Ibm

IBM Cognos Review
Good product with bad performance. We had licenses for all the features but users almost exclusively used the molap cube file.


Valuable Features:

The molap cube file is valuable.

Improvements to My Organization:

We shared the molap cube file, which means can do analyses on their desktop.

Room for Improvement:

Cognos is a good product with bad performance. We had licenses for all the features but users almost exclusively used the molap cube file.

Use of Solution:

I've used it for 12 years, since v7.

Deployment Issues:

It has poor integration with other IBM products. In v7 (not in the IBM family), I overrode their upfront UI to integrate it into a websphere portal. In earlier versions, I still find these kinds of problems.

Stability Issues:

Almost all the time, we have a problem with Cognos stability and even our maintenance vendor can’t do anything to resolve it.

Scalability Issues:

I think Cognos have problem with their dispatcher performance.  In v7, I overrode their impromptu options, to catch performance requirement, while in v8, again, I had to override their dispatcher management with our code.  Finally, in v10, we didn’t make any changes to improve Cognos, and now our users leave it.

Customer Service:

It's good.

Technical Support:

We are more expert than they are.

Previous Solutions:

We used Infobeacon and now we are using Cognos and Microsoft BI. We are currently looking for a better product and are evaluating Tableau.

Initial Setup:

In 2003 It was very complex, and we must survive by ourselves because we can’t find any technical support for us.

Implementation Team:

We always implemented it through a vendor, but until now, we are more expert than they are.

Other Solutions Considered:

We evaluated Microstrategy as another option.

Other Advice:

Quality of data, flexibility and the performance of the tool(s) are the critical aspects of BI. Test the critical aspects by doing a proof of concept using a huge amount of data.

Disclosure: I am a real user, and this review is based on my own experience and opinions.
1 visitor found this review helpful
Anonymous avatar x30

5 Comments

Nyc
DerenVendor

Hey,

You should take a look at Sisense as well. Sisense is a good mixture of MicroStrategy and Tableau. You will get the best of both worlds.

Like (0)30 April 15
Anonymous avatar x30

In my experience performance is a function of design. Well designed solutions perform well, poorly designed solutions don't.

For example a poorly designed data structure mapped to a DMR model in Framework Manager may not perform well. A well designed PowerCube (molap) can perform very well, especially if you understand what hierarchies will be challenging and address these through design. Well designed Dynamic Cubes can handle large data volumes and perform very well.

If you want maximum performance for your data sources, with scalability, then Dynamic Cubes are the way to go. Even better if you put them on top of a BLU database, and better still if you use the columnar database options.

I am sorry to hear you can't find a skilled IBM Cognos partner in your part of the world to help you. There are many good ones, and some not so good. This is true of any vendor's partners.

Good luck with your investigation of Tableau. Just remember you are comparing apples and oranges since they don't do the same thing. IBM Cognos has many capabilities and is pretty good at them all. Tableau is great at one thing only.

The closest comparison would be authoring dashboards in Cognos Workspace Advanced with RAVE visualizations, and delivering them using Workspace, all on the web, vs doing this client server with Tableau.

(I am an ex-Cognos product manager, and currently work with an IBM Business Partner - about 17 years of Cognos BI experience)

Like (0)30 April 15
Anonymous avatar x30

The shortcomings you describe likely are the results of 1) COGNOS' MOLAP architecture and 2) no single metadata.

MOLAP is fine for fixed or slowing changing reporting, where what you are looking at is well known and you have plenty of time to respond to new needs. It is less than ideal for analysis because by its very nature, the analyst doesn't necessarily know everything they are going to question, thus the need for analysis. To get over this hurdle, you should look at a tool with a ROLAP architecture.

COGNOS has been built over the years through acquisition and "integration" of disparate products and has always had a problem with trying to integrate the metadata from each. This lack of complete integration not only contributes to errors in the data across functions, but also contributes to stability problems.

An additional problem, from an analyst's point of view, is that as your MOLAP-based system becomes larger it becomes ever more difficult to manage the number of cubes and decide which cubes may be needed for analysis. And then, if you don't have a/any cube(s) to answer the analyst's questions, a new cube needs to be built, which negates the speed advantage of using a cube. This is much less of a problem with a ROLAP system because the developer is not pre-calculating answers, but is waiting until run-time. While this can cause a major performance hit, ways have been developed to overcome that.

If you have not followed the BI industry closely but have only followed developments with COGNOS, get hold of a recent copy of Gartner's Magic Quadrant for Business Intelligence and Analytics Platforms, a Business Application Research Center (BARC) BI Survey, or a BI Scorecard.
You can usually get these as marketing literature from the various BI players.
Disclaimer: I'm a MicroStrategy-certified BI consultant.

Like (0)30 April 15
Anonymous avatar x30

I have to dispute some of the comments above. I believe the above author's comments are outdated.

Firstly, IBM Cognos has a great ROLAP offering in the form of Dynamic Cubes.

Cognos Series 7 was a set of products most of which almost all were built in house. Cognos Query was acquired. PowerPlay, Impromptu, Upfront, Architect, Transformer, Visualizer were all built in house. It is fair comment that there were differences as PowerPlay exclusively used muti-dimensional sources, Impromptu exclusively accessed relational sources, and did an OK, though not outstanding job, on star/snowflake schemas.

Cognos 8/10 was built entirely in house, and all components share the same query engine. It has none of the legacy of Series 7.
Cognos TM1 was acquired from Applix back in 2007,

As I posted earlier, I am an ex-Cognos product manager.

Like (0)02 May 15
Aamer shariff pmp li?1424191760

I have been working on IBM Cognos for over 8 years from the earlier versions to the most latest.This product has evolved a lot over time and has always been in pace with the business needs and data reporting requirements which have changed tremendously over the years.IBM Cognos is far more mature product than most of the competitors , but keep in minds its expensive.The product is now moving more towards self serve analytics and predictive analytics.

Like (0)28 September 16
Anonymous avatar x30
Guest

Have A Question About IBM Cognos?

Our experts can help. 222,498 professionals have used our research on 5,689 solutions.
Why do you like it?

Sign Up with Email