Why bother with end user documentation for Web Applications?

In Rahel Bailie’s excellent presentation at the STC Conference (“The New Face of Documentation“), she looked at the “No Documentation” approach to software user assistance. This, she summed up, as the “we don’t document it; we just fix it” view of software development.

She argued that a “No Documentation” approach doesn’t lead to no documentation. Users soon start to share their tips, tricks and information. They generate the content they need. The consequence of this is that the software developer loses control of user documentation – what is said, and which pages users view when they search in Google.

She made a good case for the need for user documentation where:

  • The application or system is complex
  • Training is needed
  • You want to guide users to additional features or services
  • The concept or process is not familiar to users
  • Assistance needs to be embedded in the User Interface

I think that’s a great analysis.

She covered a number of topics we’ve looked at in this blog: the impact and role of Twitter; Web 2.0; component based authoring of re-usable topics; user generated content; and an ecosystem approach to user assistance.

It’s clear that technical authors can produce more than just paper manuals. I’m sure in the next few years we’ll see technical communication evolve, as software developers embrace and master these new technologies, and user assistance, in some form or another, will still be needed.

“The worst Help system I have ever seen”

Sarah Maddox reported from the WritersUA conference that Microsoft’s April Reagan gave a frank presentation on the planning and design that has gone into version 3 of Microsoft Help.

She was quoted as saying the feedback on the Help 2 (used in Windows Vista) was poor. For example, “This is the worst help system I have ever seen”. 

At a previous WritersUA conference, Joe Welinske reported Microsoft implemented a couple of changes when it developed the online Help for Vista. The biggest changes were (a) they developed a new Help viewer and (b) they used technical journalists instead technical authors to write the Help topics. They chose journalists because they wanted Help topics to be closer to knowledge-base articles. I’m not aware of any other major changes.

I wonder if the change in writing  style was the main cause of such negative feedback towards Vista’s Help. Users often just want to do things, and they can be best helped by short, clear chunks of text focused on getting the job done.

It will interesting to see if Microsoft changes its approach to writing, as well as the Help viewer itself,  in future releases of Windows.

Help files – A question of trust

Last month, Forrester Research released results from a survey on how much consumers trust different sources for information. They didn’t include online Help or knowledge bases in the survey, so we don’t know how well or badly they would have come out in the survey.

They found independent (non-corporate) information were the most trustworthy sources. Top of the list was information contained in emails sent by people we know. Interestingly, the survey showed that only 16% of consumers trusted corporate blogs and only 33% trusted wikis (such as Wikipedia).

Commenting on this report, Dominic Jones has claimed that, for corporate blogs in particular, “it’s ALL about credibility and trust.”

So are Help files credible and trustworthy?

How could online user assistance be made more trustworthy and credible? If the independent sources for information are the most trustworthy, should online assistance contain links to independent, external sources of information?

The research shows that message board posts are trusted by only 21% of consumers, so would user generated content be seen as independent, impartial and trustworthy?

Josh Bernoff of Forrester wrote a post about corporate blogs in which he stated that blogs themselves are not the problem, but rather consumers are being turned off by how companies are using them. I would have thought that would be true for other online tools such as forums, online Help and knowledge basis.

However, I can’t help feeling that online user assistance is one of the most credible information sources provided by organisations, and that by integrating it more into company Web sites its trustworthiness could be put to greater and more wider use.

PS Happy New Year!

When you need Help, do you press F1 or search Google?

I was at two clients in Cambridge earlier in the week, where the conversation turned to the Internet generation and how they search for information. Both publish their user documentation on their Web site, as well publish traditional PDFs and online Help files.

The first client said that they’d found from their recent customer usability studies that the first action for 50% of those studied was to go to Google – even when they knew there was F1 online Help available. They said it was essential the Help content was available on the Web – otherwise, users would be relying on information outside of the client’s control.

The second client said they’d analysed the key word searches on their Web site and from Google. As a consequence, they were re-writing the Help topic meta data and titles so that they’d appear higher up the rankings. In some cases they were having to create content containing mis-spellings, which they said went against the grain.

It surprises me how few organisations make their user documentation findable through the search engines, and how few do any Search Engine Optimisation. As our first client indicated, if they don’t use your information they they may well be using someone else’s.

RoboHelp Packager for Adobe AIR

Back in the Autumn I posted about Abobe Air and how it could be used to provide a new medium for online Help. Today, I came across RoboHelp Packager for Adobe AIR, which is currently in beta.

It converts RoboHelp 6 or 7 generated WebHelp files into a single AIR file, which can be shipped to the user as an alternative to WebHelp. Air is similar to PDF, in that it will work across different operating systems in a consistent manner.