Item: IBM Cognos Analytics
Rating: 10
Author: Paul Mendelson

Use Cases and Deployment Scope

As an IBM business partner, I implement IBM Cognos Analytics with Watson for various IBM customers. These customers span almost every industry, from manufacturing to government, research to finance, and everything in between.

My current client is a large pharmaceutical company who uses IBM Cognos Analytics with Watson across multiple divisions. I've been brought in to help with the finance team.

IBM Cognos Analytics with Watson is flexible enough to make their reporting do everything they want. A major focus was ensuring high visibility reports work on mobile devices. Because IBM Cognos Analytics with Watson allows JavaScript and lets authors directly define HTML, I can make responsive reports that behave perfectly for any sized screen.

Another focus was expanding the default functionality in IBM Cognos Analytics with Watson through a series of JavaScript controls. Writing back directly to TM1, sending users emails from a button inside the report, or exporting reports to PowerPoint, the API available in IBM Cognos Analytics with Watson makes writing these controls quick and easy.

Pros and Cons

Pixel perfect reports. If a client can provide a pixel-perfect mockup, I can make the report indistinguishable from the mockup.
Data security. It doesn't matter who the user is; when they log in and run a report, they will only see the data they are permitted to see.
Performant reports on large data. Many tools have issues running on tables containing only a few million rows. Cognos can consistently run reports on multi-billion row tables without issue.
Embedding reports in third-party tools. With Cognos Mashup Services, it is possible to build a report in IBM Cognos Analytics with Watson and call it programmatically. Need straight HTML? No problem. JSON? Atom? All good.

Data modules still can't match the flexibility of the framework manager. When dealing with large data models, or multi-grain, FM still wins.
The new native graphs, while pretty, still can't match the version 10 graphs in terms of flexibility or options. It is possible to import third-party graphs, but it's a pain.
Some functionality, like version control, is only available through third parties. Motio or BSP.

Return on Investment

I've rebuilt a few metadata models to make the associated reports run at around a fifth of the time.
I was able to write and sell an IBM Cognos Analytics with Watson extension pack to add additional functionality and to make report author's lives easier.

Self-Service

The AI assistant is still fairly new. My current clients have evaluated it but have not yet implemented it. The big issue is their primary data source is based on a complicated TM1 cube, and the AI ontology cannot accurately match requests to specific tuples.

Once IBM implements it, we'll reevaluate. There is plenty of interest in it, though.

Data Security and Quality

Governance is extremely important to my client. The reports are grouped into different sections, access to which is defined through the users' AD groups. Individual reports are designed to filter differently based on who the user is and even present different aggregation levels based on user permissions.

TM1 security is used for row-level security, while object and query structure security is handled through IBM Cognos Analytics with Watson.

Reporting

This is my bread and butter. The flexibility offered is simply unmatched by any other tool. Custom graphs can be imported, and additional interactive functionality can be added with JS.

PDF outputs can be tricky, but with enough effort, it is possible to make an entire book.

In HTML, the reports support pretty much any CSS style you can think of. If it's not available through the GUI, just write it as straight CSS using an HTML item.

Most objects support conditional formatting or conditional rendering—highlight specific bars in a graph. Hide entire report sections if they're not applicable to the prompted values.

Other Software Used

IBM Planning Analytics with Watson, Microsoft SQL Server, Oracle Database

Key Insights

Do you think IBM Cognos Analytics delivers good value for the price?

Yes

Are you happy with IBM Cognos Analytics's feature set?

Yes

Did IBM Cognos Analytics live up to sales and marketing promises?

I wasn't involved with the selection/purchase process

Did implementation of IBM Cognos Analytics go as expected?

Yes

Would you buy IBM Cognos Analytics again?

Yes

Likelihood to Recommend

IBM Cognos Analytics with Watson is especially good with any type of pixel-perfect reporting. While it's most commonly used for financial reports, I've used it for building models for plenty of other purposes. The one that sticks out was a road safety organization that used IBM Cognos Analytics with Watson to research road accidents. The ETL process merged data from police, hospitals, insurance companies, high management, and several other sources. The combination of canned reporting and ad-hoc querying tools allowed researchers to find insights and possible causes of accidents that weren't previously considered. On the flip side, IBM Cognos Analytics with Watson is often used for things it's not really perfectly suited for. Data entry is a big one. You can write back to a database through it, but it's always a bit hacky. In the latest version, Query Studio is being deprecated. The dashboarding tool is significantly more complex. Sometimes you just want a quick and dirty query, and QS was perfect for it. Finally, there is a learning curve. You absolutely must learn how to use the tool, or you will have a miserable time. Once you get past that curve, you'll find it powerful and capable of meeting almost any need (even the hacky write-back requirements that people love to throw at me).

IBM Cognos Analytics with Watson - the best enterprise reporting platform

Modules Used

Overall Satisfaction with IBM Cognos Analytics with Watson