Consumer Goods & Retail20 days of report assembly, down to hours — across 17 countries
Nielsen sales data arrived from 17 countries in 17 different shapes: product names spelled differently in each market, three currencies to reconcile, and hierarchies that had to be picked apart before anything could be compared. Assembling one static report took 20 days, by which time the decision it informed had usually been made.
- 17
- countries harmonised into one format
- 1B+
- rows processed monthly
- 20 days → hours
- to a usable sales picture
The workflow
What replaced the manual work
- Nielsen data, 17 countries
- Mixed formats, naming and currencies
One pipeline
- Connect
- Clean
- Automate
- Share
Recorded once, re-run on a schedule.
- One standardised sales dataset, refreshed monthly
- Power BI through BigQuery
Outcomes
What changed
- Report assembly fell from 20 days to hours, so insight arrives while the decision is still open
- Over a billion rows a month processed and standardised without a dedicated data engineer
- 17 countries reconciled into one format, including product naming and three currencies
- Business rules visible and editable by the analysts who own them, rather than buried in IT
- The reporting layer separated from the source, so a vendor change no longer means a rebuild
What they used
The parts of Mammoth involved
- Data Engine
- Scheduled refresh
- Automated validation
- Role-based access
“Going from weeks to hours for reports changed everything for our decision-making.”
Start with one source and one question.
That's how this deployment began too. Twenty minutes on the free plan is a fair test.
- 21-day Pro trial
- No credit card
- Viewers always free