4 thoughts on “Where Do You Store the Content Metadata?

  1. Chris Campbell says:

    “Depends” is the correct answer. In the end, it depends on what you are going to do with all that metadata. In my mind, a properly maintained and organized collection of metadata can serve one of two purposes.

    The first is using the metadata to tag, label and describe documents. Perfect for searching, organizing and maintaining. Very ECM 1.0. This is all kept within the ECM system.

    The second is to take the metadata a step further and almost completely detach the metadata from the documentation. The document is still a record, but more of an “instance” of the metadata. The metadata can become a master data repository. Rather than many documents that have the same metadata entry, the model is entirely flipped. The *metadata* entry now has links to many documents.

    A practical example is having 100 documents with the author field value of “John Doe”. The typical query to find all documents written by John is: select document_name from dm_document where author = ‘John Doe’. If you’ve taken the ECM 2.0 approach, you instead query the metadata with: Select document_name from author.John_Doe. (Or something along those lines). Major performance enhancement.

    So where do you put this ECM 2.0 metadata? It’s one of the big reasons that EMC purchased XHive and is incorporating this into Documentum as fast as possible. XHive (as I see it) is going to be positioned as a way for other applications to access metadata while keeping it under the comfy warm security blanket of Documentum.

    If you don’t have XHive, then you’re looking at another solution. Depending on your industry, compliance rules and metadata usage the answer can change. If you’re selling simple Widgets and using the metadata across your entire company, from customer information to billing to manufacturing, then perhaps your metadata would be better stored in a ERP application like OneWorld or SAP. Financial institutions, medical facilities and universities can probably stick with keeping everything within their ECM.

    Personally, I’m weighing the options, but taking a real close look at XHive as something that bridges the gap nicely.


  2. Great comment Chris. Let us not forget the old standbye. The investigation is represented by a folder with custom meta data and the documents in the folder are the case documents.

    Simple, yet effective.



  3. When possible we always store as much metadata with the content to the manageability limit.
    One of the reasons being we often have requests from customers to build the business application itself on top of ECM an then integrate with other parallel business apps.
    Having metadata inside ECM helps also with the “central repository of information” marketing/sales pitch.
    We also have several solutions where metadata is at minimmum (reference id, security, retention info).
    I would support the idea of having information outside the ECM store only if we can think of these as 2 decoupled systems in terms of business process (not technology).
    So, it depends. IT is art, not science.


  4. I have worked on a number of solutions that use meta-data outside the content repository, also a lot more with it in the content repository. For me, depends is the correct answer as it really does come down to what you are using your repository for…

    A number of projects I have worked on used a repository simply to image enable another system, e.g. CRM. Retention periods could still be set based but the system was never used for anything else (well those class of documents). In this case, I had no problem with not storing meta-data in the system.

    Another example was that the meta-data needing to be stored was simply too complex for the content management system to handle (a failing of their chosen provider to be honest). However, in this system I did insist that certain meta-data, key to the files, were also stored. This was simply becuase the retrival system was, be it only from time to time, still being used to gain direct access to these files, not just from the third party system.

    So for me, there isnt a hard rule to follow. You need to make sure your TA or ETA has a good understanding of what the business requires and the way in which the system will be used, not just now, but also in the future…..If you dont have a good ETA then you may well have some troble….


Comments are closed.