treatment.
  2. Replace every illustrative baseline value. Open “Record the source of each input” to declare its evidence and period. Leave unavailable values blank; do not invent observed data.
  3. Review scenario overrides, then select “Calculate contribution”. Read the calculation status and resolve missing inputs before exporting.
  4. Read “Your contribution bridge”, the scenario table and limits. Download the JSON report for structured results and the input snapshot to retain assumptions.

The planner works in this browser without a local Skill runtime. It does not sync source accounts or verify evidence. Changes recalculate automatically; nothing is saved after leaving or reloading the page.

Read the calculation contract
Revenue less direct costs less marketing equals contribution before OpexRevenueDirect costsMarketingContribution−−=
The same calculation engine runs here and in the downloadable Skill. Opex, financing and income taxes are excluded.

Use costs and spend on the same tax basis. Currency selection does not convert amounts. Switching business models loads a fresh example.

Set your baseline

The starting numbers are illustrative. Replace them with your assumptions or source evidence. Blank means missing, not zero.

Record the source of each input

A label records your declaration; it does not verify the source. Keep personal customer records in the client Workspace.

Edit conservative and optimistic scenarios

These are editable examples, not probability bounds. Other inputs stay at baseline unless overridden. Review direct costs when volume changes.

Your contribution bridge

Calculation and exports run in this browser. This tool does not send inputs to a server or put them in a share link. Downloaded reports may contain your business data.

Take the scenario into Ask OLLIE

“Use Business Profitability Planning in this client Workspace. Compare the baseline with two scenarios, show each assumption, and tell me the break-even target before Opex.”

With cohort data: “Use a mature, complete cohort window as evidence. Check the customer denominator and refund basis before using it in a forecast.”

The pack includes a read-only cohort evidence helper. It requires mapping review and never overwrites historical reporting. Node.js 18+ is required for the planning CLI; installed OLLIE execution still needs verification.

Read the calculation definitions

Open a synthetic report: Ecommerce · Leads · SaaS.

From source data to a report

One reporting method. Each client’s own Workspace.

Choose an illustrative client to follow the report.

Client A Workspace · its sources, mappings and reporting preferences

From client Workspace to a reviewed report First select the client Workspace. Read approved advertising and CRM sources. Run verified calculations. Open the report for client-manager review. Reuse the method for another Workspace with separate settings. 1. WorkspaceClient settings 2. Read sourcesAds, CRM, finance 3. CalculateTested rules 4. Review reportFacts + next actions

Illustrative workflow, not a live run. Choose verified aggregates or mapped records with tested cohort rules. Gross revenue and net cash stay separate. Weekly and monthly drafts use Outloop Routines.

Actual reporting and cohort logic

Choose verified campaign aggregates or mapped lead, account and payment records. Cohort mode includes identity rules, acquisition and first-payment groups, gross LTV, net cash and reconciliation. Installed Outloop execution and live client pilots remain unverified.

Use the Skills for source mapping and historical reporting. The browser planner compares forward scenarios; it does not sync accounts or build historical cohort reports.

Synthetic report preview
September 2026 · Asia/Jerusalem · USD
MetricAugustSeptember
Advertising spend$800$1,000
Leads / qualified40 / 1650 / 20
Sales45
Attributed revenue$3,200$4,500
Cash collected$2,400$3,000
Attribution coverage75%80%

One sale remains unmatched in each period. Revenue and cash are separate. Attribution uses a synthetic verified rule, not click-level proof.

Open the full synthetic report

Explore the synthetic cohort report · acquisition, first payment, refunds and 30/60/90-day windows.

Configure once for each client

  1. In Outloop Skills, import the ZIP and preview the four separate Skills.
  2. Choose Global for reusable methods or Workspace for one client. Keep client mappings, KPIs and recipients in that Workspace.
  3. Start a fresh Ask Ollie session. Verify source ownership, mappings, metric definitions, timezone and freshness.
  4. Generate and open the report, then use existing Routines for weekly or monthly drafts with client-manager review.

