In human computer interaction, the gulf of execution is the gap between a user's goal for action and the means to execute that goal.[1] Usability has as one of its primary goals to reduce this gap by removing roadblocks and steps that cause extra thinking and actions that distract the user's attention from the task intended, thereby preventing the flow of his or her work, and decreasing the chance of successful completion of the task.[2] Similarly, there is a gulf of evaluation that applies to the gap between an external stimulus and the time a person understands what it means.[3] Both phrases are first mentioned in Donald Norman's 1986 book User Centered System Design: New Perspectives on Human-computer Interaction.[4]
This can be illustrated through the discussion of a VCR problem. Let us imagine that a user would like to record a television show. They see the solution to this problem as simply pressing the Record button. However, in reality, to record a show on a VCR, several actions must be taken:
The difference between the user's perceived execution actions and the required actions is the gulf of execution.
Example of Gulf of Execution:
An example of gulf of execution is, if a user wants to save a document in a word processing software, but they are unsure how to access the "Save" feature or they cannot find it easily due to unclear labeling or hidden menus, it creates a Gulf of Execution. The user's intention to save the document is not aligned with the system's interface or available actions, causing frustration and making it challenging for the user to complete the task.
Example of Gulf of Evaluation:
An example of the Gulf of Evaluation can be seen in the context of a voice-controlled virtual assistant, such as Amazon Alexa or Google Assistant. Imagine a user giving a command to the virtual assistant to play a specific song from their music library. After issuing the command, the virtual assistant responds by playing a different song or fails to understand the command altogether.
In this scenario, the Gulf of Evaluation is wide because the user may have difficulty understanding why the virtual assistant played the wrong song or why it didn't recognize the command. The user's mental model of the system's response and behavior may not align with the actual outcome, leading to frustration and confusion. The system's feedback, in this case, is not adequately helping the user evaluate whether their desired action was successful or not.
To bridge the Gulf of Evaluation, designers could improve the feedback provided by the virtual assistant. For example, the assistant could respond with a confirmation message, such as "Playing song 'X' from your library" to ensure the user understands what action it will take. If the assistant misinterprets the command, it could provide an informative error message, such as "I'm sorry, I couldn't find the song you requested. Please try again."
By providing clear and meaningful feedback, the Gulf of Evaluation can be narrowed, enabling users to better understand the system's response and assess the success of their actions.