Rowan & Pike

A retailer with 140 branches and eleven spreadsheets. We rebuilt its sales data model and cut weekly reporting from three days to ten minutes.

Retail

Retail analysts reviewing printed sales charts
10 min
Weekly reporting, down from three days
140
Branches on one trusted data model
£310k
Stock write-offs avoided in year one

The challenge

Rowan & Pike ran 140 branches, and every Monday a team of three spent most of the week stitching eleven spreadsheets into a sales report. By the time it reached the board, it was out of date, and two directors regularly quoted different revenue figures.

What we did

We started with a two-week discovery sprint, interviewing branch managers, finance and the buying team. We agreed one definition of revenue, margin and stock cover, and wrote them down.

Then we built a tested data model in BigQuery with dbt, loading till, stock and supplier data every night. A single Power BI report replaced the spreadsheets, with a one-page summary written in plain English each Monday.

The result

Weekly reporting now takes ten minutes of review instead of three days of assembly. Buyers spotted slow-moving lines earlier, which avoided an estimated £310,000 of write-offs in the first year. The board now argues about decisions, not about whose number is right.

Rowan & Pike is a fictional client created for this demo site. All figures are illustrative.

Next case study

Elmfield Clinics

Missed appointments were costing a clinic group hundreds of thousands a year. Forecasting and reminders brought the rate down by a third.

Hire us for your next data project · Est. 2014 · Hire us for your next

Any dataset? Any question? Any deadline?
Our analysts will help you out.