2 min.
Search engine ergonomics – part 1/3
1L’art de la gestion de projet2Un projet à succès commence par une bonne gouvernance3Cascade, agilité, demandes de changement?

Search engine ergonomics – part 1/3

Client experience & UX


Search engine ergonomics is not in its infancy and, at present, ergonomic recommendations for search engine design are relatively well documented (for example in the usability.gov book ) but are generally based on relatively old studies.

Despite the age of these studies, most of the criteria are still valid, because the behavior and cognitive abilities of users are not changing so quickly. What is changing rapidly, however, is our understanding of users and their environment. (The environment in terms of usage needs and technological capabilities of web development).

The purpose of this post is therefore to make a small inventory of articles that present a new look at this field. This post will be separated into three parts which concern respectively:

  • research behaviors,
  • The "advanced" search
  • The paging.

14 best practices associated with research behaviors:

In January 2008, boxes and arrows published an article on types of search behaviors . What emerges from this article is, on the one hand, a better understanding of the constraints that influence users and, on the other hand, the identification of six types of behavior with associated ergonomic recommendations. They therefore identified six important constraints: 

  • Expertise in the field (wanted),
  • research experience,
  • Cognitive style (global or analytical thinkers),
  • Type of purpose (research for navigation, informational or transactional),
  • Research method (precise or exploratory queries),
  • Research context (environmental parameters that influence users).

I do not dwell on these constraints because, unfortunately, ergonomists and developers have no control over them. The most interesting aspect of this article is therefore the identification of six types of behavior for which they provide design recommendations. They can be summarized as follows:
Alternate between browsing and searching:

    1. It is therefore important to properly link the different contents of a site in order to help users find what they are looking for once they choose a result.
    2. A breadcrumb and/or list of similar content may be used.

Minimize the result set:

    1. This is the role of filtering functionalities by categorizing results.
    2. To know the effectiveness of the filters, it is necessary to present the number of results found by the initial request and the number of results displayed thanks to the filters.
    3. The choice of query keyword binding operator is important. The author advises using the AND operator by default, but in this other article ( Advancing Advanced Search ) the author suggests that the OR operator may be more efficient in certain contexts.

Overview of the results  (The user reads the titles of several pages then makes his choice):

    1. To facilitate this, the titles of the results must be clear (especially for non-HTML documents whose filename is not always meaningful).
    2. It is important to highlight the words of the query in the description and/or the title of the results,
    3. If the results are paginated, users should be allowed to change the number of results per page.

Immediate judgment  (The user looks at the first results and determines if his request is successful):

    1. Optimize results for the most frequent queries (From search logs, test the most frequent queries and assess the quality of the results provided, then optimize the content of these pages to improve their ranking in the results pages).
    2. When the pages can no longer be optimized, it is possible to create “manually” chosen results and present them in context according to the query entered.

Formulation Difficulties  (User has problems formulating an effective query):

    1. Consider setting up help tools for query entry or offer suggestions. (one could also add, offer spelling correction and suggestions when the query looks wrong or doesn't return results).
    2. Setting up autocomplete is also very relevant in some cases.

Back and forth  (The user quickly walks through the results to directly see their content before choosing the best one):

  1. It is necessary to identify the links of the results already visited.
  2. Consider offering the option to open results in a new window/tab for easier comparison.

Conclusion of part 1  :

I must say that the recommendations provided in this article, although relevant, do not present major innovations because we already use most of them during our analyzes. What is very interesting, however, is the identification of research models and constraints. This can turn out to be extremely useful when you have to do ergonomic consulting for sites that target a specific category of users. In this case, based on these models, it becomes easier to focus the design on the important aspects for such and such a category (such and such a type of behavior) and thus create a search engine which will maximize the customer experience.  

to follow, a review of advanced research. (Keep you posted: )