Ads Format Confidence
Ads Format Confidence
The goal of Ads Format Confidence is to call out useful, relevant, curated and reliable information (Qualifier) from ads content in a simple and efficient way so that users can quickly:
Identify what they are looking for or
be inspired with a stronger decision. This helps bolster trust and make users feel more confident with Google ads.
Short-term Value to identify what critical information to highlight for a better UX. Long-term Value to get Fundamental learnings in how varying visual elements impact parsability and reduced clutter (simplicity).
From our research, we found out the more useful and higher quality the content is, the higher prominence it should have to draw users attention. This is how we help users quickly achieve their goals and find our products useful. However, if every single content is highlighted in a high prominent way, it will add much clutter which blocks users from finding what they want.
Hypothesis
01
Highlight a qualifier is critical and will have a positive impact on users and user experience
Original: No qualifier
Highlight qualifier
02
A system-level UI consistency of the qualifier across result types will help improve parsability and reduce clutter, especially when it’s more visual
I defined UI consistency criteria from four categories: Format, Color, Typography and Placement. For example, from the format perspective, a qualifier could be (1) plain qualifier only has text or icon with lowest prominence and scaled size to fit varying scenarios; (2) badge qualifier is a badge styling that has a size limitation for its content but could include text extension if more info display needed; (3) rich qualifier can be paired with media and compensate for visuals that used when it needs to be informative and high prominent cases
03
Text heavy results should be compensated for more visuals to make content outstanding and appealing
Text heavy results with no qualifier
More text, less visual
---------->
Less text, more visual
Visual minimal results with no qualifier
More text, less visual
---------->
---------->
---------->
Less text, more visual
When visual is already rich enough, there is no need to keep compensating for it to avoid clutter and keep a balance to maintain parsability
04
Qualifier formats should be scaled and customized based on visual density & user intents to keep visual balance and fit user needs
High density page
Low density result
minimal info & less prominent
---------->
---------->
---------->
Low density page
High density result
More detailed info & more prominent
Experiments
Based on our four hypothesis, I had proposed groups of experiments candidates to validate our assumptions.
Appendix: Prominence levers
Placement
Users are more likely to see information the higher up (top vs bottom) it is and most cases tend to read left to right (leading vs following).
In search, we heavily use position to communicate which result is most likely to be important and useful (content with higher value & quality).
Richness
Richer patterns usually correlates with high quality content. On today’s prod we identified couple format elements that increase richness:
- icon,
- badging style,
- medium/bold text,
- capital letters,
- structured text,
- interactive text
….
Size
Bigger formats usually contain more elements, which also attracts more attention from users.
In search, height has powerful effects on metrics because it excludes other results from the screen.
Color
The human visual system is highly tuned to pick up differences in luminance and hue.
Use of color can attract attention over desaturated elements, even without adding content or size.
Since most of search is white and desaturated, any use of color attracts user attention.
Divergence
One of the things the human visual system is most attuned to is contrast. If one item is different from its cohort, it immediately stands out.
The irony is that if everything is different, nothing will stand out - it is just visual noise for users.
Interactivity
Low prominence formats help reduce clutter and make it easier for users to skip. However, sometimes we want a format’s prominence changed based on users’ showing their intention/action