How help search should work

Thanks to Internet search engines like Google, software users have become more comfortable using search as their primary means of accessing help content.

Every help authoring tool seems to have a different approach to presenting search results. Offerings range from ranked results to alphabetical lists, with additional features thrown in such as the inclusion of chunks of topic text with highlighted search keywords. Each method of presenting search results offers different benefits to users.

Since help tools offer mixed approaches, I feel perfectly comfortable throwing my own opinions into the mix.

Two kinds of search users

In my experience, there are two types of searchers.

The first type is looking for the "needle in the haystack." They already have some expertise using the software, and use search to comb every chunk of content that contains their keywords. Overview information is not important to them; they are searching for a specific detail that could be anywhere. This type of user depends on full-text search functionality; search highlighting and summarized descriptions showing the context of the keywords are important to them.

Let's call these searchers "haystackers."

The second type of searcher uses search to develop a basic understanding of how to use the software. For example, when using a photo editing application, they may search on "editing images" to find overview information and procedures to help them get started.

This type of searcher is looking for topics that teach the basic concepts related to their keywords. Desired results might include "Overview of editing images," "Importing images for editing," and "Saving edited images." This type of searcher does not want to wade through a huge list of topics; they only want to see the topics that focus on the keywords they entered.

Let's call these searchers "generalists."

Building a better mousetrap

So how do we present search results that are helpful to both haystackers and generalists?

First, we have to summarize the needs of both types of searchers.

Haystackers want:

  • A complete list (all topics containing the keyword)
  • An alphabetical list of results to keep track of which topics they've already combed for details
  • Context summaries that hint whether a topic contains the details they're after

Generalists want:

  • Ranked results so they know which topics are the most relevant
  • Clear indication of overviews and important procedures (scannable results)

Now that we understand the needs of both types of users, we can build search results that meet those needs.

While ranking topics is important for generalists, I think that results must be presented in alphabetical order. Ranking means nothing to haystack searchers, and alphabetical presentation is essential for keeping track of which topics they've already searched. Therefore, a great search tool should rank topics while maintaining the alphabetical organization.

Instead of reordering topics to show rank, visual cues should be used. Highly relevant topics can be shown via bold formatting, or with a numerical rank following the topic title.

For example, if a user searches on "editing images" in a photo editing application, we could present results in the following manner.

Search results:

 Designating a network publish location
 Displaying the contents of the image recycle bin
 Editing image colors [2]
 Entering image titles manually
 Overview of the image editor [1]
 Sending images via email

This method of presenting images meets the needs of both types of users. Haystack searchers get a full list of results, organized alphabetically. Generalists can find the "big picture" topics quickly, and rely on numerical rankings for guidance.

We could go a step further and add context information for haystack searchers. For example...

Search results:

 Designating a network publish location
  "...so that you can publish your edited images to a network..."

 Displaying the contents of the image recycle bin
  "...allows you to retrieve image edits that were deleted..."

 Editing image colors [2]
  "...you can edit the color of your image before saving..."

 Entering image titles manually
  "...can edit the title of your image manually by..."

This method of presenting search results gives context information to user searching for specific details, but still allows new users to find basic overview information quickly.

The key to presenting great search results in your help system is to assess the needs of various types of users. By balancing those needs, you can guide users through your help topics and give them a positive software experience.

Perhaps you have thoughts on how search functionality should work. If so, please leave a comment and share your thoughts or any other technical writing tips with other HelpScribe readers.

More posts like this: