It's a tough tool for developers. Right now in the market, I think Informatica MDM is number one and is used a lot. 

IBM MDM was developed by some other company and IBM developed it more. There are definitely some flaws. The tool is difficult to understand from a customer perspective and for developers. Generally, customers can understand Informatica MDM and its functionality easily. If a developer is facing issues with IBM MDM, then a customer will definitely not understand most of the functionality. That is the main concern.

IBM is introducing some Next-Gen MDM, where they are using a lot of current technologies and machine learning. Most of the problems we may be facing right now are probably solved in that version.

IBM InfoSphere MDM's UI is complicated. It also needs to improve its on-prem configuration. 

There are not really many areas of the product that need improvement. They have already moved it to the cloud and that is available at the moment. It is not really an additional feature because that capability is there, I just have not used it myself. I left just before the last company was with was moving to the cloud environment. That would be useful in my opinion if it had been available to me.  

They had just upgraded the UI for it in a previous release, which was a major improvement. Three or four years ago, I would have instantly said the Stewardship UI needed an update if asked for an area that needed improvement and it would absolutely have been my first thought. But that has been done recently and they did a nice job of it.  

They have good interfacing with other databases within the suite, and that is a fine feature. Because of the complete integration of the InfoSphere suite, the integration with the ETL (Extract, Transform and Load) and the other parts of InfoSphere work really well together.  

I really do not think there is anything else I need. It does what it needs to do for the MDM market and its users.  

Though probabilistic search is supported, this feature is not an out of the box API. We need to code and customize product to retrieve the results. Thanks to the IBM team, this was fixed in v11.4.

The master data, and the reference data management in the toolset are not good. The solution isn't giving us what we want, so we are considering building something internally ourselves. If IBM can improve the way that they did data reference management and master data management then I'd say we'd consider staying with the solution.

We also want better cloud integration, so we could do things in the cloud if we needed to. This would be a good feature to add in a future release.

• Documentation
• GUI
• Training

It needs a better database.

At the time global name recognition was purely batch-oriented, as it needed to become transactional/real time. The stack architecture seems too complicated as a result of having too many components that needed to be built together.