DCL/Conditional Content: Tips for Avoiding Pain

When Conditional Content Goes Wild: 4 Top Tips for Avoiding Pain – Part 1

By Noz Urbina, Content Strategist and Founder, Urbina Consulting

After presenting and writing a lot on branding and content marketing lately, this article takes me straight to my techcomm roots. That said, it should be valuable for many types of communicator. In part one I’m going to:

  • Define conditional content and its role in delivering contextually appropriate content
  • Explain why market trends push us to deliver contextually appropriate content
  • Illustrate why we need to – and a bit of how – we can minimise use of conditional content

All of this applies to users of FrameMaker, MapCap Flare, SDL Architect, all the many DITA tools, or other any tool that supports or encourages conditional reuse. This article is in no way a criticism of any of these tools, but a cautionary tale about abusing the power they offer.

In part two, we’ll go deeper into how conditions can be avoided with judicious use of content modelling and content strategy and also look at a manager-friendly explanation of its issues to help justify investment in changing a creaking implementation.

What is conditional content? A quick intro for communicators

(If you’re already familiar with conditional content you can safely skip to the next section.)

Conditional profiling – often referred to simply as "using conditions" – exists to address a very valid business need: the ability to specify small fragments of content as being specific to given target scenarios, products, platforms or other audience subsets.

In the days of the early web and print production, when we needed to create content for multiple audiences, we simply added more content and duplicated it across outputs. Copy, paste, rinse, repeat. Everyone receiving everything was the only feasible way to get the job done. We still see this today in appliance and gadget manuals where we have to read through line after line of:

“If you have an AveeMaster Bravio BX920, your connector panel will look diagram 1.2, and you should do step 1. If you have an AveeMaster Bravio CX720 full hd, then your panel will look like diagram 1.3 and you should do steps 1 and 2. If you have an AveeMaster Bravio CX420…”.

It is a painful way to consume content and clearly done for the sake of the publisher’s process and costs rather than the user’s experience.

Conditional content offers the ability to mark individual fragments or modules of information for a specific audience, e.g. Bravio BX920 owners vs maintainers, and then use an automatic process to filter the appropriate content out so that users need only read what is relevant for them. It makes content more reusable because you can use one module of content in several situations, and just filter it as needed for the context. This helps us deliver a more contextually appropriate experience.

Content Module

Conditions allow us to mark parts of a document as context-specific and let a computer do the filtering for us

Why (and how) you should minimise or avoid conditions

I am known on the conference circuit for my “cautious futurist warnings” about upcoming industry challenges. Cautious because audiences tend to not want to prepare more than a year two ahead of time. After all, who knows what the future will bring? That said, here are some two predictions I am willing to wager significantly on:

  • Users have become accustomed to the web’s speed and personalisation ability, and as a direct result, all communicators have been pushed to tailor content more and more – why would they want to wade through irrelevant content in one case when they know that technology protects them from needing to in others? As new channels like wearable technology, embedded content and augmented reality rise in popularity, we will be faced with more demand for different outputs that are appropriate for use in ways and contexts for which we hadn’t written or planned.
  • The industry will continue to struggle to even address the demand for personalisation that exists today. Although we have been talking about it for years, most organisations have not developed content strategies or information architectures suitable for scalable, highly tailored publishing.

The difficult truth is that we need contextual content more than ever. Although this may suggest we want to use conditions more, there are some good reasons we should not and many ways to avoid them. You can use these as a check-list to validate your content as you go.

1) Remember that conditions aren’t the only way to reuse

The best way to start avoiding conditions is to simply question each instance you might use them:

  • Could you accomplish the same thing with transclusion? Transclusion is the use of text from a shared source makes it appear duplicated. Like a copy and paste operation that keeps a link to the source in all the places it’s ever been pasted. In DITA this is called a conref or conkeyref, and in Unstructured FrameMaker a text inset or variable.
  • Could you break up the content into smaller modules instead? If you break content up, you could include only the relevant content in your deliverable.
  • Could you write around the differences? Maybe you could rewrite the content to make it more general and avoid the need to use a condition at all. Sometimes you can just say “your TV” instead of “Your Bravio BX920” with no loss of quality for the user.
2) Take the conditions acid test: should the content be reorganised?

