The Art and Science of Content Models


The problem of personally organizing content certainly reminds us of the challenge of organizing data.

In the data management world, the original disruptive innovation was the personal calculator-with-a-memory. But even more influential, many would argue, was the spreadsheet.

If you leave out storytelling, building spreadsheets is one of the few ways that most people have ever stepped beyond lists and outlines when they wanted to represent a logical grouping of subject matters. The “natural” appeal of spreadsheets is easy to describe: I have several things that I care about (columns) and I care about each of them in several ways (rows). It is not very difficult to navigate the collection of data.

The simplicity of that approach is why it is so easy to stretch beyond organizing “facts” (the typical spreadsheet ingredient) into organizing “ideas” (the stuff that content collections are made of). have several ideas that I care about (columns) and I think about each of them in several ways (rows).

A typical frame of reference (or framework) tackles the task of organizing ideas about a given subject area.  Usually, the first challenge is to decide what is important to understand about the subject “this time”. We “model” the subject area — literally, we represent it — with those decisions.

If there are three main things to know this time, the framework considers them to be three “dimensions”. The most familiar example of this is with objects, where we can  pay attention to the height, width and depth — 3 dimensions. We might add a fourth dimension — time — and even a fifth dimension such as cost. You can probably imagine that fewer dimensions are easier to manage and think about than are more dimensions…

Finding Content Value

One of the biggest challenges we have now is from having so much unrestricted access to information. Whether we see information as data or as ideas, nearly everyone experiences an overload of it. The problem is not something that just happens to us; much of the time it is because we are actively looking for information, using fairly powerful tools; and in the process we’re simply getting more than we needed or knew what to do with.

When we get around to taking the resulting collection of information seriously, one issue is to separate the valuable stuff from the rest. But we are easily reminded that, like clothing, particular information has varying importance according to how and when it might be used. We might keep a wide range of things just in case”

Often we find that if someone else goes through our collection of information they come away thinking of different dimensions than we do. The difference is usually attributable to their point of view and their need. That is, we may not always be aware of how the information can have value until someone or something shows us. Punchline: there is not only one “right” way to understand something. There can be multiple ways. The dimensions that you choose make up your model of the subject area at that time. You can also make or get other models for other occasions.

One of our colleagues from the “data science” world, Joseph Pusztai at the company Cubewise, talks about this same flexibility. He writes:

“…accurately representing the dimensionality of your model is where “modeling art” meets “modeling science”. Many data mining (e.g. classification and clustering) algorithms purport to identify the naturally occurring dimensionality and hierarchies in your data, but often human intuition can do a better job, as well as enable you to introduce new dimensions into your model that did not historically exist (e.g. a car company launching their first electric vehicle will have no historical sales or production data for it). Humans are also very good at understanding that there is rarely one giant monolithic model behind a set of data, rather, we are usually dealing with dozens or even hundreds of smaller independent models, holistically interacting with each other along common dimensions. “

In that statement, we can substitute a few words (switch “data” to “content”) to recognize how the statement applies to ideas (concepts) as well as to facts.

An algorithm is like a filter; the way the filter is constructed will allow some things to be kept on top or forward, while other things fall through or away. We might literally discover a range of different things that have something in common keeping them caught, as one group, by the filter. As an alternative to algorithms, human intuition, usually reflecting experience or belief,  can also provide a filter and  a “commonality” that groups things. That one thing found in common is like a dimension.

Human understanding, especially including different flavors and levels of subject expertise, easily accounts for why one framework can satisfy an audience while other frameworks are also valid for the same content. With eXie, one collection of content can be seen in various ways, for example by several of the different personas that are found among eXie users.  A historian, a teacher, and a designer may have different respective frameworks for the same collection of content.

Do It Yourself

For your own purposes: when you pick a subject area, decide what level of expertise you want to use, and choose your “dimensions” (called themes in eXie), you are then well on your way to cataloging your curated (selected) content items for an audience. (Remember that your primary audience may be yourself in some future moment!)

By using themes to point out what you care about in the subject, your framework columns get named. Then you can point out the main ways that you usually care about those themes, which become names for your framework rows (called topics).

In data modeling, there is a lot of excitement about the powerful new tools that analyze facts and discover ways to organize them in models. With content collections, subject matter experience is held in an extremely fast and powerful tool — the mind — and modeling the collection of concepts takes place with both some new discoveries and as reinforcement of what is already familiar.