Bill Hackos, Comtech Services, Inc.

CIDM conducted a survey of XML and DITA productivity online in the summer of 2009. About a hundred DITA and XML users responded. CIDM executed the survey after a number of users mentioned to us that they had suffered losses in productivity rather than gains after moving to XML and DITA.

In this article I present the primary findings from the survey and I attempt to provide some insight into why some companies achieve large productivity gains while others do not or even suffer losses in productivity.

The Survey

The survey that we conducted was not a scientific sample because the responses were self-selected. Participants took part voluntarily as the result of an email request. Nonetheless, the results proved to be interesting and provided opportunities for more selective, in-depth study of the question.

We were most interested in determining whether loss of productivity was just an aberration with a few organizations or was more widespread. We wanted to know the following from our respondents.

  • How long have you been using DITA (or XML equivalent)?
  • What fraction of your publications group is using DITA or XML?
  • How do you measure success?
  • How do you measure productivity?
  • How has your productivity changed since moving to DITA or XML?
  • What factors account for the productivity change?
  • What are you doing to improve productivity?

The survey results showed that productivity problems are indeed an issue for organizations implementing topic-based authoring using XML but also gave us insights into the problems and questions we can use to extend the study.

Survey Results

The survey was put on the Web using Zoomerang. Ninety-five participants completed the survey. The results were not truly scientific because the participants may not be an accurate sample of all XML/DITA users. Nevertheless we obtained some surprising information.

Sixty-one percent of the participants have been using XML/DITA for less than two years while 17% have been using it for more than three years. About half of the participants are in organizations that have almost all writers using XML/DITA. The others have a limited number of their writers involved.

Forty-two percent of participants reported that they achieved a moderate or significant increase in productivity as a result of implementing XML/DITA. However, 44% reported a moderate or significant decrease in productivity. Length of use makes a difference. For organizations using XML/DITA for two or more years the situation is significantly positive. For organizations using XML/DITA for two or more years, 59% report a moderate to significant increase in productivity; 30% report a significant increase in productivity. Only 20% report a moderate or significant decrease in productivity. Ten percent of those report a significant decrease.

For the questions “How do you measure success?” and “How do you measure productivity?,” we found little consensus. About half the participants measured percent reuse, reduced writing time, reduced production time, reduced translation costs, and reduced production costs as metrics for success. Productivity was measured in terms of completing an entire manual or an information development project including multiple manuals.

For the question that accounts for productivity change, about half of the participants cited decreased production time, while most participants reported increased first draft time and increased planning time. Participants felt that productivity increased because of the reuse of topics, content references within topics resulting in fewer topics, and increased collaboration.

More than half of the participants felt that productivity would continue to increase in their organization as more training was given to authors, tools improved, and authoring guidelines improved.

A perplexing result was that about 14% felt that productivity would not improve and 6% felt that productivity would continue to decrease. Why would an organization continue to use a process that results in continually decreasing productivity?


These results show that start up costs, training, and conversions in the first two years limit the benefit of using XML/DITA. This outcome is to be expected because the shift to topic-based, collaborative authoring is not simple. In subsequent years, the benefits of using XML/DITA result in a significant ROI. It is important that writers, publications managers, and upper management are aware of the effect of these startup costs.

It is surprising that 10% of participants using XML/DITA for two or more years report a significant decrease in productivity. Why are they still using XML/DITA? A possible answer to this question is that not all of the questionnaire participants are managers. Many are writers who may view productivity only from the viewpoint of the author. If authoring is slower, the writer may perceive a decrease in productivity, but because of reuse, production, and the great benefits to translation, the overall productivity increase of the implementation may be substantial.

The success of a reuse implementation is dependent on many factors. Strong leadership and good management at the department level and at the upper management level are essential to success. Good processes and standards are equally important. Many implementations have failed because of shortcuts being pushed by managers, upper management, and tool vendors. Unless care is taken to structure content for DITA reuse, implement better workflow processes, and ensure project planning, the implementation will fail.

Many departments try to convert their content as is without structuring or minimizing. The result is a conversion to DITA without the opportunity to reuse topics creatively. The content repository is also burdened with topics that are of little or no value to customers but are nevertheless updated and translated over and over.

Upper management may insist upon resource and time constraints that are unreasonable. Management sometimes expects almost instantaneous ROI. These constraints put stress on the implementation and can lead to failures that would have been successes with more reasonable expectations.

Today’s insistence on offshoring by upper management can be deadly to implementations that require experienced, high quality staff and close collaboration.

However, despite all of the pitfalls that we see that lurk to thwart DITA and XML implementations, it is refreshing that we see so much success. We intend to continue an in-depth study of the productivity issues to help managers achieve success more reliably.