Test #1: This person thought that the title of the event was actually a button that you could click on. The person tried to click on the non-existent button for about 3 seconds and then figured out that it did nothing. The rectangular shape and the gradient could mislead people; however, the other participant did not have this issue.
Test #2: This person thought that the threshold number (in this screenshot '2') was too small. He thought that this was an important piece of information that was not given enough emphasis.
Consent forms were included in the original submission here.
Here is the usability script that was used to maintain some level of consistency between sessions:
Things to keep an eye out for:
Changes implemented in this iteration
Changes started in this iteration and will be continued in the next iteration
Changes for next iteration
These show redesign possibilities for people that were slow or hesitant to use the close and rsvp button. The far right shows the button as is (skinny). Second and third from the right shows a variant that is wider. And the variant all the way on the left shows a variation on the wide button where the button that opens the shelf stays in a constant position. This was a a problem during the in-person test because participants were having trouble closing the shelf when the close button slipped below the fold of the phone. You can see how this can be a problem in the variant second from the left but how it it might be improved in the variation on the left.
The top image here shows the app as it currently is with a box around the title. The second variation shows that boxed removed to show a more simplified and traditional title design. One person in the in-person test thought that the box was an affordance for a link. The second variation attempts to eliminate this affordance cue mistake.
Experiment #1: Can I click it? Yes you can?
Our in-person testing data showed the some people perceive the event title, which is currently contained in a gradient-shaded blue rectangle at the top of the event, as a clickable button. This may be a perception by few or many people, but the small sample size of the in-person testing group doesn’t help us generalize and answer this question. Thus, we will turn to online testing to increase the sample size, gather more data, and see if this perception is held by a significant population of people. In order to test this, we will have two interfaces to A/B test. One interface - the "A" - will be the original implementation of the button as it was in our user testing with the title contained within a blue-gradiented rectangle and the alternative - the "B" - will be a sort of addition by subtraction: removing the rectangle and making the title simply black text on white background in a size suitable of a header (slightly bigger than the body text). The "B" version draws on the the notion of a visual hierarchy (larger text + high on the page) to convey that it is the title, while the A version relies on style differentiation and positioning (different color + high on the page) in order to convey its function. Hopefully our online test will let us know which version fewer people click on. Clicks to this area is a negative in this test, as clicking on the title is the user behavior that we are trying to minimize in this experiment.
Experiment #2: The RSVP shelf button. (Size? and Stationary?)
We saw that opening and closing the RSVP shelf was a bit clunky in the in-person tests. So we designed some alternatives. In one variation we decided to make the shelf button bigger so as to provide a larger target area for our users to click on. In addition we also went a step further in a second variation to include not only a larger button but also an RSVP shelf button that stays in the same position when it is opened and closed. We noticed that it was a bit unnatural in the current implementation for people to have to click in two different locations on the screen to open and close the shelf. We hypothesize that keeping the button in one location but changing the text from RSVP to close will increase users propensity to click and close the shelves which is key to using the application. In this test we have two variations that will both be compared to each other and the original control design.
Note that experiment 1 and 2 will be run in the same session. Since both variations ultimately will be compared back to the control design we can combine the experiments into so that we can get more data with fewer people in less time.
Also note that this test provides small incremental variations that build upon one another. This is to avoid confounding variables and to focus in on specific causal relationships between the independent and dependent variables of our test. If we wrapped all of the changes into one we wouldn't know which changes were the ones that caused the results. So instead we split the changes into 3 variations that all can all be compared back to the baseline original design (1 variation for the header experiment and 2 variations for the RSVP shelf).