Social Campaign Engine
Ninety days of marketing that runs itself.
A full product marketing campaign — hundreds of posts, three a day, across LinkedIn and Facebook — generated up front, reviewed and approved as a batch, then published on schedule by a machine. One switch pauses everything.
Write the quarter once, then let it fire
Instead of a person improvising content every morning, the entire 90-day arc was generated as a coherent campaign — technical posts, commercial posts, and personal-voice posts in a deliberate mix — then approved as a body of work before a single one went out.
📝270 posts, one arc
Three posts a day at morning, midday, and evening slots, sequenced from technical depth to commercial pitch to personal voice — a campaign with a shape, not a random feed.
📡Three channels, one queue
Each post cross-publishes identically to LinkedIn and two Facebook pages — 810 total sends from a single approved queue, with the live post URL recorded per send.
✅Approval is the gate
The scheduler only ever publishes items that are explicitly approved and past their scheduled time. Draft content physically cannot leak — unapproved means unpublishable.
A scheduler you can trust unattended
The engine is deliberately boring in the best way: a tracked queue, a publisher that checks in every ten minutes, and pacing limits so it never behaves like a bot storm. It survives reboots, catches up gracefully after downtime, and logs every action.
- Queue states move one way: pending → approved → posted, with errors surfaced, never swallowed
- Per-run send cap and a spacing gap between posts — polite to the platforms, safe against runaway loops
- Missed slots late-post automatically when the machine wakes — downtime shifts the schedule, it doesn't drop content
- Channel connections are health-checked before launch, not discovered as silent failures mid-campaign
- Pause everything with one command; pull any single post back out of the queue without touching the rest
By the numbers
Marketing that survives a fact-check
Before launch, every post was swept against the platform's own verified claims register — the same verify-before-document discipline used across Synergy engineering, applied to marketing copy.
🔍Claims sanitized pre-launch
A deep technical audit falsified two claims the drafts had picked up — a stale benchmark figure and an overstated audit-log property. Every affected post was corrected, and one built entirely on the bad claim was regenerated on a true topic.
🧱The fact sheet is the source
The generator writes from a hardened fact sheet, so future content inherits the corrections — the campaign can't quietly drift back into claims the engineering team already disproved.
The rule underneath: an automated voice publishing on your behalf must be held to a higher truth bar than a human one, because it will repeat a mistake 270 times without blinking. Verify first, then schedule.
Set-and-forget schedulerApproval-gated queuePublishes to LinkedInPublishes to Facebook PagesChecks the queue every 10 minEvery claim fact-checked