What is included

We build from business questions: first the list of reports that are needed, then the data model to serve them.

Discuss the scope

Gathering requirements

What questions the report must answer and who will use it.

Sources

The accounting system, the website, the CRM, spreadsheets. We decide what comes from where.

Data model

A single structure in which products and counterparties are reconciled across systems.

Loading

Scheduled refreshes that do not load the production systems during working hours.

Marts

Prepared datasets for specific reports so the analytics runs fast.

Connecting reporting

Dashboards in a visualisation tool or an export in the format people are used to.

How the work runs

The first mart takes four to six weeks. After that new ones are added faster on the ready foundation.

01

Priorities

We pick one or two reports that are currently assembled by hand and take the longest.

02

Design

The model, the mapping of reference data between systems and the refresh schedule.

03

Implementation

Loading, the mart, and a check that it matches the manual calculation.

04

Growth

New marts as requests come in, built on the existing foundation.

The first test of a warehouse is whether it matches the manual report. If the figures differ, nobody will trust the system and people go back to their spreadsheets. So reconciliation against the existing calculation is mandatory, and any differences are resolved before go-live.

Frequently asked questions

Reports in an accounting system show that system data. As soon as you need to combine sales from 1C with website traffic and leads from the CRM, the built-in tools fall short. On top of that, heavy reports on a production database slow users down.

Not necessarily at the start. The first marts can be exported to spreadsheets, which is familiar and quick. A dedicated tool is justified when many people use the reports and interactivity is needed.

For most management reports once a day overnight is enough. Real-time refresh is rarely required, costs more and loads the systems. It is best to start with a daily cycle.

We will build your data warehouse

Tell us which reports you assemble by hand and from which systems. We will start with one mart and check that the numbers match.