John Frazzini, VMware, Inc.

People like to have information presented in a way that resonates with their own needs. A hurried system administrator is usually looking for the exact answer to the current problem he or she is facing. An application developer might want to browse all of the capabilities of an API before deciding on an application’s implementation. Different media lend themselves better to specific types of information. For example, showing a user a video might be a better way to describe a complex process with multiple steps, rather than laboriously explaining the process in a text-based instruction.

A year ago, the documentation team at VMware decided that text was not enough for today’s sophisticated information consumers. We needed to enhance our portfolio of delivery strategies to include video among all of the other ways that we communicate with our users.

With the dramatic success of web-based video hosting systems such as YouTube, companies are learning that video can widen the audience of users that use the documentation. Linking the documentation to videos or vice versa can be used to point viewers to more information that they might be interested in. Using these linking strategies can provide a useful measure of user engagement and provide insight into how to further encourage users to read or view more documentation.

Criteria for Videos

Other teams were already producing video assets, and it was challenging to differentiate the type of content the documentation team was going to produce. For example, the corporate marketing and the technical sales team produced videos to demonstrate product features or illustrate competitive advantages. The support organization produced how-to videos for common product problems to reduce the number of calls to customer support.

To ensure that we fulfilled a specific need, we created a set of criteria for our videos:

  • Videos should be instructional.
  • Videos should be related to one or more topics that are already covered in our documentation so that we can easily link the two formats, allowing users to consume the information in the form they want.
  • The script of a video should link a concept topic with the procedure so that a user can implement the concept.
  • Concepts that can easily be illustrated through graphic animation, such as time-based processes or communication between multiple systems, are good candidates for videos.
  • Long and complicated processes that would take more than three to five minutes to explain are not good candidates for videos, unless you can divide the topic into a cohesive series of videos.
  • Videos should provide some value beyond the written documentation, such as listing benefits for a feature or describing a use case that the feature satisfies.

Using these criteria, we evaluated our existing documentation for candidates and prioritized the videos to create. We then assigned each video topic to a writer based on subject area expertise.

Developing Expertise

At first, the idea of creating videos was daunting for the writing team. No one was particularly sure how to proceed, and it was easy to get bogged down in all of the video production details. To determine how to begin, we started a pilot project with a small team of eight writers. The team met for eight weeks. The goal was for each person to produce a complete video. We learned several valuable lessons during this process.

The most important lesson is that writers are particularly well-suited to create scripts and storyboards for videos, but they are not usually adept at video production techniques. We decided to focus on producing high-quality scripts and storyboards and to hire a video production person or contract out the tasks of filming the presentation, editing the video, and creating animations.

Secondly, through trial and error, we developed our criteria for selecting topics to translate into scripts. Our criteria are still evolving, based on each new video that we produce, but having defined criteria helps to easily identify topics to produce as videos.

Lastly, we already knew that review is an important part of information development from our experience writing documentation. Aside from the usual review process for technical accuracy of content, we developed specific techniques for reviewing video scripts:

  • All ideas for video topics are presented at “pitch meetings” to a team of peers to get immediate feedback on whether the idea meets the criteria and how it can be improved.
  • Approved ideas go through the development phase to produce a script, a storyboard that has the general flow of the video, and preliminary graphics.
  • During the development phase, the script is read aloud to identify any problematic phrases or awkward descriptions and to get a sense of the general length.
  • After the video is shot and edited, a review board from the management team carefully examines the video for any problems, such as using copyrighted graphics or having continuity errors, and gives the final approval.

Along the way, we identified several tools that we needed to evaluate and obtain:

  • a presentation tool that has basic graphics and animation features to rapidly prototype and review scripts and storyboards
  • a tool to edit video segments after they are filmed
  • a tool to create animations and to produce the correct video format for our web hosting platforms
  • a sound editing tool to improve or edit the audio track, when needed
  • a variety of cameras, lights, and audio equipment

Finding the Time to Implement Video

Traditional documentation and release deadlines make it difficult to keep up the pace of delivering video. However, we make it a departmental goal to include video as part of the package of deliverables for a release.

We have found that when we get close to a release, spending a few weeks identifying video topics and producing scripts can easily become part of our documentation process. As long as this is a planned activity, it becomes part of the frantic rush to get a release out the door.

Measuring Success

We use a number of measurements to determine how well our deliverables are performing. First and foremost, the number of videos is important. For our first release, we created 12 videos. Over the course of the year, that number has grown to about 30, with new videos released each month. For the next release of our product suite, we plan to add 20 videos for a total of 50 videos produced in the span of a year.

While the number of videos we have created might not sound like a lot, the goal is to continuously produce more content to draw an audience to our offerings and keep that audience engaged. As such, another key measurement is our subscriber base and video view numbers on YouTube. Our YouTube channel currently has about 2000 subscribers. Some videos are more popular than others, but each video generally has between 1000 to 5000 views. Considering that we have done very little outreach and have made few attempts to draw the audience to our video content, we view our one-year view rate as a success.

Another key measurement we look at is the average viewing time of each video. We keep most videos at three minutes or less because viewer attention span is short. We track the amount of time each viewer spends watching each video. If 60 percent or more of the viewers keep watching through more than three-quarters of the total video length, we judge that as a successful video. If, for some reason, viewers start dropping off in large numbers toward the middle of the video, we see that as a sign of a less successful video. We try to determine what keeps viewers interested longer and incorporate those characteristics in future videos.

What’s Next?

Certainly we want to expand upon our success in the first year of our effort and continue to produce more video content. We also want to study our successes and failures and tune our video criteria to produce more engaging content. Also, we want to more tightly integrate our traditional written content with our video content by embedding our videos in our documentation and providing links to related content.

We have made several decisions about how to present our content. For example, all of our videos have an on-screen presenter because we want to provide users with a face to relate to a feature. This decision has resulted in a particular look and feel for the videos, but we are constantly evaluating whether these decisions result in better or worse user engagement. If you want to see the results of our efforts, please visit our YouTube channel at