We released a great primer on “What is a metadata model?” earlier this week. The article also thoroughly covered why they’re an important component to a successful digital asset management system. Our special guest contributor, Ralph Windsor, Editor of DAM News, continues the discussion and dives deeper into helpful tips when actually deploying a metadata model.
This is one of the big debates in metadata design and there are pros and cons to both methods. In general, it is better to try and work out the fundamental elements of your metadata model in advance because otherwise the system configuration might need to be significantly changed or even assets re-catalogued again just because not enough prior consideration was made to the metadata model at the outset. With that said, it’s clearly not possible to plan for every scenario and you will be evolving the model over time by changing, adding and possibly removing fields, so you need to anticipate the need for some flexibility too.
From my experience, academic, preservation or public sector digital asset management implementations tend to include too many fields so that each asset takes a long time to catalogue (and costs more as a result). By contrast, commercial organisations will over-rationalise the range of fields to simplify the cataloguing task, they can sometimes later find that assets lack enough information to make them useful or easy to find, which also reduces the ROI of the system. Those are obviously generalisations, but depending on what type of organisation you operate, you can deduce from that where your issues are more likely to originate from. It will be different for every organisation and the best advice is to test early, often and get a wide range of feedback.
You will almost certainly not get it 100% right first time, but the more data you can collect (both before and after implementation) the easier it will be to optimise the model over time and adapt it based on real feedback rather than guesswork.
You need a cross-functional team of people to come up with a good metadata model, it can’t be left entirely to the vendor or a consultant because they don’t know your organisational characteristics as well as you. Equally, the vendor still needs to be involved because they understand what their system is capable of and can help devise strategies to realise your model within the scope of features that their application offers. You might choose to involve some external or metadata expert assistance also to help identify potential issues with the model which could be improved. It is essential to go through the proposed model with end users and if the system makes it easier to test proposed models first, preferably by getting them to enter some actual asset cataloguing metadata to see if it is effective or not.
Metadata models are one of the more sophisticated elements of Digital Asset Management, yet they can make a big difference to the ROI obtained. A well thought through metadata design draws a good balance between having enough detail to allow assets to be easily found in searches but without it taking too long for cataloguing users to enter metadata. Similarly, once the model is complete, although you will be adapting it over time, ideally it should not need to change a great deal so you can avoid a lot extra work changing it.
This article has only scratched the surface of this subject, but hopefully it should have given you some ideas for how to critically evaluate your own proposed DAM metadata model and incrementally improve the ROI you can generate as a result.