Data Update Cadence
| Data Type | Update Frequency | Notes |
|---|---|---|
| Profiles | Daily | Updated via ETL when creator activity changes |
| Content | Real-time | New content appears as it’s created |
| Analysis | Periodic (~30 days) | AI analysis is regenerated periodically |
| Audience | Periodic (~30 days) | Based on recent content and engagement patterns |
Timestamps
Profile Timestamps
updated_at: When the profile data was last refreshed in our systemlast_active_at: When the creator last posted content (if known)
Audience Analysis Timestamp
meta.last_updated_at: When the audience analysis was last generated
Recommended Cache TTLs
| Data Type | Recommended TTL | Reason |
|---|---|---|
| Profiles | 24 hours | Updates daily via ETL |
| Content (single) | 1 hour | View/engagement stats may update |
| Content (lists) | No cache | Real-time additions |
| Analysis | Weeks to months | Regenerated infrequently (typically monthly) |
| Audience | Until last_updated_at changes | Regenerated infrequently (typically monthly) |
Best Practices
- Cache profiles for 24 hours: Profile data typically updates daily
- Cache analysis/audience aggressively: This data is regenerated infrequently (typically monthly)
- Don’t cache content lists: New content is added in real-time
- Check
last_updated_at: For audience data, cache until this timestamp changes - Refresh on demand: Only re-fetch when needed, not on every request