What needs improvement with Microsoft MDS?

Microsoft MDS isn't getting strong support because Microsoft is focusing more on cloud solutions. Profisee initially made MDS, which Microsoft bought. Profisee now handles the cloud version. There are changes because MDS is the source for master data, and other tools like Dynamics 365 could replace it but aren't promoted for that purpose. I don't have any specific feature requests for MDS when it comes to the next release because Microsoft isn't actively supporting it. Any improvements would probably come from Profisee, the company that acquired MDS. Microsoft still owns MDS but isn't adding new features. Master data management is moving to the cloud or handled on-premises by Profisee MDS.

The tool is very comprehensive; it has a lot of features and yet is easy to use. It is multi-domain which means that no matter what industry you are in, you can use the tool to create master data or reference data. One of the cool features of MDS is the fact that you can reference entities in different models in the MDS DB. MDS uses the model concept to encapsulate different domains in the MDS DB. From my understanding, MS, as of 2021, will not maintain the product going forward. That said, the original creator of the product is doing an excellent job building a new service on top of MDS called Profisee MDM.

It is a dinosaur of software. It is very outdated and not user-friendly. It is extremely outdated and not user-friendly at all. I have been working on a project with MDS, but I found it so difficult that I started looking for alternatives. Unfortunately, I couldn't find any. So, I had to continue with MDS. However, after a couple of days, I realized that I really needed to find an alternative because working with MDS is just too frustrating. In the next release, I need something that is easy to navigate and user-friendly. The way it has been created; apparently, the software was ages ago, like fifteen years ago, it was released by Microsoft. It might have been tolerable back then, but it doesn't meet modern standards right now. It's not convenient to use, and it's really rigid.

The solution can be improved by minimizing processing time for records. Currently, there is an issue because JMD repositories have records that come in thousands and millions, which takes a long time to process via the Excel plugin. Users get frustrated with that approach.

The stability could be improved.

The Microsoft license cost could be lower. We are a data company. We have loads of data and we are finding that, while it's working fine, if you're looking for reporting of big data, it has its limits. We'd like to have a NoSQL solution in Microsoft MDS.

I have faced data primary violation in MDS when the primary key was full and that is not good. There are occasions when the solution maintains SD and we get duplications of MDS. When I use the SIS package for batch activity and it's full, I have to delete the batch manually by going into the database. The solution can sometimes be slow. We are in a healthcare environment and I've faced some synchronizing issues with the test model. I've faced some issues as a consultant of this solution because of the above. It can be slow sometimes. I rate the solution six out of 10.

I would like to separate the website from the database side. By default, you have the webserver running on the same server as it is the database. This forces us to make a bigger VM which means more costing for the VM. We tried to go with a less expensive route and every now and then we run out of memory, you have to reboot the machine. It is okay in DEV and QA, but ideally, we should not have this happening. I do not like using Silverlight and Internet Explorer. The new 2019 version gets rid of that, which is one of the reasons why we are looking to switch. Additionally, the overall performance could improve.

The version that I work with is on-premises and it required an additional virtual machine on the cloud. It would be a better option to have an on-cloud version.

The only drawback is that it does not have the matching, merging, and all true MDM components. For these, you have to use another competent called Data Quality Services (DQS). You need to plug it in and use it along with MDS for true MDM. Both of these are integrated together, but you have to do them separately, whereas, in Profisee, there are a couple of screens where you can configure the matching process, create matching rules, and other things, and everything is in one product, which is not the case with MDS. In order to implement a true MDM, you need MDS, DQS, and SSIS. You have to use MDS to store your golden records, DQS to configure and standardize all your rules and matching percentages, and SSIS to load the data to DQS and MDS. At the same time, you also need Melissa Data to clean up your addresses to validate and standardize the addresses. That's the main component of true MDM. It would be good if they can create a true matching component inside MDS and merge MDS and DQS.

They do not offer the SQL Server Management tool via the installation. It is a separate tool I use when I'm trying to configure the mirroring with SQL Server Mirroring. This is not supported and I'm getting some errors on the database mirroring. So sometimes I use the 2014 management console and the 2017 server for that. If the customer does not agree to that, I use a query for the database mirroring connections. Additionally, I think some kind of machine learning related feature should be included. This is because technology is moving fast and all of the customers are getting it easier. So developers are making machine learning products. That's why they should include some kind of a machine learning feature here, too.

I would like to see better pricing - a drop in price would be great.

