We currently do not use micro content for chatbots or other applications that require extra programming outside of a Flare project. However, we do use micro content extensively for featured snippets in search results of the output. See the full topic to learn what our approach involves. Our approach involves the following.
[Menu Proxy — Headings — Online — Depth3 ]
Our projects are integrated with MadCap Central, which gives us access to analytics on our outputs
We also use a filter to only show the terms for which we do not yet have micro content.
Although much of our micro content is created proactively (i.e., as we write topics), we rely on analytics results to help us fill in missing gaps.
Number of Micro Content Files
Based on the search phrases we noticed in our Central Analytics reports, we decided to make three categories of micro content files for each product, storing them in a Resources > MicroContent subfolder.
Definitions This micro content file holds simple definitions, sometimes with an image included, but usually not.
Quick Steps When viewing analytics, we noticed phrases containing verbs (e.g., apply condition tag, import Word), indicating that users are trying to find information on how to complete a specific task. Therefore, we created a micro content file and named it “Quick-Steps.” The response for each of these micro content phrases is a very brief set of steps, usually a shortened version of the main topic for that task. It also usually includes a visual element to the right, such as an image or animated gif.
Mini Landing Pages We discovered that the vast majority of search phrases in our reports are simply words or short phrases (e.g., skin, master pages, output), rather than actual questions or even phrases containing a verb. Therefore, we created another micro content file and named it “Mini-Landing-Pages.” Since these search phrases are somewhat vague (e.g., if people search for “skin,” we don’t know which aspect of skins they want to know about), the responses associated with the micro content phrases are designed to be mini landing pages, containing a short snapshot of the subject. This usually includes a table with links to processes, tasks, or other information that we think the user is most likely trying to find, as well as visual element to the right, perhaps even a tutorial or video link.
Structure and Style of Responses
In many cases, we wanted to include both text and either multimedia or images in our micro content responses. With text usually appearing on the left or top, and the multimedia or image on the right or bottom, we decided to create a few different responsive layouts; this allows the content to stack when viewed on a small screen
As far as the skin is concerned, we used the default settings for most areas of the micro content. However, we did change the following.
We decided to increase the height of the responses because most of our content was requiring users to click the expand button to see all of the information. So on the Styles tab of the Skin Editor, we expanded Search Micro Content Response > Size, and we changed the default Maximum Height from 200 px to 400 px. Although some of our micro content responses still require users to use the expand function to see everything, the majority of our responses can now be seen without doing this on a large screen.
Before the Change
After the Change
Response Links From Topics and Single-Sourcing Content
In the vast majority of cases, we create micro content from selected text within a topic. This lets us single-source content while automatically including a link back to the topic in the search results.
We also created some special styles that let us hide content in the micro content response, while displaying all of the content in the full topic. In addition, we have styles for changing the look of content when shown in the micro content response. The idea is that the micro content response should be a smaller representation of what a user sees in the full topic, without having to create two sets of content. See Single-Sourcing Micro Content Using Styles.