Think Info

Exploring the information space

How do we manage this Context thing, anyway?

When people discuss content, another word is bandied about with impunity: context.

context [kon-tekst] n. the set of circumstances or facts that surround a particular event, situation, etc.

dictionary.com, based on the Random House Dictionary, © Random House, Inc. 2011.

Everyone is in agreement that we want and need it. When we communicate with others, it is vital that we take context into account. Context serves two purposes:

  • to provide a baseline for understanding, the assumptions of meaning
  • to make the choices to communicate the appropriate message to the other party, in the best possible way

I am going to deal here with context in the second of these forms: the process of determining how to communicate with another party. Most people do this instinctively in person (a salesman reads his customer’s body language and responses, and adjusts his pitch accordingly). But when we move context management into the digital realm – when we have to hand these decisions over to a computer – things get tricky.

It’s complicated

Context is a big problem; it is convoluted.
The problem with context - who is responsible / qualified?
Adobe’s CTO for Customer experience, David Nuescheler (@davidnuescheler) said in an interview with CMSWire that “the context travels in places where the content repository can never reach” (first video; @4:35).

Now, I agree with David on how complicated this subject is. However, I find his statement worrying from a CMS vendor: if context is so complicated that it extends beyond the ability of the content store, how can the CMS be configured to make use of it in a meaningful way?

Context, at the end of the day, is a tool used by marketing to pitch segmented information at customers, whether potential or existing. What we need is for our CMS’s to gather information – behavioural, demographic, sociographic, preferences – and make decisions about how to serve our information in the most effective way. (Yes, analytics need to feed back into context management for continual improvement.) We take criteria-axes and use them – where the individual interacting with our system falls on each axes – to determine what to display.

Why is it so messy?

There are two reasons why context, when considered closely, scares most people.

The first is simple to understand: when there are more than two or three axes being taken into consideration, the combinations get out of control; multiply the number of options on each axis together and you will see the problem. This is a brute-force form of contextualisation; it relies on someone making all the decisions in advance. There are other approaches to this issue, such as the associative recommendations made by Amazon, using meta-data, based on what you have recently looked at; better, but still incomplete.

The second reason – and even those who do the best job so far in contextualisation miss this – is that context is non-commutative. What do I mean? Simply that the order in which the elements of context came into effect is itself contextual. If I look at your product page and make a purchase, my context is of a customer who has found what he wanted. If, however, I purchase your product, then come back and look at your product page, my context is of a customer who cannot understand your product (or can’t get it to do what I want). The elements are the same, but the order means the difference between pleasure and frustration.

Can we get this context thing right?

In order to do context properly, we need two things:

The CMS: Context management system

Firstly, we need content management systems that are also context management systems. These CMS’s need to understand that context is not about an intersection of segments, or having demonstrated a particular penchant. We need CMS’s that can cope with the complexity of non-commutative context; that work off relative emphasis rather than absolute options.

C3IA: Content - Context - Container

C³IA: Content - Context - Container

While we have CMS’s that – at least in theory – separate content from presentation (or let’s call it container), we need CMS’s that go a step further and treat context as a third, distinct, element in the process. We need an information architecture based on the 3C’s: content, context and container. This is not a simple case of context driving content, but a triangular relationship where any two poles will influence the third.

The Context owner role

We also need context managers – people who can get their heads around the options and variations. People in marketing are, generally, too focused on their singular push – this week’s hot new product – to consider the wider variation of contextualisation. Developers may enjoy coding the engine, but applying the rules is not their cup of tea. Content owners and creators can understand that they need to create contextual variants, and they can craft words accordingly, but often find contextual algorithms too technical.

We need a class of people who enjoy puzzles, who can think in multi-dimensional abstraction, to manage contexts (and probably also to design the context engines for the new CMS’s). This is a new role within the digital workspace; an interface between marketing, content strategy, editors, developers and analytics.

Advertisements

2 responses to “How do we manage this Context thing, anyway?

  1. Pingback: The Enterprise as the Context « Think Info

  2. Pingback: What does your CMS actually do? « Think Info

%d bloggers like this: