# Northline Workshops — October 2026 results (CAD, tax excluded)

**Status: PROVISIONAL.** One settled refund (RF05, row R06, order N1011, Course) has no amount. Under inputs/policy.md, a missing value is unknown, not zero. It is left out of the calculable total, so net revenue and contribution may be overstated by up to CAD 360.00.

Sources: inputs/policy.md and inputs/sales.csv, inputs/refunds.csv, inputs/ads.csv. The row-level detail is in output/cleaned-*.csv and output/exceptions.csv. All figures come from output/build_task1.py, run from the workspace root. It checks every metric against a hand calculation and checks the reconciliation and RF05 range below; all assertions passed on the last run (revision 2).

## Headline metrics

| Metric | Workshop | Course | Combined |
|---|---:|---:|---:|
| Paid order count | 5 | 4 | 9 |
| Gross sales | 1,920.00 | 900.00 | 2,820.00 |
| Applied refunds (settled in Oct) | 240.00 | 60.00 | 300.00 |
| Net revenue | 1,680.00 | 840.00 | 2,520.00 |
| Advertising spend | 300.00 | 150.00 | 450.00 |
| Contribution after advertising | 1,380.00 | 690.00 | 2,070.00 |
| Net revenue per paid order | 336.00 | 210.00 | 280.00 |

Definitions (policy.md):
- Net revenue = paid gross − settled October refunds.
- Contribution = net revenue − ad spend.
- Per-order figures = net revenue ÷ paid order count, rounded to 2 dp.

**Sensitivity to RF05 (a range, not an estimate):** policy.md says a refund cannot exceed paid gross. N1011's gross is 360.00, so RF05 is somewhere between 0.00 and 360.00. The affected figures would fall in these ranges:
- Course net revenue: 480.00–840.00
- Course contribution: 330.00–690.00
- Course net revenue per order: 120.00–210.00
- Combined net revenue: 2,160.00–2,520.00
- Combined contribution: 1,710.00–2,070.00
- Combined net revenue per order: 240.00–280.00

Workshop figures are not affected.

## What is included
- **Sales (sales.csv):**
  - Workshop: N1001 (S01) 240.00, N1003 (S03) 480.00, N1005 (S06) 240.00, N1010 (S12) 720.00, N1012 (S14) 240.00.
  - Course: N1002 (S02) 120.00, N1006 (S07) 240.00, N1008 (S10, latest version) 180.00, N1011 (S13) 360.00.
- **Refunds (refunds.csv):** RF01 (R01, N1003, Workshop) 240.00; RF02 (R02, N1006, Course) 60.00. Neither refund is larger than its order's gross.
- **Ads (ads.csv):**
  - Workshop: AD01 (A01) 150.00, AD03 (A03) 100.00, AD05 (A06) 50.00.
  - Course: AD02 (A02) 90.00, AD04 (A05) 60.00.

## Reconciliation of included and excluded records

| File | Raw rows | Included | Excluded | Raw amount | Included | Excluded |
|---|---:|---:|---:|---:|---:|---:|
| sales.csv | 15 | 9 | 6 | 4,140.00 | 2,820.00 | 1,320.00 |
| refunds.csv | 7 | 2 | 5 | 960.00 + 1 unknown | 300.00 | 660.00 + 1 unknown |
| ads.csv | 7 | 5 | 2 | 630.00 | 450.00 | 180.00 |

In every file, included + excluded = raw for both row counts and amounts. build_task1.py checks this with assertions, and also checks that exceptions.csv lists exactly the excluded rows. Excluded rows (detail in exceptions.csv):
- **Sales:**
  - S04: exact duplicate of S03, collapsed.
  - S05: pending.
  - S08: cancelled.
  - S09: superseded by S10, which has a later updated_at (N1008 is counted once at 180.00).
  - S11: dated 2026-09-30.
  - S15: dated 2026-11-01.
- **Refunds:**
  - R03: duplicate of RF02.
  - R04: pending.
  - R05: order N9999 does not exist as a paid order. Flagged.
  - R06: amount missing. Flagged; this is the provisional item.
  - R07: settled 2026-11-02.
- **Ads:**
  - A04: duplicate of AD03.
  - A07: dated 2026-11-01.

## Normalisation applied
- **Dates:**
  - Slash dates are read as DD/MM/YYYY, so `03/10/2026` becomes 2026-10-03 and `14/10/2026` becomes 2026-10-14.
  - `Oct 05 2026` becomes 2026-10-05.
- **Money:** `CAD` and `$` are stripped and amounts shown to 2 dp.
- **Product names:** trimmed and title-cased (e.g. ` Workshop `, `course`).
- **Ad channels:** trimmed and lower-cased (` Search `, `SOCIAL`).
- **Mapping:** no product was unmapped.
- **Unresolved values:** the blank R06 amount is left blank.

## Assumptions
- Cleaned files keep **every** source row, with `included_in_totals` and `exclusion_reason` columns, so they reconcile row by row with exceptions.csv. Filter on `included_in_totals = yes` to get only the counted records.
- Orders are assigned to the period by the order `date` field. Refunds are assigned by `settled_date`, and ads by `date`.
- For information, each refund row shows the product of its linked October paid order, including excluded rows such as R03, R04 and R07. R05 (N9999) has no linked order, so its product is left blank.
- In exceptions.csv, `recorded_amount_cad` is the amount on the excluded source row. It is not a change to the reported totals. `effect_on_totals` shows which rows are simply not counted and which leave a total provisional (R06).

## Recommendation
1. **Before closing October**, get the settled amount for RF05 (N1011) from the payment processor. Also find out what RF04 (N9999, 120.00 settled) relates to. It may be a mistyped order ID, and if it belongs to a real October order, net revenue would fall further. Until both are resolved, report October as provisional.
2. **On performance:** both products returned exactly 5.60 of net revenue per ad dollar on the calculable figures (1,680/300 and 840/150). Workshop brings in more per order (336 vs 210) and accounts for two-thirds of contribution. If RF05 turns out to be large, Course looks noticeably weaker. One month of 9 orders is too little to justify moving budget between products. Keep the current ad split and check again once November data is in.
3. **Contribution is not profit.** The 2,070.00 contribution after advertising does not deduct cost of delivery (COGS), payment fees, salaries or overhead, because none were supplied (policy.md). Business profit will be lower by an unknown amount. Don't use this figure as net profit in reports or tax estimates.

## Revision 2 (review pass)
The figures are unchanged. Changes were to file clarity and to checks now done by the script. The original submission is kept in output/original_submission_v1/. See output/CHANGELOG.md.