Most of the Microsoft partners, especially digitally, are separate. Personnel are business people, and they do not have technical expertise, so you end up as a company spending a lot of money training your staff and your engineers. This is something that Mircosoft needs to improve on. They must have personnel who have each shared capacity for their staff, so when it comes to implementation, it doesn't take a lot of time. The implementation takes months, and it could even take as long as a year. The implementation is a small issue because Microsoft does not come to the ground, it is the personnel. The only issue is pricing because technical support is there. Our engineers are well trained in Microsoft, and we have the Microsoft BI Office which is very helpful but pricing is a bit of a challenge. There is not much to be corrected other than just improve continuously on the business intelligence section. That is important for businesses.

In the next version, I would like to see integration with Azure. The duplication feature is not so good. The merge function should be improved to include better automation. For example, if two people have the same name, address, or mobile number then the records should be automatically merged. As it is now, I have to merge these records manually. I would like to be able to generate a report of current operations.

The solution could be better integrated into Microsoft's other products. One of the key features that it doesn't have is a data modeling feature. Everyone uses Microsoft, but not for modeling. They use IBM, SAP PowerDesigner, etc. It would be really good to combine this solution with a modeler. For me, that's what is missing. In a future release, I would like to see a better link to SSIS, Microsoft's business intelligence product. I would also like to see improvements in the user interface.

* Largest shortcoming: Business rules cannot be used to prevent bad data from getting entered. Rules can only apply to data already saved. And even then it requires the data steward to push a validate rules button. If they forget they won't notice (and notifications won't go out) regarding rule violations. In my 30+ years in IT I've never experienced a tool that only allows you to create reactive rules that identify bad data but never prevent it from getting saved. In a large enterprise I would rule this as a show stopper. * Bottom line: Business rules need the ability to prevent data changes, not just validate saved changes * Second largest shortcoming: Hierarchy integration views don't support custom attributes. This capability (which would be awesome if it worked right) is completely worthless without supporting custom attributes. * Explorer GUI is clunky and awkward. * Tool needs to get more frequent enhancements. For example, since 2016 there has been no enhancement except replacing Silverlight in the 2019 version. * It needs the ability to publish only a single business rule. Publishing business rules requires publishing all rules and results in requirement for re-validation of all rules. Until that is done the validation status on all affected records is in an "awaiting validation" state which can adversely affect data usage and validity. * Needs ability to remove menu options that may not be relevant to data stewards such as Entity Dependencies, Collections or Changesets. * Needs ability to default user group's first screen. For example defaulting to a specific hierarchy display. * Resolve a bug in the domain based validation business rule feature: the web interface will not allow correcting domain based lookups in the child properties (only works using hierarchy drag/drop or within Excel). * Hierarchy names should allow exceeding 30 characters. * Selecting child members within the navigator in a hierarchy results in the GUI just ignoring you... it should allow selecting children in the navigator or at least give an error message. (Note that 2012 version resulted in blank middle panel when this occurred rather than just ignoring you). * Needs to allow renaming the code and name attributes within the publication views (only allows renaming in the UI). * Derived hierarchy views should optionally include all attributes from all respective levels, not just the code and name attributes. * Needs to allow setting the CODE attribute datatype as numeric rather than forced datatype of string (which turns into nvarchar(250) behind the scenes). This is a significant headache as simple sorting on the CODE attribute with numbers causes numbers without leading zeros to not sort correctly and it requires adding business rule to enforce numeric datatype. * Needs to support setting view datatypes to varchar rather than forcing nvarchar so subsequent extracts do not have to do datatype conversion if not using unicode in your target database. * Versioning lacks granularity as it only supports model-level versioning. * The deployment migration GUI tool doesn't support data, only meta-data making the GUI export tool fairly useless. (Including data requires use of command line to support data) * The deployment migration tools do not support business rules requiring manual maintenance when using multiple environments (e.g. dev/sit/prod). * Excel plug-in needs to support hierarchies. * Needs to support email notification to more than one email group/user. * Staging process is non-intuitive and overly complicated. * Repository should utilize a prefix or unique schema for each model so that entities and views do not have to be uniquely named across models. * Needs ability to integrate seamlessly with data quality products. (DQS has some level of integration with MDS but the integration is overly complex and DQS's days are likely numbered). * Need to be able to do nNotification without an AD user first accessing the model. Once you setup the notification for an AD user or group it should just work without every user logging on at least one time first.