Case Study: Facebook Story Midcards
An early sharing suggestion midcard.
Part 1: Early Days
The initial design was focused on functionality and minimizing risk. Our sharing suggestion tech was very basic, and we worried users would be alarmed to see their private photos on what is normally a surface for shared content.
Users could edit every aspect of the suggestion (photo, song) directly on the midcard.
Editability was essential given primitive suggestion tech. Most suggestions consisted of the most recent photo you took combined with a top 40 song.
Content strategy focused on minimizing risk. We emphasized that nothing had been shared yet, and they could change anything before sharing.
The midcard was a success, sparking a gold rush to scale the feature. A dedicated department was created for creating new midcards.
Part 2: Gold Rush
To scale the feature, Product leads decided to focus on editing people’s photos, hoping it would make the suggestions more shareable. These edits included effects, AI backgrounds, stickers, and pre-written captions.
Users could no longer edit everything directly on the midcard. The approach was too hard to scale in a “throw things at the wall” phase of product development. Users now had to make edits later in the flow.
Content strategy emphasized the unique edits to the photo. Since we could no longer hedge the suggestion with editability, we hedged by treating the card as a sort of demo of the various edits people could use, often adding a “Try it” button, in addition to a “Share.”
These experiments were very hard to ship, and almost all of them failed. The lack of design standards for this new format meant that every aspect of every midcard design was subject to scrutiny. The lack of A/B testing made decisions difficult and political.
Part 3: Quality Push
Around this time, I began establishing and implementing design standards for the growing story midcard system.
Larger images, with dynamic sizing to fill the screen. Data showed that larger images increased shares.
Smaller headline fonts, to allow for larger media previews and more space for content.
Consolidated privacy strings from 7+ variants down to 3, all fitting on a single line, even on the smallest devices.
Finally had support to run a few small content experiments on CTAs and headers.
Design and content standards implementation → 17.4% lift in publishing rates
Part 4: Optimization
Dropped subheads in most cases to further maximize preview image size.
Shifted from benefit-led copywriting to direct "Share a …" formula, emphasizing the “what” of the card before the “why.”
Deemphasized music and effects, adding focus to the media and minimizing clutter. Data showed that bad song suggestions were not a blocker to sharing.
The simplified headlines alone led to a 6.8% conversion lift. The approach was soon adopted across all midcards.
Another early midcard.
Part 5: Experiments
Around this time, I began establishing and implementing design standards for the growing story midcard system.
Advocated for smaller header font size so media could take center stage and more characters fit per line
Consolidated privacy strings from 7+ variants down to 3, all fitting on a single line on the smallest devices — after months of negotiation with Legal
Added first-name personalization to headlines — highly effective for engagement, but eventually flagged as too click-baity at scale
Ran content tests on button copy; established "Share story" (ready-to-share) and "Create story" (requires input) as standards
Design and content standards implementation → 17.4% lift in publishing rates
We added the ability to share with one tap, but it required lengthy privacy text.
Subtle refinements to the design and content increased sharing by 17.4%.
What Sets Us Apart
People-First Approach
Everything we do is built around understanding your needs and helping you succeed—because when you thrive, so do we.
Long-Term Relationships
We’re not just here for the now. We love creating lasting relationships with our clients and growing with them over time.
Proven Process, Flexible Execution
We bring structure where it counts and adaptability where it matters. Our methods are clear, but always responsive.