1) From Youtube, Instagram along with enterprise collaboration tools, we’ve spent the last decade seeing the rise of authoring tools that require less time and skill to create content for teams to learn their jobs.
For many of us who remember back in 2008, this was the rise of the rapid authoring tools and the beginning of making it easier for subject matter experts to create learning content themselves. From the corporate learning departments specific authoring tools made it easy to create a powerpoint that you like, convert it to a course. The resulting trackable course was ready to run on your company’s LMS. Times have changed though. Those systems were priced in the region of $500- $1000 per author and so this limited both the amount of content produced by the few number of “authors” who felt comfortable using them and was a natural barrier to only allowing those considered worthy of such tools in the company to create new materials in the first place.
Now, we are seeing the rise of all manner of content creation vehicles. Collaboration and social tools like Youtube, Instagram, Google docs along with stalwarts like Microsoft Office have only added to this. Not only are there more and more learning systems which include authoring within them other non-learning tools such as blogs, web page based manuals, pdfs, video from phones and even Slack, email or text messages people create to answer a question are being used in 1-1 or one to many job training. Not to mention how slide decks, PowerPoints and Google Slides are not diminishing in use but growing as well. Much of this content may not be in the form of trackable courses but because of how much more easy it is to share and gain direct participation these collaboration tools are really how many people chiefly interact with their colleagues and by extension learn their jobs.
2) Knowledge management and learning vendors are loath to admit; there is a LOT of bad content mixed with the good without an effective way to find it.
We as the vendors in managing as well as creating company learning content and knowledge derived from existing company content we never want to think that the content produced by our customers is not necessarily ideal to learn or provide support to do their jobs. However, creating bad content is the flip side of unleashing extremely easy to use authoring and collaboration tools. Or worse, becoming bombarded with so much content that the goal of determining useful and accurate from the rest becomes even more difficult. From a recent post "The productivity pit: How Slack is ruining work" its clear when we take into account the exponential growth of all produced content, the many versions of those pdfs, word docs, Google sheets is bad enough, now when collaboration tools such as Slack, Microsoft Teams or Atlassian Confluence impacts come into the picture, the problem is even worse. Which content or content fragment is useful to me right now? It's been a tough problem to solve.
3) What the knowledge management and learning content vendors never like to emphasize; most current approaches to content curation needs human intervention.
Back when LMS vendors and 3rd party content vendors were primarily focusing their content on compliance/ generalized training using on demand learning as a replacement for existing classroom training, few were concerned with the effort required to maintain and update the learning content produced. It still needed to be kept updated there was simply a lot less of it. Knowledge management vendors took the role of curating content to mean automation was applied to content management, but someone with the right credentials still had to give a "thumbs up" or down on produced content. Even if it was as simple as liking a post on Facebook. Today so much more content is used as micro learning events; such as videos, images, documents and all that needs to be maintained as well. But as we all well know, few of us have time to do so. Even 3rd party learning content vendors run into this issue. As their library grows so to does the effort and time required to keep all their titles relevant.
4) Current methods of measuring and optimizing a company's content are not effective. They are good practice its just getting harder to follow them.
Adopting a content strategy where your team measures when content has reached the tipping point of usefulness or optimizes the content through revision or reviewing sounds like a great idea. But it's a matter of the time such content strategy takes that is the limitation. There are increasingly not enough team members with enough time and resources to effectively do this on the myriad of content within your company.
5) The big content collaboration vendors recognize this already.
Back in September of 2017, Box.com one of the worlds largest Enterprise Content Collaboration Platforms announced they would be applying Natural Language Processing (a form of Artificial Intelligence) to all 30 billion files that they managed for their clients. The goal is to specifically identify, classify and categorize automatically the billions of photos their clients have and be able to quickly organize these unstructured forms of content. Apple, Google, Amazon and Dropbox of course have similar goals and processes. All of them realize expecting us “humans” to manage this content deluge is not reasonable.
We can look no further than Amazon specifically in their successes in managing a vast amount of varied content. As a shopper on their site we have access to millions of book and product SKUs. They apply 1) Social rating – what you experience when using social sites like Facebook by measuring “likes”, shares and votes. 2) Collaborative filtering based on users past actions and 3) Semantic Analysis which understands the meaning of small segments of the content and then applying machine learning analyzes relationships between all these segments for future suggestion or recall of specific content. Amazon uses a combination of all three techniques to position the right book or product based on our behavior, peer experiences as well as having a semantic understanding of the product page we’re viewing. This has doubtlessly taken many thousands of man hours and peta bytes of data to analyze to improve these algorithms. But over the past few years we are seeing additional “unsupervised” AI driven algorithms that are able to automatically categorize and organize content without the need for training data and the required time involved. These are known as unsupervised AI / deep learning approaches.
Companies like getguru.com are making similar strides to further automatically curate the peta bytes of content produced in Slack, Confluence and even Google Sheets. Here at Feathercap, we apply AI driven semantic curation and search to find the right sentence or video segment within any of your team's content to answer their questions. Content such as pdfs, Word, Powerpoint, Video, audio and our own authored learning content.
Thankfully, as the above examples illustrate, this overall content curation and optimization problem is being tackled on many fronts.
See our AI and the Augmented Workforce primer on how a workforce and technology can effectively work on tasks together. Including curation.