One daily report across every SAP system worldwide.

Where the customer stood

The company had just been carved out of a larger group and could no longer call on the parent's IT and BI teams. That left finance and sales without reliable reporting. The data sat in several SAP systems at sites around the world, each naming products and customers differently. Other consultancies had already tried to connect them.

Until then, the team had got by with raw data exported to Excel by hand. It was laborious and still gave only a static picture. Every site ran its own SAP system.

The goal was a single view across all of them for the first time, right down to individual orders. It had to refresh itself every day and need little upkeep, so that it would last.

What we did

We started by finding the key metrics in every SAP system and agreeing a common definition for each. Clustering algorithms brought product and customer names into line. On that basis we built reporting that refreshes itself every day. Management gets a finance dashboard with revenue, order intake, gross profit and order backlog. Sales gets a pricing workbook that goes down to individual orders. Workshops introduced staff to both.

The logic behind the scenes differed from one SAP system to the next. We had to work through it site by site before any figure could hold true across all of them.

The pricing workbook lets sales answer their questions in a few steps. The filter bar folds away, and a presentation mode helps in meetings. Notes explain outliers in the data, and everyone can save their own views as a default.

What came out of it

Reporting now refreshes itself every day across all sites. Where whole teams used to pull data together, the figure is simply there in the morning. Prices can be compared worldwide down to the individual order.

After the workshops, staff built their own analyses. Because the data refreshes itself, upkeep stays low. The project later became the basis of our Sales Cube package for companies running several SAP systems.

What proved harder than expected

The technology wasn't the hardest part. Agreeing what each metric meant at every site was harder, and only then could the systems be connected.