Symptoms and causes

The request usually sounds like a need for a nice dashboard. Almost always one of the problems below sits behind it, and that is what needs solving.

Departmental reports do not match

We introduce single definitions for the metrics. The cause is that one metric is calculated by different rules.

The same counterparty appears three times in the database

We clean the reference data and set the rules. Duplicates appear where nothing is checked when a record is created.

The report takes three days to prepare

We build a warehouse that refreshes automatically. Assembling it by hand from spreadsheet exports eats days of an analyst time.

The data is out of date by the time of the meeting

We set up scheduled data loading. A manual monthly export cannot keep up with the decisions being made.

Accounting is drowning in source documents

We connect document recognition. Details and amounts are typed in from paper by hand, line by line.

A dashboard built on dirty data is more dangerous than no dashboard at all. While a report is compiled by hand, mistakes get noticed along the way. A polished automatic chart looks credible, and decisions are made from it without a second look.

How we build the analytics

We start with one department and one group of metrics. Building a warehouse for the whole company at once takes long and almost always ends in nothing.

01

Agreeing the metrics

We agree what counts as revenue, a shipment and an active client. The dullest and the most important stage.

02

Audit of the data sources

We see where the data sits and what condition it is in, and estimate how much cleaning is needed before loading.

03

Data loading and the first mart

We set up regular data loading and build the first mart around the specific questions the manager asks.

04

Expansion

We connect the next sources and departments. Each one costs less than the last because the foundation is already there.

Frequently asked questions

If all the data lives in one database and nobody questions it, then yes. A separate warehouse is needed when there are several sources: accounting, the website, the warehouse, advertising. Merging them inside 1C is awkward, and heavy analytical queries against the live database slow users down.

With the data in reasonable shape, three to four weeks including agreeing the metrics. If the reference data is a mess, most of the time goes into cleaning and the timeline shifts. That is exactly why we look at the data first and quote timelines afterwards.

For standard printed forms accuracy is high; for handwritten notes and poor scans it is lower. That is why the workflow always includes human confirmation: the system fills the fields and an accountant checks and posts them. The time saving holds even with partial checking.

We do, as part of the support service: we watch the data loads, react to changes in the sources and add new metrics on request. Alternatively we hand it over to your specialists together with the documentation; the arrangement is agreed at the start.

We will sort out your reporting

Tell us which reports are compiled by hand and where the numbers disagree. We will review the sources and propose the first data mart.