- winner if posterior probability > 50%, unsuccessful otherwise
- Removes inconclusive — binary outcome based on Bayesian signal
- Label "Loser" renamed to "Unsuccessful"
- Updated all stored experiment verdicts in DB
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
FR/UK/DE sheets have a 'Status' column that appears before 'Reason',
causing the dynamic detector to pick the wrong column.
Use confirmed indices: FR=col S(18), UK=col J(9), DE=col N(13).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Inventory sheets like 'INV ITALY' were being incorrectly matched as
buybox sheets, adding all their ASINs as false BB lost entries.
Now only sheets starting with 'BB' are processed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace fragile header/column-index approach with pattern-based ASIN detection:
- Scan all cells for Amazon ASIN pattern (B[0-9A-Z]{9}) to auto-detect the ASIN column regardless of headers or position
- Detect country from sheet name with flexible matching (no hardcoded sheet name list)
- Detect reason column from header keywords, no hardcoded column indices
- Remove all hardcoded fallback indices (col 13, 18, 9, etc.) that caused empty results
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Force reasonIdx to column N (13) for BB_ES sheet, same as DE/FR/UK, so dynamic header detection can't override to a wrong column
- Fall back to 'BB Lost' when reason cell is empty instead of skipping the ASIN entirely
- Add B095K8948Z to debug logging alongside B0D3874WSD
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The previous fix only handled the unambiguous case (day > 12). Dates like
"5/1/26" or "12/1/26" (Jan 5 / Jan 12, 2026) were falling through to
JS's new Date() which parsed them as US format MM/DD/YY, assigning them
to May and December instead of January.
Now any 3-part slash date with a 2-digit year in position 3 and a valid
month in position 2 is treated as DD/M/YY (Spanish/EU convention),
fixing weeks 2-3 of January (days 5-18) showing zero sell-out.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add support for European day-first date format (e.g. "23/2/26" = 23 Feb 2026)
in the sell-out CSV pipeline so months and years are correctly extracted from
column C of Amazon Sell Out 2023-2025.csv.
- normalizeMonth: detect DD/MM/YY when first part > 12, return "Mon-YY"
- mapRowToRecord: add 'C', 'Date', 'Fecha', 'DATA' to month column aliases
- validateSellOutHeaders: accept date columns as substitute for YEAR + MONTH
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add date? field to BSRRecord (ISO string YYYY-MM-DD)
- processBSRExcel now extracts and stores the Date column as an ISO date
(handles both Excel serial numbers and string dates)
- VendorDataView groups chart data by date (daily) when dates are available,
falling back to weekly buckets — eliminates straight-line charts caused
by only 2 weekly data points
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Fix critical regex bug: /\\d+/ was matching literal '\d' instead of
digits, causing all rows to be rejected (week always = 0)
- Add VendorCSV format support: 'Top Level Category (Rank)', 'Date', etc.
- Derive ISO week number from Date column when no Week column exists
- Process all sheets (not just first) to handle multi-tab workbooks
- Log column names on load to aid future debugging
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
VendorDataView was refactored to receive BSR data as a prop but App.tsx
was not updated to pass it, causing the component to crash on undefined
and show a blank screen. Also fixes supabase.ts type errors and removes
dead code.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Change getWeekStartSunday to getISOWeekMonday for timestamp calculation
- Ensures week alignment between data aggregation and ISO week range calculation
- Fixes incorrect treatment_after average calculation
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Add ISO week calculation functions (getISOWeek, getISOWeekMonday, calculateISOWeekRange)
- Expand experiment dates to full ISO weeks (Monday-Sunday) instead of exact date ranges
- Use actual data.length for averaging metrics instead of theoretical durationWeeks
- Baseline now takes same number of complete weeks immediately before experiment
Fixes incorrect units sold average calculation (was 6.5, now correctly shows 4.33 for 13 units over 3 weeks)
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
The generic regex /sale/i in ads column detection was matching "Advertising Cost
of Sales" (the ACOS % column) before reaching "7 day total sales" (the actual
attributed sales amount). This caused attributedSales30d to receive the ACOS
percentage instead of the real sales value, making experiment ACOS = 0 or 100%.
- Replace /sale/i with specific regexes: /\d+\s*day.*sale/i, /total\s+sale/i, etc.
- Replace /cost/i with regexes that exclude "cost of sales" columns
- Replace /unit/i with specific regexes for unit columns
- Add diagnostic logging for ads Excel column headers and first-row values
- Revert ACOS 100% fallback (root cause now fixed)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Change ?? to || for salesAds/units/revenue in weekly aggregation to guard against NaN
- Add || 0 fallback to ads-only records in mergeSalesAndAdsData
- Fix parseCurrency to check isNaN on all EU-format return paths
- Handle ACOS edge case: cost > 0 with adRevenue = 0 now returns 100% instead of 0%
- Add diagnostic console logging to computeDiD for data flow tracing
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Allow pasting multiple ASINs separated by spaces in any MultiSelectDropdown
search field and pressing Enter to auto-select all matching options.
Also switch ads CSV/Excel parsing to header-based column mapping for
robustness across different file formats.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace the basic CRUD experiment tracker with a scientifically rigorous A/B testing system:
- Add DiD analysis engine (services/experimentAnalysis.ts) that computes treatment vs control
group comparisons across 7 metrics (units, sessions, CVR, CTR, ROAS, revenue, ACOS)
- Implement Bayesian verdict system (Winner/Loser/Inconclusive) using posterior probability
with normal CDF approximation (Abramowitz & Stegun erf, no external deps)
- Build counterfactual time series for trend charts (actual vs estimated without change)
- Rewrite ExperimentsView as single component with 3 inline sub-views (list, detail, create)
replacing the previous modal-based ExperimentDetail and ExperimentForm
- Add control group support, change annotations (before→after diffs), and SEO experiment type
- New types: ExperimentChangeAnnotation, DiDMetricResult, DifferenceInDifferencesResult,
ExperimentVerdict
- Simplify App.tsx by removing experiment modal state (5 useState hooks eliminated)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
