Reducing time for task completion by 47%

OVERVIEW

Custom Forms was reported to be the feature that a major portion of world-wide customers would be willing to make a purchase for. This was much more than a Google Forms clone, it would need to allow for complex expressions based logic and integrate with our suite of products seamlessly. I decided that we needed to start by getting the base interactions clean and solid.

An old version of the design that I was tasked with improving the interactions on.

A revised first stab at simplifying the field addition interaction.
A revised idea at simplifying the field addition interaction.

THE MESSY MIDDLE

After running some ideation workshops with internal stakeholders, I created this alternative to the Speed Dial FAB. After internal tests, it was discovered that the new addition symbol was too small to be easily noticed. This, combined with other feedback led to my creation of the next version.

Our reworked navigation panel including the field addition interaction.

After multiple rounds of testing and iterations, I had refined the base to a silky smooth puree of pure interaction glee. The time to complete each task had gone down demonstrably and there were no clear observable trends of problems with the designs.

Next, I needed to build on the meat and potatoes of the basic addition of form fields to include layout and logic options.

After a few more rounds of testing and iterations, I arrived at the following designs.

A blank form field
A blank form field
The same field with logic added to it to hide it in certain situations.
The same field with logic added to it to hide it in certain situations.

RESULTS

The result was an expression language similar to Excel. I knew going into this project that the usability would be less than an ideal solution, but for timeline reasons it was deemed the most appropriate course of action. Indeed, during the testing, the results were not very encouraging on our final round of testing. However, after quantifying the results of the tests, we saw a clearer picture. Our new iterations had reduced time to complete tasks by up to 47% over old designs, resulting in a huge gain. Task accomplishment rate was at an astounding 0% when asked to do this with 0 knowledge (including needing to guess the perfect syntax for the expressions), but after reviewing the help articles and training available to them, success rates jumped to 100% on follow-up tasks.

Overall, this was a project that pushed my ability to collaborate with team members to its limits. What started as a murky web of interconnected feature requests and unvalidated assumptions ended up in the timely delivery of massive improvements not only to the visual design but also to the speed and efficacy of our users.