I think one of the most glaring flaws in the whole conditional reuse paradigm is often overlooked because it’s about users, not us. It feels obvious once you think of it: Conditions can’t reorder or reorganise content in any way.

Consider these questions:

  • Is simple filtering really publishing an optimised result for both audiences? For example, to turn a user’s manual into a maintenance manual?
  • Do the two intended audiences approach things in the same order? Will they look for things in the same groupings or chapters?
  • Is the deliverable hierarchy appropriate or might some audiences get better use from different nesting?
  • Would a different information flow be appropriate for one user group versus the other?

The ease of filtering content and the inability to reflow or reorder helps us, consciously or not, brush these concerns under the rug in our pursuit of maximum reuse and efficiency. In this way they facilitate our being self-centred, instead of user-centric.

I have seen organisations doing user and task analyses with real users have their assumptions in this area strongly challenged. If you address any of the questions above and find that a simple filtering isn’t ideal for users, then conditions are not the appropriate way to handle the reuse between these outputs.

3) Don’t create “unlimited value” conditions

If you are going to use conditions, use them for things that don’t grow indefinitely in number. The worst conditional nightmares I have seen are those that conditionalise on things like:

  • Products (in companies with large numbers)
  • Product versions
  • Specific customers (when these are numerous)
  • Product releases

Take the example of labelling things as specific to a product: If you are going to keep to a small number of conditions, and your company supplies over 20 products that might use the same procedure or reference, then you have already got a problem. If you only sell 3 products, but you are conditionalising for every release or version, again, conditions that were easy to manage at first will be overwhelming within 3 years. Customer-specific content can be even worse.

Try to keep conditions on things that won’t increase too often in number. Regional markets are safer, depending on your industry (in areas like medical devices even this can be difficult). Audience profiles like “maintainer” or “novice user” also tend to be smaller and more stable in number.

4) Pull the chute sooner, not later

Pull the Chute

What I did on my summer vacation – Noz in the skies over Holland

The conceptual simplicity of the conditional filtering approach is what makes it so popular and appealing; however, it suffers from major scalability issues.

I like to try to limit to 10 potential conditions. So in any one deliverable (site or document) you should have a maximum of 10 potentially “special” fragment types on which to filter. At CSApplied 2013, Ann Rockley said her recommended number is a mere 6.

When I have started a project, many of my clients have upwards of 20, with some reaching 60 and even 90. This much mark-up makes creating content slow, painful and error prone – something I will be illustrating in more depth in part two of this article.

Backing out of this level of use can require changes to process, platforms, and people which take a very long time, and involve:

  • User and task analysis
  • Persona and scenario development
  • Content audit and (gap) analysis
  • Information architecture
  • Content conversion and (extensive) rewrites
  • Retraining and updates of style and writing guides

However, it is in many cases conditions are “the devil we know”, and thus we delay a change. This is exacerbated by tool vendors promising that their tool or new upgrade will make this problem manageable without any real changes. These tools demo nicely with the 4 or 5 conditions that demos tends to use. If they showed you what year 2 or 3 might look like, you might not have been so enthusiastic.

Conditions – a symptom of a bigger issue

I fear communicators are in a vicious cycle today. As the change in our market accelerates, the longer we avoid taking on revolutionary changes in search of simple short-term incremental changes, the bigger our long-term risk. Short term simple can be medium-long term awful. The risk increases with every delay that in 2 years’ time, management or the market will push us to deliver something in a matter of months that would have needed a 3-7 year transition process to prepare for. This is a current reality for many organisations for whom I have worked.

In the next part of this article we will be looking at how to explain the issue of conditions in a manager friendly way, and digging deeper into some of the alternative approaches that I have mentioned here. But before then, please let me know about your biggest conditional content horror stories, best tips, and other thoughts on Twitter or by email so I can try to address them in the next post.

Noz Urbina, Urbina Consulting

Noz Urbina is a Content Strategist and is the Founder of Urbina Consulting.