The Data Warehouse in telecommunications

2012 — 2014 at Seven Sky (Iskratelecom, CJSC)

Project role — Solution architect + PM

About the company

Company "Seven Sky" was established in 2008 by merging of the two local broadband telecom operators in the Moscow region (Russia) with joint coverage of 30 thousands of households. In 2012 we had over 150 thousands of the Internet and IPTV subscribers in Moscow and Moscow region.

The strategy was to expand further through acquisitions. For this reason the IT landscape was dynamic.

Project goals

The goals of the data warehouse project were:

  1. Ensure a complete match in the reporting of different levels.
  2. Ensure unchangeable financial reporting for the past periods and auditability of changes, happened due to the late data or error fixing.
  3. Ensure due diligence KPI monitoring (count of subscribers, churn, growth, revenue) at all the stages of an acquisition, including retrospective analysis independent from the current state of the legacy IT systems.
  4. Ensure contradiction discovery in the data of the different IT systems, including systems of the acquired businesses.

The progress and results of the project

In 2012 the demand had been identified. At the business level, the analytical department was founded, it included IT and a marketing team. That way the small IT team under my leadership was reoriented to the data warehouse and reporting, the previous project was reprioritized.

For a platform, we had picked Microsoft SQL Server because of our good expertise in its database engine and the basic expertise in its reporting and integration services.

In 2012 we were solving the business tasks with the tools described above but without proper methodology. Nevertheless, the current business demand for the reporting and analysis had been covered.

We raised our own level and decided to use the principles of Dimensional Modelling (R. Kimball), also to divide the storage into the primary loading zone (Staging), normalized storage (NDS/ODS) and storefronts (DDS).

In 2013 we had built our ETL practices, which helped us to solve performance issues, to obtain consistency of execution for independently implemented ETL processes, to solve specific source issues, to solve the issues of coordination for changes and regeneration of storefronts. We had expanded reporting, we built cubes (Analysis Services) and trained business units to use them.

In spring 2014 all the requirements for retrospective analysis (this became relevant after replacing of the billing system), for auditability of changes, for the data quality, for the drill-down (from any top-level indicators to individual events) were completely covered.

← Back to portfolio