The Iterate-or-Kill Call: Auditing Sula's Video Performance Export
Objective: Given a five-video TikTok performance export with 3-second retention, completion rate, shares, and saves, decide which videos the algorithm is rewarding, which are dying at the hook, and what to do about each.
You're the content lead at Sula Vineyards, India's largest listed wine producer (NSE: SULA), reviewing the first five TikTok videos from a new tasting-room content series.
Five videos, one export: views, 3-second retention %, completion %, shares, and saves. Decide which videos earned wider distribution and which never made it past the algorithm's initial test batch.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, sufficient for a five-to-twenty video weekly audit
The process
1 step
Step 01 of 01
The algorithm distributes a new video to a small test audience first. If that group watches to completion and engages, distribution widens; if they scroll away in the first three seconds, distribution halts. Watch time and completion rate outrank likes and comments as ranking signals.
Which of these five tasting-room videos got wider distribution because of what happens in the first three seconds, and which ones stalled?
Procedure
- Import the export: V1 (12,400 views, 71% 3-sec retention, 48% completion), V2 (890 views, 31% 3-sec retention, 9% completion), V3 (34,200 views, 68% 3-sec retention, 52% completion), V4 (1,100 views, 29% 3-sec retention, 11% completion), V5 (8,600 views, 66% 3-sec retention, 44% completion).
- Sort by 3-second retention, the strongest predictor of whether the video escaped the initial test batch.
- Flag any video under 40% 3-sec retention as a hook failure, regardless of its production quality.
- Cross-check completion rate against views to confirm the retention pattern matches the distribution outcome.
SULA TIKTOK PERFORMANCE AUDIT (5 videos) ESCAPED TEST BATCH (3-sec retention > 60%) V3: 34,200 views, 68% retention, 52% completion V1: 12,400 views, 71% retention, 48% completion V5: 8,600 views, 66% retention, 44% completion STALLED AT HOOK (3-sec retention < 40%) V2: 890 views, 31% retention, 9% completion V4: 1,100 views, 29% retention, 11% completion
Healthy
Videos with 60%+ 3-second retention correlate with 8-30x the view count of videos under 40% retention.
Unhealthy
Treating a low-view video's problem as a distribution-luck issue instead of a hook-quality issue visible in the first three seconds.
What this means
Views are a downstream effect, not the diagnostic signal. The 3-second retention number tells you whether the hook itself is the problem before the algorithm ever gets to test completion.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Two of five videos stall under 1,200 views with sub-40% 3-sec retention | Re-cut V2 and V4 to remove any brand intro or slow setup before the first 3 seconds, then re-post as new videos | half day |
| High-performing videos aren't being analyzed for what worked | Document the opening 3 seconds of V1, V3, and V5 as a hook template for the next batch | 30 min |
Final deliverable
A retention audit splitting the video slate into 'escaped test batch' and 'stalled at hook', with a re-cut recommendation for each stalled video.
See a reference example
UTKARSH SMALL FINANCE BANK TIKTOK AUDIT (excerpt) ESCAPED TEST BATCH V2: 21,000 views, 64% 3-sec retention, 41% completion STALLED AT HOOK V5: 640 views, 22% 3-sec retention, 6% completion Cause: opens with a 4-second branded logo animation before any content
Success criteria
You're done when you can:
- Correctly separates videos by 3-second retention threshold rather than raw view count
- Ties each stalled video to a specific hook-level fix, not a vague 'needs more views' note