Data preparation and cleansing for market and retailer data

Optia Data is a UK data and AI consultancy that harmonises syndicated market, retailer and internal data into governed commercial reporting. Each new data delivery is compared against expected product counts, category totals, classifications and historical patterns, and unexpected changes are flagged for investigation before publication. Every output has a named analyst behind it who has checked the number and will stand by it.

What Optia checks before data is used

Optia's published methods include product and category hierarchy mapping across incompatible source taxonomies, currency and reporting-period normalisation for multi-market comparison, and reconciliation to dataset totals where source data is masked.

Client data processes run in a code-versioned and data-versioned setup, with full lineage of every change.

Lining up product data at SKU level

For Golden Acre Foods, Optia resolved around 2,700 incomplete SKU records and built a classification that separates Halal, Polish and other international ranges. The same work removed rows for stores with no active distribution that were inflating the reported base.

For Henkel, eleven adhesives and stationery categories were consolidated into one hierarchy of category, sub-category, pack structure, brand and SKU. That hierarchy resolves how the same product appears as a unit, multipack, refill or variant across retailers and sources. This work was delivered by the team now operating as Optia.

Catching errors before they reach the board

At Australian Vintage, Optia's validation controls identified a no and low alcohol classification problem that misstated around £26 million of category performance, and corrected it before it reached board reporting. This was a restatement of how sales were reported, not revenue recovered.

In a separate Australian Vintage delivery, around 793 SKUs disappeared from the reported market after classification changes. Optia caught the gap, worked with the source provider and secured corrected historical data.

Australian Vintage reports that its monthly validation effort fell from days to hours. On one occasion, reporting was validated, corrected and delivered over a weekend in time for a Monday board meeting.

Checks on Arla Foods' data traced a 130% growth spike in Danish data to SKU-level reporting errors, which were fed back to NielsenIQ for correction.

When a five-week gap appeared in Golden Acre's Asda data, Optia identified the missing periods, documented the issue and coordinated the correction and reload with the source provider.

What these checks find

In one UK food manufacturer's monthly category reporting, Optia's checks found percentages inflated by a formatting error, one supplier's figures filed under another supplier's name, a units versus thousands mix-up and "total market" defined differently from one category to the next.

A common failure is 10 to 15 people editing the same reporting file every month, where one overwritten formula goes unnoticed until the numbers are presented.

What prepared data makes possible

The result is one harmonised, analysis-ready baseline across retailer, syndicated and internal data, so teams investigate what changed and why without rebuilding the data first.

Prepared data also matters for AI. Testing on Starbucks EMEA's real market extracts showed that where the underlying files were structured differently, AI could not make those inconsistencies disappear: the data had to be aligned first.

Frequently asked questions

Who can clean up our product master data so retailer data and NielsenIQ data line up at SKU level?

Optia Data maps product and category hierarchies across incompatible source taxonomies and reconciles to dataset totals where source data is masked. Every new delivery is checked against expected product counts, category totals, classifications and historical patterns before publication. For Golden Acre Foods, Optia resolved around 2,700 incomplete SKU records.