The Problem
Usage data showed many users tried mobile social posting, but very few returned. The friction was structural: cross-posting required repetitive setup per platform, account management lived in settings instead of the creation flow, and platform nuances like character limits and tone were handled entirely by the user. What should have felt lightweight felt procedural.
Before redesigning anything, I audited native posting on Facebook, Instagram, and LinkedIn. Each platform optimized for immediacy. Our experience tried to consolidate cross-posting but introduced complexity instead. That gap pointed to three specific areas: how accounts were managed, how content was adapted, and how users handled not knowing what to say.
Rethinking Account Management
If a user realized mid-composition that they needed to adjust a connected account, they had to abandon their draft and navigate to settings. On mobile, that interruption reliably ended the session. Account management was treated as a configuration concern, which was architecturally accurate but wrong for how people actually used the product.
I moved account controls directly into the publishing workflow. Users can add, remove, or switch accounts without leaving the post. The draft stays intact, the context stays intact, and adjusting an account becomes a low-cost action.
Reducing Cross-Platform Friction
Sessions with small business owners made clear the struggle wasn't where to post. It was the time required to adapt content for each platform. Users were either posting the same copy everywhere or manually rewriting per platform and giving up partway through.
I shifted the design goal from distribution coverage to adaptation efficiency. The system adjusts a base message per platform automatically, adapting tone, formatting, and structure while keeping manual refinement available. Cross-posting became faster without removing user control.
Guiding Content Creation
A recurring pattern from usability sessions: users stalling because they didn't know what to say. Blank composer, cursor blinking, session abandoned. This was a consistent drop-off point visible in both session recordings and qualitative feedback from UXR.
I introduced contextual prompts tailored to the selected platform and business category. These evolved into lightweight AI-assisted recommendations that generate starting points without replacing the user's voice. The goal was to reduce the cost of beginning. Getting users past the blank state meant significantly more sessions ended in a publish.
Results
After multiple UXR testing rounds, the team launched to 20% of users. Two weeks in, no incidents and consistent positive signal across completion rates and qualitative feedback. We moved to 100% rollout within that window.
Post completion increased 144%. Feature adoption grew 81% in the first three months.
Results
Measured post-launch across UXR-validated rollout from 20% to 100% of users within two weeks.
+144%
Post completion rate
+81%
Feature adoption (first 3 months)
Reflection
The numbers came from fixing structural friction. Repositioning account management, building adaptation into the system, and reducing the cost of beginning all targeted moments where the workflow was working against users. The work was informed by close partnership with product, engineering, and UXR throughout.
The AI content layer opened up something worth continuing. Helping users past the blank state without scripting their voice is a narrow but meaningful target, especially as posting behavior diversifies across video, short-form, and platform-specific formats.