Bulk actions is a typical UI feature in web applications. This article highlights 2 parts in design process. They are “understand and validate design question” and “Test”.The overall process is showing below.
The story starts from product owner sending me the requirements. Based on customer feedback, they want bulk action on node list screen (the object is called “node” in our product) because they have to execute some actions like “deploy” and “remove” repetitively.Before moving forward to generate design solutions. First, I validated whether that was a real design question. I asked PO to send customers following questions:
After validating those questions, I know the design question is real. And by measuring the value and effort, Product owner and I both believed it worth to solve. Moreover, in ideation phrase, I kept reminding myself to think about more solutions than “bulk action”. Because usually, the problem we see is superficial. It exists due to another underlying problem. For example, in this case, the underlying problem is “machine can not execute automatic or smart action”.
Define User Tasks:
1. Find how to add nodes
2. Find where to remove nodes
3. Find where to refresh
4. Find where to deplay applicatopns on Serial No.1, 2, 3
Define Testing Objectives:
Find out which design can achieve below metrics, If no design meets, ask follow up questions to find out why and then iterate the design.
Effectiveness: 100%
Time Based Efficiency: Highest
Task Level Satisfaction: Highest
Testing Approach:
A/B test to large audience is the best way to test Time Based Efficiency and Task Level Satisfaction. But to reduce the test cost, it is good to have some one on one usability test to test the effectiveness and get some idea about the other 2 metrics.
Testing Procedure:
Note: Because users may have preconceived prejudice. To reduce the deviation, the prototype order for each user is different.
Design 1: Action Sections
Design 2: Dynamic Actions
Design 3: All in Model
Task Completion:
The usability metrics should be used flexibly and contextually. In this case, Effectiveness should be highly considered. Time Based Efficiency and Task Level Satisfaction would be secondary metrics because there are only a few users attending the test, which may cause bias.Consequently, design 1 is the best option for this specific design question. Don’t copy, steal.