case study
one pipeline, four use cases, real numbers
how dropspace publishes across 4 use cases, 11 social accounts, and 6 platforms — with full engagement data.
the setup
i'm a solo founder building an app studio — one app every two months. each project has its own social presence, its own voice, its own content strategy.
without automation, maintaining 11 social accounts across 6 platforms would be a full-time job. instead, one ai agent handles all of it through the dropspace api — the same api available to every user.
here's the real data from 7 weeks of production usage. not a demo, not a projection. actual numbers from actual posts.
259
posts
26,653
impressions
210+
reactions
41
comments
11
accounts
6
platforms
since january 28, 2026 · total cost: ~$172 ($3.50/day)
the four use cases
Dropspace
saas for multi-platform content distribution. the agent runs brand marketing — facebook posts, linkedin updates, twitter posts, reddit posts, and ai-generated instagram and tiktok slideshows. fully autonomous, no human in the loop.
230
posts
15,711
impressions
0.2–0.4%
engagement
lowest engagement but zero effort. 86 tiktok posts account for 12,416 of the impressions. facebook: 24 posts, 0 impressions — effectively dead.
JCLVSH
build-in-public content for a solo founder building an app studio. the agent drafts posts about the apps being built, sends to slack for review. josh approves or edits before anything goes out.
14
posts
8,710
impressions
~1.2–2.2%
engagement
linkedin averaged 998 impressions per post with 9.9 reactions each — 14x twitter per-post performance. "575 commits and $0 in revenue" got 3,840 linkedin impressions alone.
ynho
dj set clips from josh's music project. the pipeline identifies transitions between songs in live dj sets, cuts 30-second clips, and captions with artist credits. josh picks which clips to post and manually uploads to tiktok with trending sounds.
10
posts
1,897
impressions
4.9–5.4%
engagement
highest engagement across all use cases. tiktok clips with original sound averaged 190 views and 5.4% engagement — 27x higher than fully autonomous content.
Studio Jams
a music community in nyc that hosts concerts, open mics, and workshops. the pipeline takes event photos from a google drive folder and turns them into tiktok slideshows with face-aware text overlays — fully automated.
5
posts
1,355
impressions
1.4%
engagement
consistent 208–297 views per post. real event photos from real events — not ai-generated. automated from google drive folder to tiktok with no manual steps.
the engagement pattern
the data makes a clear argument: human curation drives engagement. automation drives scale. the right mix depends on what you're optimizing for.
| mode | use case | platform | engagement | effort |
|---|---|---|---|---|
| human-curated | ynho | Instagram, TikTok | 5.4% | high |
| human-approved | JCLVSH | LinkedIn, Twitter | ~2.2% | medium |
| automated | Studio Jams | TikTok | 1.4% | low |
| autonomous | Dropspace | Facebook, LinkedIn, Twitter, Reddit, Instagram, TikTok | 0.2–0.4% | none |
human-curated content (ynho) gets 5.4% engagement. fully autonomous content gets 0.2%. that's a 27x difference. but autonomous content requires zero ongoing effort.
key insights
linkedin is 14x twitter
for the jclvsh account, linkedin posts averaged 998 impressions per post with 9.9 reactions. twitter averaged 145 impressions per post for the same content. if you're a dev tool and only posting on twitter, you're leaving reach on the table.
facebook is effectively dead
24 posts on facebook. 0 impressions. 0 engagement. it costs nothing to keep running, but the data is clear.
total cost: $172 over 7 weeks
$3.50/day in ai image generation (fal.ai at $0.08/image). the api calls run through dropspace's own infrastructure. 259 posts across 6 platforms for the cost of a coffee per day.
the agent handles infrastructure. you handle taste.
the pattern across all four use cases: automation handles scheduling, formatting, posting, and analytics. humans decide what's worth amplifying. that division of labor is where the leverage is.
try it yourself
this is the actual setup running in production. the same api, the same mcp server, the same webhooks.