Media and social growth case study

AI Motion-Graphics Content Engine
5M+ plays, 35K+ followers

A content engine where every frame is generated from code instead of edited by hand, tuned by a retention analysis of the channel's own published data, and now measured at over 5 million plays across Facebook, Instagram, TikTok and YouTube.

Media & Social Growth Motion GraphicsContent AnalyticsRetention Modeling
5M+
Plays
across the published catalogue
35K+
Followers
across every platform
2M+
Best single short
one composition, one render
5.4x
Median reach lift
978 to 5,330 plays per short
Published on FacebookInstagramTikTokYouTubeTelegram
Industry
Media & Social Growth
Timeline
Ongoing in-house build, 88 shorts published to date
Outcome
5M+ plays, 35K+ followers
Result snapshot

5M+ plays, 35K+ followers

The format change lifted median reach per short 5.4x, from 978 plays to 5,330. The best single short has passed 2 million plays, the catalogue has crossed 5 million, and the channel grew past 35,000 followers across every platform. Because every frame is drawn in code, not one upload has ever been claimed or taken down.

A repeatable format with a measured pre-publish gate instead of guesswork
Median reach per short up 5.4x after the format change
Zero copyright exposure, because nothing is borrowed and every frame is generated
A catalogue that now works as a search and email asset, not only as a feed
/ The challenge

Where the bottleneck actually was

A football channel was posting reaction and hot-take clips across Facebook, Instagram and TikTok, and the results were flat: a median of 978 plays per post and a median of zero shares. Every video was cut by hand, so runtime and pacing changed from one upload to the next, and nobody could explain why one clip reached 200,000 people while the next reached 3,000. Using broadcast footage was also a permanent copyright risk on every single upload.

Reaction and hot-take posts earned a median of zero shares, so nothing compounded.
Every short was cut by hand, so published runtime and pacing were unpredictable.
Nobody could say why one video reached 200,000 people and the next reached 3,000.
Broadcast footage put a copyright claim risk on every upload.
/ What we built

A system built around the real workflow

We built a content engine rather than a content calendar. Every frame is rendered from a versioned code composition, so a short is a program, not a timeline, and no broadcast footage is ever touched. Then we closed the loop: a pull from the Meta Graph API snapshots plays, watch time, shares and the full second-by-second retention curve for every published short, and an analysis pipeline turns 80 settled videos into build rules. Those rules changed the format. Watch-through rate became a hard pre-publish gate, the first numbered step now lands on screen before second 3.5, the trailing summary beat was cut at source, and the topic mix moved from individual techniques to whole-team systems. The catalogue was then published as a crawlable library site, so the same diagrams also work as search and email assets instead of disappearing into a feed.

Module 01
Code-generated motion graphics: every frame rendered from a versioned composition, zero broadcast footage
Module 02
A Meta Graph API pull that snapshots plays, watch time, shares and full retention curves per short
Module 03
A retention analysis that reads each curve second by second and separates the two distinct failure modes
Module 04
Build rules fed back into the compositions: first numbered step before second 3.5, no dead tail, whole-team systems over isolated techniques
Module 05
A pre-publish quality gate on watch-through rate, the one signal that actually predicts reach
Module 06
A crawlable library site that republishes the same diagrams as static pages, with email capture and its own analytics
Build profile
Stack
RemotionReactMeta Graph APIPython analysisAstroSupabaseNetlify
Proof source
Naurra in-house build
Code-generated content engine with a measured feedback loop
Related pages
Next step

Want this outcome in your business?

We scope custom AI systems around the workflow that is already costing you time, margin, or speed.

Book a free consultation
/ More case studies