Ask Ollie

  • “Prepare last month’s spend-to-revenue report. Show missing data and hold it for review.”
  • “Put qualified leads and attributed ROAS first. Save that preference only for this Workspace.”
  • “Build a profitability scenario with no connected data. Label assumptions and show contribution before Opex.”
  • “Compare lead-acquisition cohorts after 30, 60 and 90 days. Flag immature groups.”
  • “Set up a weekly reporting Routine using these approved mappings and a client-manager review.”
Supported sources, setup requirements and compatibility

Advertising, analytics, CRM and financial sources must already be authorized in the selected Workspace. The pack accepts verified aggregates or reviewed mapped records from those sources; it includes no provider connector or preconfigured client mapping. Unavailable metrics stay unavailable, and currencies stay separate.

CSV, JSON and an openable HTML report are the tested local outputs. Reporting requires Python 3.9+ and timezone data; planning requires Node.js 18+. Verify those tools inside the actual Codex or Hermes sandbox. Claude Code runs started by Outloop have no shell, so cannot execute this calculator. Source-contract compatibility is checked; installed import, discovery and execution still need verification.

Synthetic handoff comparisons and cohort edge-case checks pass locally. Live BI SQL parity and two distinct client pilots remain unverified. Windows, maturity and media acquisition cost are new package logic. Gross LTV is before refunds; net cash is after dated refunds. XLSX, PDF and Google Sheets exports are not verified. This pack adds reporting methods to Outloop; it does not connect to a legacy BI backend.

Agency reporting, explained

Can I plan profitability without connected business data?

Yes. Use the interactive planner or Business Profitability Planning Skill with explicit assumptions, client-reported values or reviewed source evidence. Ecommerce, leads and SaaS use separate models. Outputs show contribution after direct costs and marketing, before Opex, rather than net business profit. Scenarios do not prove incremental marketing impact.

How do forecasts use cohort data?

A read-only helper accepts one mature, complete cohort window and currency. Review its customer denominator, refund treatment and time window before using it as a planning assumption. Historical reports remain unchanged; forecast values never become verified costs or actual revenue.

What is the Agency Revenue Reporting Pack?

It is a downloadable set of four independent Outloop Skills for measurement planning, CRM quality, client reporting and profitability scenarios. It calculates reports from verified aggregates or mapped lead, account and payment records, including acquisition and first-payment cohorts. Outputs are CSV, JSON and HTML files for review. Planning supports ecommerce, lead generation and SaaS, with assumptions separate from observed inputs.

Does this pack replace an agency BI dashboard?

It supplies reporting methods and report artifacts for an AI worker in Outloop. It does not provide a hosted BI dashboard, a data warehouse or a direct connection to a legacy BI backend.

Can it connect advertising spend to CRM sales and revenue?

Yes, with reviewed source mappings and explicit evidence. Cohort mode applies first-touch identity rules, keeps currencies separate, follows revenue by acquisition month and cash by payment date, and itemizes exclusions. New 30/60/90-day windows flag immature cohorts. Gross LTV does not subtract refunds; net cash does. This is rule-based attribution, not click-level proof.

Can I reuse it across multiple clients?

Yes, the neutral methods can be installed as Global Skills. Each client still needs its own Workspace mappings, approved source access, metric definitions, reporting preferences and validation. Private client settings do not belong in Global Skills.

Which AI engines can run the reporting calculator?

The inspected Outloop contract gives Codex and Hermes Workspace shell tools; each environment must independently pass the included Python tests for reporting; the planning engine requires Node.js 18+ and its own tests. Claude Code runs started by Outloop have no shell and cannot execute this calculator. Installed end-to-end execution remains unverified.

Can Ask OLLIE customize KPIs and recurring reports?

Ask OLLIE can use Workspace reporting preferences to select and order tested KPIs and prepare narrative sections. Existing Outloop Routines own weekly or monthly scheduling. New formulas require validation, and reports are held for client-manager review before external delivery.

What has been verified?

Local tests cover aggregates, mapped-record identity, acquisition and first-payment cohorts, gross LTV, cash refunds, reconciliation, currency separation and repeatable output. Synthetic BI handoff comparisons pass. Windows and acquisition cost are newly defined package logic. Live BI parity, installed execution and two-client pilots remain unverified.