The digital age is changing lots of traditional industries. Neither education nor publishing are exempt. Many educational publishers have recognised this and are beginning to adapt their content into Smart Content using metadata tagging.
There are 3 things to take into account when picking your metadata management tool:
- Understanding the product
- Standards used by the system
- The accuracy and specificity of the tool
What is Metadata Tagging?
Metadata tagging involves assigning tags to digital content, making it easily searchable, clearly organized and ready for new digital business models. Traditionally, metadata tagging was done manually, by employees laboriously labelling each new piece of content. However new AI technologies mean this process can be automated, saving time, money and unhelpful errors.
Typically, in technical world it looks something like this:
Even though automatic metadata tagging is a relatively new field, but there are already several tools available and look less scary, for example:
Source: EDIA’s plug-in to Alfresco
As an educational publisher with little technical background, it can be difficult to choose which tool to use, so here’s our guide to several things to look for when choosing a metadata tagging software:
1. Pick a product you understand
While it’s not important to understand the minutiae of machine learning or AI algorithms, it’s important to be clear on what each tool can and can’t do. Automated metadata tagging is a complex process and any good company should explain how their tool works in language that a non-professional can understand. AI and machine learning needn’t be incomprehensible; they just need to be explained properly.
Watch out! Many companies nowadays claim that they do AI, but in reality they still use simple mathematical formulas or outsource the work overseas to do metadata tagging manually.
Conclusion: Our take on this is that if a company can’t explain clearly what their tool does, it makes it unsure whether the tool is effective or usable.
2. Look for broad standardization
The publishing industry is specialized because it involves huge amounts of content which must be tagged according to several criteria simultaneously. Metadata tagging content allows for better organization and searchability. Some metadata tagging algorithms focus on keywords, some on barcodes, some use Natural Language Processing to determine language levels. As educational publishers share content with a broad range of clients, it is important to work with a standard that everyone recognises.
Here is the list of common metadata categorization standards across publishing industry making content helpful for publishers and their clients:
- Flesch–Kincaid readability ease
- CEFR tagging
- Bloom’s Taxonomy
- Topics and subjects
- Tagging against curriculum
- Language etc.
Watch out! Some metadata tagging software companies use their own system for classification. These systems are often unique to the tool, limiting its usefulness for your organisation.
Conclusion: When searching for a metadata tagging software for educational content, consider which standard to use and whether it is broadly applied.
3. Focus on accuracy and specificity
Specificity and accuracy are the reasons why AI tools are superior to manual work for metadata tagging, so make sure you take advantage of this.
It’s worth looking at rates of accuracy for your chosen metadata tagging tool. If there’s no information on a website about tool accuracy, feel free to ask. This is data that companies in this field should have to hand and gives a good marker for how useful their tool is.
Watch out! 100% accuracy isn’t realistic but above 75% is a great start considering the level of modern technologies.
It’s also worth asking about the specificity of an algorithm. Truly Smart Content is separated from its context, meaning it is possible to search for and extract content within books, chapters and articles without accessing the entire text. However, this level of metadata tagging requires detailed tagging, rather than just assigning broad themes to the content, like ‘Maths 13+’.
Watch out! Consider asking about the data that was used to train the algorithm.
Conclusion: Be curious about the nature that was used to create the metadata tagging offering. We also recommend checking out this brief course for getting basic understanding of metadata tagging for book publishing.
Whether you need an off-the-shelf or a customized solution, it’s best to work with companies who understand the specifics of your work. Deciding on a metadata tagging software as a non-expert can be complicated. A lack of knowledge may hold publishers back from investing in AI solutions. However, Smart Content is becoming the industry standard, particularly for companies who work with personalized learning. Considering what is important in an automated metadata tagging tool is an essential step towards this future.If you enjoyed this article, check out this article on 5 Companies Providing CEFR Tagging Tools .