JoAnn Hackos, Comtech Services, Inc.

CIDM has recently completed a productivity survey. In the survey, we asked if people are seeing positive or negative changes in productivity as they have implemented DITA and content management. As of today, we have had about 100 people respond to the survey. Even more interesting, however, is that more than 250 people have viewed the survey and decided not to respond. I will report on the survey in this newsletter after the Best Practices conference.

But right now, I’m really curious about the non-responders. The survey asks a lot about measurements. I wonder if people decide not to respond because they don’t have solid data on productivity changes, whether up or down.

Every DITA or CMS project I’ve reviewed in the past 10 or 15 years begins with a business case that includes predictions of productivity gains and cost savings. But, just what happens to these promises? How many of the same managers who predict cost savings are also able to show that the predictions have come true?

Measuring change starts with data. Managers need to know, at the beginning of a change process, exactly what the current state might be. If we measure productivity by the number of pages we write and publish in a year, then we must collect data about the time spent producing those pages.

First, we need to know how many pages we actually produce. In traditional, book-based metrics, we count pages published in final deliverables. In traditional help system development, we count the number of help topics written (given an average length of help topic). We also need to distinguish between the number of pages published and the number of pages touched in a particular release. Many pages are never touched at all. Others are completely rewritten. Many new pages are developed. We need to count all of these possibilities to understand the output of our teams.

In my book, Information Development: Managing your Projects, Portfolio, and People, I provide examples of accounting for new pages, minor change pages, major change pages, and pages without change. Enacting more detailed accounting of the present state gives us a way of really knowing what changes to expect through our change initiative.

Based on this as-is analysis of productivity, you can effectively calculate the affect of content management on productivity. If you believe that your current process results in duplication of content, you can predict that the number of “pages” worked on for a release will go down. At least, your prediction would account for unnecessary rework on duplicate content. Unfortunately, a decrease in the number of “pages” produced will result in showing that your productivity has decreased rather than increased.

Productivity is a measure of efficiency. An increase in productivity tells us that fewer people can produce the same work as before, or the same people can produce more work. Your calculations of the benefits of moving to XML-based authoring and content management must show either

  • you will need fewer people in the future (a cost savings on salaries), or
  • you will be able to produce more work in the future with the same number of people (a decline in the rate of growth)

Let’s look at an example:

Today, 10 writers produce 10,000 pages of deliverables in one year. Of that number, 2,000 are new pages and 5,000 have changes in the content. The remaining 3,000 pages have no change.

You calculate that among the 10,000 pages, 25% are duplicates. That is, 2,500 pages of text in the final deliverables appear more than once. That means that your 10 writers actually produce only 7,500 pages in one year. By eliminating the duplication, your overall production of unique pages (with no other changes in content) would actually decrease, negatively affecting your productivity. You can gain by retaining only three-quarters of your original staff.

However, you also calculate that your volume of content has been increasing by 2,000 pages per year with the same 25% of content duplication. That means you produce 1,500 additional new pages on average each year. That means that within two years, your existing staff (with no layoffs) will be producing more content without duplication than they did originally.

You can use the extra one quarter of your staff to do the migration to the new content design. You may also find that you have continuing additional overhead due to the need for more collaboration and planning of managed content. Your actual savings per year may be closer to 15% of the original. After a few years, you will come out ahead.

Consider that you may have more than 25% duplication in your content, which means you will recover your start-up costs more quickly. You may also have a large organization in which the additional overhead is amortized across more people, leading to a faster return of your investment. However, small groups are often easier to manage through collaboration.

Another consideration is the time everyone spends formatting. We have senn studies that report desktop publishing and production activities can easily consume 50% of writer time. By automating the production, eliminating all desktop publishing, and placing a specialist in charge of final production, you may eliminate more wasted time than you had originally calculated through reuse alone.

Don’t mislead your management by predicting instantaneous savings. It takes time to work out all the details and to recover the costs of your initial investment. Remember, as well, that minimalism trumps everything. Writing less and delivering only what customers really need is always the first and best goal to pursue.