From Management Today, June 1997

On-line documentation seems to offer a lot of advantages over traditional hardcopy manuals: instant distribution, easy update, no document control problems, endless possibilities for cross-references and hyperlinks...

What a pity it doesn't work. The sad fact is that no-one reads it.

Of course, on-line documentation is not alone in that. No-one much reads hard-copy documentation either.

It seems almost too obvious to be worth saying, but clearly, documentation adds no value to an organization if nobody reads it. And with only rare exceptions, corporate documentation - quality manuals, standard practices, work instructions, SOPs, and all the rest - has literally zero readership. With good reason. Most corporate documentation is ill-conceived, badly written, disorganized, ugly, and boring.

It is very difficult to get people to read anything. Documentation is a particular challenge. The problems faced by documentation managers are similar to those faced by any editor: busy readers, competing interests, and poor copy. On-line documentation doesn't solve any of these problems; and worse, it adds to them.

For one thing, most people won't read documents on screen. It's harder on the eyes, it lacks the convenience of a piece of paper, and a document on screen seems about a third longer than the same document on paper. As Bill Gates put it -

...Reading off the screen is still vastly inferior to reading off of paper. Even I, who have these expensive screens and fancy myself as a pioneer of this web lifestyle, when it comes to something over about four or five pages, I print it out and I like to have it to carry around with me...And it's quite a hurdle for technology to achieve to match that level of usability.

Bill Gates, it may safely be assumed, is computer literate. The same assumption cannot be safely made about an organization's workforce in general. There are still many people - a majority in most organizations, as soon as you get away from the desk-jockeys - for whom computers are largely a mystery. As anyone who works a help desk will tell you, a great deal of time is still spent explaining which is the 'enter' key or what 'point-and-click' means. Even amongst office workers who spend their days working at a PC, there are many whose computer skills have been learnt entirely by rote. They know just enough to do their data entry work or send their email; but anything much beyond that is a mystery and a terror.

Given the emphasis on documentation from the quality movement (quality accreditation such as ISO 9000 or 14000 is based explicitly on documentation) it is bizarre that so little thought goes into the quality of the documentation itself. To most quality practitioners, documentation quality means control - making sure that each user has a complete and up-to-date set of documents. Never mind whether the documents have ever been looked at. And if no-one reads it, who cares whether it is up-to-date?

'Documentation quality' should mean: how well does the documentation deliver information to its intended audience?

In quality terms, the reader is a customer. The information is the product; the documentation is the package. To have value, documentation must be readable, immediately to hand, convenient to use, and short. Like most products documentation has competition. Most obviously: go and ask someone, guess, or leave it till the next shift. Unless consult the documentation is easier than these alternatives, the documentation is worthless.

Readable means that the documentation has to be within the readers' literacy levels. In many cases - indeed most cases in non-office environments - this rules out narrative prose. Most people's reading skills are at their peak the day they leave school, and go downhill from there. About a quarter of adult workers can't read a newspaper article to make sense of it. Comprehension is lower still for technical information: the 1989 Survey of Australian Adult Literacy found that 50 percent of adults couldn't correctly interpret a Berger paint chart. There is no question that the vast bulk of corporate documentation is beyond the literacy skills of the vast bulk of readers. There are ways to approach this problem; but on-line delivery isn't among them.

Immediately to hand means that the reader doesn't have to go looking for the documentation. One aspect of literacy is the ability to find information - to use a table of contents or an index, or to make sense of headers and footers to guide you to a section and page. People who never read don't have this skill. The 1989 Survey found that 30 percent of workers can't use the Yellow Pages. The ability to find your way around a complex document requires an understanding of the underlying information model. If readers and writers have trouble with this on paper where the classifications are generally straightforward (if not particularly well formulated), then still less will they manage the abstract linking and drill-downs of on-line documents. And as many internet users have found, once you start following links you very quickly get lost. After two hyper-jumps, most users have forgotten what they were reading in the first place.

Convenient means that the documentation is in a form that the reader can physically make use of. Ring-binders, for example, are not a good choice if the readers don't have desks to open them on. (In one notorious incident, a company's quality accreditation hinged on whether it was reasonable to expect refractory maintenance workers to lug a three kilo lever-arch file into the blast furnace.) By the same token, on-line documents are no use to the majority of workers who don't work at computers.

Short means that reading the document is quicker than asking someone. For most purposes, that means one page. If you can't say it in one page, then in practice you can't say it at all. Not through documentation, anyway. A common mistake is to try to give the readers all the information that management would like them to have. It just doesn't work. By giving the reader everything, you give them nothing.

In a Management Today article Rob Johnson quotes a technical writer who 'is scathing about ... on-line systems that merely automate paper.' His contempt is appropriate, but misdirected. The fault lies not with the design of the on-line systems, but with the original paper. Managing an organization's knowledge, preparing it, producing it, and delivering it the people who need it is a difficult and complex task. It can be done well - and this is a distinguishing feature of good management - but technology offers no quick fixes.

The technological potential of on-line documentation is very seductive. Technologists and IT people nurture a powerful image of hyperlinked hierarchies of processes, with drill-down on-line resources, live performance indicators, improvement discussion forums, audio visual components, and all the other Internet and intranet bells and whistles. But it's almost entirely nonsense. The extraordinary and spectacular capabilities of the technology simply bear no relation to the capabilities of the users.

The cost savings are largely an illusion, too. By delivering your documentation on-line you save on the physical printing and distribution of paper, and you might reduce the administrative control costs. But in its place you need the software to make it run and a competent IT person to set it up and maintain it - which is no trivial task. You need training and help-desk support for the end users, for the auditors, and for those who are writing and changing the documentation. And since many readers - including Bill Gates - will print out the documentation in order to read it, the printing and paper cost is incurred anyway. On-line documentation is often a major investment for a negative return.

Of course there is a role - and a growing one - for on-line documentation. Some information changes so quickly or is simply so voluminous that there is no other good way to distribute it. Some information is targeted to competent computer users (help files are a good example). There is documentation that entails audio-visual and interactive elements, or that is part of a workflow automation system.

But for ordinary management purposes, with ordinary readers, if you actually want people to read what you have to say, on-line documentation can be worse than useless.

by George Kesteven