DEMONSTRATION CASE · SYNTHETIC DATA
Turn a blocked product CSV into a review-ready plan.
This transparent example shows the useful middle ground between “the file opens” and “import everything”: scan locally, apply only deterministic repairs, and isolate decisions that need a human.
Original 24-row workflow illustration below. The separate six-row downloadable report was checked on 23 October 2026. Go to the new report sample
THE INPUT
A realistic mix of technical faults and business ambiguity.
The demonstration file contains 24 invented product and variant rows. It combines fixable formatting problems with missing values that should never be guessed: duplicated option combinations, text inside price or stock fields, a row with no usable product identity, and image URLs that still need a live reachability test.
Scan locally
Parse the file in the browser and classify blockers, warnings, and notes without uploading row data.
Separate safe repairs
Allow deterministic cleanup such as recognized header mapping, blank-row removal, or trimming structured values.
Isolate decisions
Leave SKU, handles, missing titles, prices, relationships, and ambiguous option values for explicit human review.
Prepare a test import
Deliver a new working copy, issue report, exception list, and a representative subset for Shopify preview.
BEFORE
The starting file hides unlike problems together.
- Duplicate handle and option-value combinations
- Currency symbols and text in numeric fields
- A row without a usable Title or URL handle
- Image links not yet tested from Shopify’s side
AFTER
Each output has one clear job.
- Untouched source retained for rollback
- CSV issue report with row, field, severity, and suggested action
- New working copy containing only approved deterministic repairs
- Exception list and representative import checklist for final review
$49 REVIEW FORMAT · SYNTHETIC SAMPLE · 23 OCT 2026
Five file findings. Two decisions a scanner cannot make.
This separate six-row mapping example shows what an annotated risk report looks like. It is not the 24-row workflow illustration above, a paid customer report, or evidence of a Shopify import. All source values and buyer rules are invented and available for inspection.
Input checked: shopify-held-candidate.csv — six data rows, 12 columns. The local result is blocked: two errors and three warnings.
Additional review: two source-based business questions. These are human findings, with no invented scanner code.
Priority: P1 means resolve or keep the affected row on hold before import. It is a work priority; the scanner severity is shown separately.
Row numbers: CSV records including the header as row 1. Source and candidate retain the same row order; these files have no multiline cells.
R01 · P1 · Local error
INVALID_PRICE
shopify-held-candidate.csv · row 2 · Price
- Evidence
- Price is $19.90, not a plain decimal.
- Risk
- The local numeric rule blocks this value.
- Next action
- Use 19.90 only under the explicit USD selling-price rule in demo-rules.json.
- Buyer question
- Does this amount mean selling price in the target store currency?
- Recheck / remaining hold
- Reviewed row 2 has Price=19.90; the local INVALID_PRICE finding is absent.
R02 · P1 · Local error
INVALID_INVENTORY
shopify-held-candidate.csv · row 3 · Inventory quantity
- Evidence
- Inventory quantity is 6 pcs, not an integer.
- Risk
- The local inventory rule blocks this value.
- Next action
- Use 6 only under the confirmed single-item count rule; keep multi-location inventory outside this example.
- Buyer question
- Are these single-item counts for one location, rather than cases or available-to-promise quantities?
- Recheck / remaining hold
- Reviewed row 3 has Inventory quantity=6; the local INVALID_INVENTORY finding is absent.
R03 · P1 · Local warning
DUPLICATE_VARIANT_KEY
shopify-held-candidate.csv · row 4 · Handle / options
- Evidence
- dp-demo-tee + Size=S repeats row 2 because the mapping omitted Color.
- Risk
- Different SKUs do not make the represented option combination unique.
- Next action
- Represent the confirmed Color and Size options; do not delete a row to hide the collision.
- Buyer question
- Are Red S and Blue S distinct variants of the same product?
- Recheck / remaining hold
- Reviewed rows 2 and 4 have distinct Color+Size keys: Red/S and Blue/S.
R04 · P1 · Local warning
DUPLICATE_VARIANT_KEY
shopify-held-candidate.csv · row 5 · Handle / options
- Evidence
- dp-demo-tee + Size=M repeats row 3 because the mapping omitted Color.
- Risk
- The same mapping omission also affects medium variants.
- Next action
- Apply the approved two-option mapping consistently across the product group.
- Buyer question
- Are Red M and Blue M separate variants, using the same option names?
- Recheck / remaining hold
- Reviewed rows 3 and 5 have distinct Color+Size keys: Red/M and Blue/M.
R05 · P1 · Local warning
DUPLICATE_VARIANT_KEY
shopify-held-candidate.csv · row 6 · Handle / options
- Evidence
- dp-demo-tee + Size=S repeats row 2; the Green row also has an unresolved selling price in the source.
- Risk
- Correcting option columns alone would not approve this row's selling price.
- Next action
- Keep this record on hold for R06 instead of exporting it as a reviewed variant.
- Buyer question
- What selling price is approved for DP-GREEN-S?
- Recheck / remaining hold
- DP-GREEN-S remains in the source and is absent from the four-row reviewed candidate.
R06 · P1 · Human review
Cost is not an approved selling price
supplier-source.csv · row 6 · amount_kind / amount
- Evidence
- DP-GREEN-S has amount_kind=cost and amount=8.20, while the held target copied 8.20 into Price.
- Risk
- 8.20 is syntactically valid, so the local scanner cannot know it is a supplier cost rather than an approved selling price.
- Next action
- Leave the row on hold. Do not add a margin, fill zero, or assume cost equals selling price.
- Buyer question
- Which selling price and currency does the buyer approve?
- Recheck / remaining hold
- Still unresolved: source row 6 is held; no invented selling price is present in the reviewed candidate.
R07 · P1 · Human review
The product group is unknown
supplier-source.csv · row 7 · group_code
- Evidence
- DP-BLACK-M has a blank group_code. Its held target has a Title but no URL handle.
- Risk
- A title prevents the local missing-identity rule from firing; it does not establish the intended product or variant relationship.
- Next action
- Leave the row on hold. Do not infer its parent from the SKU, name, color, or neighboring row.
- Buyer question
- Is this a separate new product, a variant of TEE, or an update to an existing product?
- Recheck / remaining hold
- Still unresolved: source row 7 is held; no guessed handle is present in the reviewed candidate.
WHAT A $49 REVIEW ADDS
A human-readable risk map before cleanup begins.
- One bounded file, up to 10,000 rows
- Annotated issue report and priority order
- Written repair plan and questions that require business context
- Separate cleanup quote only when requested
YOUR FILE, SAME BOUNDED START
Start with evidence, not an open-ended cleanup promise.
Describe the problem without uploading private data. The $49 review scope, timing, and secure handoff are confirmed before any project file moves.