LIVE GMV Max wants 50-70 videos in queue: how to supply them
TikTok says add at least 50-70 videos in the In queue status for stable LIVE GMV Max exploration. One Sume bulk queue can cover it; here is the order of work.

TikTok's LIVE GMV Max best-practices page says to add at least 50-70 videos in the In queue status for more stable creative exploration (read 2026-10-03). A single Sume bulk queue takes up to 100 Format runs, so one queue can produce the whole supply, if you budget for failures and for the videos you reject.
Sume makes the clips. Moving them into the In queue status, and keeping them there, is done in TikTok's tools.
What does TikTok actually say about supply?
The page names two creative types: Video-to-LIVE, which uses the account's existing videos as promotional content, and LIVE-to-LIVE, which uses the active livestream. Only the first is a video you can prepare ahead of time.
- The page does not give duration, resolution or exploration percentages.
- More high-quality videos help the campaign identify top creatives.
- Check the status of each video before a live session.
| Item | Page says | Plan with Sume |
|---|---|---|
| Videos In queue | At least 50-70 for stable exploration | Queue 80-100 so rejects still leave 70 |
| Budget use | A modest portion of campaign budget keeps testing new videos | Refill the queue as videos drop out |
| Video-to-LIVE | Uses the account's existing videos | Source for AI clips |
| LIVE-to-LIVE | Uses the active livestream | Not prepared with Sume |
How do I queue the supply?
The number 70 is TikTok's, and the over-ask is mine: a bulk queue of 100 with a spend cap per item leaves room for clips you cut in review. The code below creates the queue and prints the poll address; it assumes your products are in a list and that your key carries formats:write.
The queue has no webhook. Poll status_url and branch on counts.
import os, uuid, requests
H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}",
"Idempotency-Key": str(uuid.uuid4())}
products = [f"SKU {n}" for n in range(1, 91)]
items = [{"instruction": f"12 second vertical demo for {p}",
"generation_spend_cap_usd": 10} for p in products]
r = requests.post(
"https://api.sume.com/v1/formats/sume/sume-product-usage-demo/bulk-runs",
headers=H, json={"concurrency": 8, "items": items}, timeout=60)
r.raise_for_status()
q = r.json()["data"]
print(q["id"], q["status_url"])What do I do when the queue finishes?
When the queue reads completed, read counts.failed before you count finished clips. Requeue only failed rows with a new Idempotency-Key; rerunning the whole list pays for the successes twice.
Then review. A clip with a wrong product shape or an on-screen claim you cannot back up should not go into the queue, and TikTok's policy checks stay on your side.
What Sume does not do
Sume has no connection to LIVE GMV Max. It cannot see how many videos are In queue, refill them, or read exploration results. Keep a spreadsheet of SKU, hook id and file name and compare it with the status list in TikTok's tools.
Sources
Related posts
More in Use cases
- LIVE GMV Max Creative Boost: $10 minimum, 24h default, prep clips
Creative Boost in LIVE GMV Max needs allowlisting, In queue or Not delivering videos and a $10 daily minimum. How to prepare variant clips with Sume first.
- Partnership ad code or account-level permission: which to ask for
Meta partnership ads come in two permission types. How content-level codes and account-level access differ, and what to log when the clips are AI-made.
- Pinterest Q5, the week after Christmas: plan creative in October
Pinterest calls the week between Christmas and New Year's Q5 and advises testing creative at Halloween. A Sume bulk-run plan, with Pinterest's figures dated.
- Pinterest self-serve A/B testing: build creative variants with Sume
Pinterest added self-serve A/B testing for creative, targeting and Performance+ on Sept 17, 2026. Build clean creative variants, one change each, with Sume.
Written by Sume