bintelog.com

IBM Cognos OLAP: TM1 versus PowerCubes

Miro Kosecek — Sun, 07 Nov 2010 00:20:27 +0000

The fierce competition of two OLAP technologies has not stop by merging their mother’s companies. Cognos in 2007 acquired Applix (though months later itself was acquired by IBM) not just to create synergy effect or broaden its customer base but to gain technological advantage too. Cognos’s PowerPlay, some ten years ago probably the most broadly used OLAP technology, has already reached its technological limits and new innovative products had superior performance and more capabilities.

In 2005 when it launched new generation of product Cognos 8 BI, the company has chosen to continue with old PowerPlay cubes under new name PowerCubes but with more broader possibilities to access the data stored in. Now for some time it has looked, that PowerCubes have their days counted. It is still broadly used and sold but with increasing integration of TM1 in Cognos BI it will probably take over. A recently launched Cognos Express , a Cognos 8 BI version for midsized companies, PowerCubes was ultimately replaced by TM1.

The advantage of TM1 is not only better OLAP performance but also its possibility to accept user data input that makes it not just analytic and reporting but planning and budgeting tool too so it can replace Cognos Planning (former Adaytum) as well.

Anyway IBM Cognos PowerCubes technology is still selling quite well within IBM Cognos 8 BI server and will continue in IBM Cognos 10 as well. If you want to use TM1 with Cognos 8 BI you have to purchase another server. Is it worth? What are the differences?

How OLAP works from end user’s perspective

Miro Kosecek — Wed, 11 Nov 2009 01:15:02 +0000

Video

Using OLAP (multidimensional analysis) brings brand new user experience but until users do not have it, they cannot figure out how they will use it. This is often a major issue when projecting business intelligence or data warehouse solutions. User are used to static paper reports and when interviewed by project analysts they describe their requirements from that point of view. So it happens that data warehouses or business intelligence solutions often try to use OLAP as purely reporting tool and users are not happy because OLAP tools often lack rich formatting capabilities. If the end users knew how OLAP works, many BI projects would be much more successful and satisfying for both the users and the implementators.

In this videocast I am trying to show to end users few very basic OLAP possibilities using IBM Cognos PowerPlay, probably the most popular OLAP tool. This is a first part. The second part will follow later where I will explain how to use OLAP for reporting.

Video

Questions to Consider When Building Business Intelligence in Retail Business

Miro Kosecek — Sun, 01 Nov 2009 17:58:05 +0000

The most important benefit business intelligence can bring to a retailer is exact knowledge about real margin achieved for every product sold in any shop in any day. Though it can seem trivial at first sight it is a major challenge when you try to do it.

Calculating the basic gross margin in fact is really easy. You just have to know the stock price and a price of the transaction.

Even when you have on stock a product purchased more times with different prices the stock price can be calculated simply either by FIFO (first in first out) method or by weighted average. In computerized warehouse you have no problem to have exact price for every item issued. Similarly for each sales transaction you have the exact price, so you can know the exact margin for it.

The difficulty is in bonuses and additional discounts that depend on sales cumulated over a certain time period. A very simple example can be a very common case of additional discount that a retailer receives at the end of the year if he sells a certain number of units. This additional discount affects the purchasing price of the product but since it depends also from the future sales in the moment of the transaction you do not know it. How to deal with it?

Data Sources for Business Intelligence

Miro Kosecek — Wed, 28 Oct 2009 21:48:14 +0000

The most basic way to deploy business intelligence is to connect it to data in existing operational databases, like ERP systems (enterprise resource planning systems e.g. SAP, Oracle Financials, Microsoft Dynamics, Lawson or other), CRM, billing systems, manufacturing control systems and so on.

This approach has an advantage that it is possible deliver particular solutions for many important business needs within weeks or months. This can serve as an eye opener for important decision makers and creates a new perspective of working with information.

Operational data however are structured and stored in a way that is not very suitable for reporting or even analytical needs. It is very probable that it will not be possible to combine some areas of data between them and therefore theirs information potential can be limited. Some reporting queries can also significantly slower or even halt the main system.

Sometimes this can be solved by staging the data or creating departmental data marts – that is creating data extracts and transform them into structures suitable for analytical and reporting purposes. Of course these data needs to be updated periodically; generally it happens at night time.

As the reporting needs grow however these data marts tends to become too complex and inconsistent. In that case it is necessary to create a so called data warehouse.

Business Intelligence Software and Systems, an Introduction

Miro Kosecek — Fri, 23 Oct 2009 02:15:46 +0000

Business intelligence is a name used for wide class of software applications that provides managers and other knowledge workers with information they need for managing more effectively their organizations (companies, public institutions, schools, hospitals, ngo’s and other). It includes tools for reporting, analysis, OLAP, dashboards, scorecards and others.

This information is important especially for better decision making and planning. It can be structured or not. The highest value information is generally not structured and is very difficult to process with computers. Current business intelligence systems therefore deal mostly with structured information. It is information that can be easily viewed in tables or charts. An example can be certain product sales performance by different regions in time.

The term business intelligence (BI) is most commonly used for formerly executive information system (EIS) used especially by SAS institute, decision support systems (DSS) and management information systems (MIS).

The core of most organizational IT systems nowadays is so called enterprise resource planning (ERP) system. The term is not very appropriate, but it became terminus for that class of software applications. ERP applications like SAP, Oracle Financials, Microsoft Dynamics, Baan and others has been collecting and storing large amounts of data with high potential of information that was scarsly leveraged. ERP vendors were not flexible enough to fulfill these needs and the niche was soon dominated by independent suppliers such as SAS Instutute, Cognos, Business Objects, Hyperion, Panorama, Aplix, Crystal, Brio and many others. SAP, Oracle and Microsoft tried keeping a pace with them both modes by acquisitions and proper development but no one of them ever become a absolute leader.

Since mid nineties the fiercest competitors in leading reporting and analysis were Cognos and Business Objects and they remained thus until two years ago. In 2007 Gartner estimated global volume of business intelligence market about $ 7 billion with consistent 7% average growth rate per year and the biggest players in IT market decided not to loose the game and this triggered major acquisitions. So Oracle bought Hyperion in spring. In summer SAP acquired Business Objects and finally on November IBM took Cognos.

How does business intelligence look today? Two major fields the BI solutions cover are reporting and analysis.