You are here

Why the data warehouse plays a key role in your CPM architecture

As Corporate Performance Management initiatives are initiated, companies should reflect on how their CPM environment will interact with the (existing) data warehouse infrastructure so that both can serve as seamless input and output to each other.
 
Many companies have deployed data warehousing and business intelligence technology over recent years in order to better understand the historical performance of their business. Often these initiatives were driven by departments trying to get a grip on the information needed to monitor, analyse and drive the performance of their own department. Research shows that over time most business intelligence projects also have been able to successfully combine data, information and insights from different departments to come to cross-functional analysis like Supply Chain Analytics integrating data from Purchasing, Production, Inventory, Distribution and Sales or Customer Analytics integrating data from Sales and Marketing. As such these projects now deliver more company wide insights.
 
Sales and Finance departments typically pioneered in introducing business intelligence in a company. Despite that, financial data is today often the least integrated data source in the corporate data warehouse, thus not allowing to completely link the operational insights with the financial impact. This is not only the case for the historical financial performance data -in its internal view as management reporting-, but even more so for consolidated financial data -the external view- and the data on future performance in the areas of planning, budgeting and forecasting, key drivers for the business going forward. As such, financial performance measurement -and especially consolidation, forecasting and planning- often still happens in a 'silo', not integrated in the corporate information architecture.
 
In recent years, the Office of Finance has gained an interest in Corporate Performance Management. This has been the result of a number of factors :
  • CFOs are becoming more business partners, taking up the responsibility to deliver the insights rather then just some financial, legally required, data.
  • Scorecarding is coming to maturity as a way of measuring and driving performance. It can be expected that also Activity Based Costing/Management will re-gain in popularity in the years to come.
  • Increased regulation from the compliance perspective requires complete access to both consolidated and detailed data ('drill down').
  • Most companies urgently needed to upgrade their consolidation methods, processes and technology and decided to broaden the scope of these projects to a broader CPM initiative.
  • More than ever -especially for quoted companies-  correctly managing market expectations by accurate forecasting and planning drives stock performance more than the actual company results themselves.
As your company embarks on a CPM initiative, make sure you reflect on how the information flow and processes will be defined between your CPM applications and your corporate data warehouse that is in place, so no new silos are created or work (eg. data-cleansing or hierarchy structure creation) is lost or needs to be redone. In case no corporate data warehouse is in place, your CPM initiative might also be a good starting point for the setup of a data warehouse initiative.

Some of the key goals of CPM are:
  • Quicker and more accurate information and where needed with the detail of the aggregated data readily available
  • 'Single' version of the truth
  • Allowing more people and levels of the organisation to contribute to accurate forecasting and planning in order to combine top-down with bottom-up; this results in the delivery of better information and more motivation of the workforce
  • Creation and input of targets in a Scorecarding concept
  • Production of looking-forward KPIs (eg. sales forecast, last best estimate)
In order to achieve these goals, a tight integration with the data warehouse is indeed required :
  • The data warehouse will have the underlying data to provide the detail, it will typically be relatively 'fresh' and will provide benchmark data for historical comparison, allowing to help discover outliers.
  • In case no tight integration between the CPM system and the data warehouse is achieved, the danger of different versions of the truth re-appears eg. data from the consolidation system versus data from management reports
  • Quality planning and forecasting processes allow easy access to recent past and latest actual data updates. Furthermore the best possible planning assumptions are coming from historical KPIs
  • Scorecards will typically define goals for each metric. These goals should be derived from the planning and budgeting systems and striving towards direct integration will reduce overhead, chances of mistakes and increase timeliness in case of changes
  • The forecasting and planning processes will themselves produce data that companies might want to analyse over time and that thus ideally are "archived" in the data warehouse in order to maintain historical versions and possibility to analyse trends (eg. sales forecast accuracy). The processes themselves result in KPIs that would typically be part of the company scorecard (eg. latest sales forecast for the month).
  • Finally, integrated CPM-data warehousing projects will benefit from economies of scales in the areas of ETL, data quality, metadata management, portal and security setup.

Bottom Line

As your company embarks on a CPM initiative, make sure you reflect on how the information flow and processes will be defined between your CPM applications and your corporate data warehouse that is in place. Forecasting and planning processes and Scorecarding initiatives heavily rely on information and processes that a state-of-the-art data warehouse will be able to deliver. Furthermore, company wide insights can only can be delivered from a data warehouse that integrates, next to historical data, also the "future" data and the externally reported consolidated data.