A single SKU needs at least 100 videos before GMV Max settles into consistent spend and holds a target ROI. Not 100 uploads in one week. A hundred pieces of creative the algorithm can test, spend against, exhaust and set aside over the life of the campaign.
Most Malaysian sellers arrive at that number the expensive way. They upload eight or ten videos, set a budget, watch two of them deliver, and conclude the product is wrong or the ads are broken. Neither is usually true. The campaign simply has nothing to work with.
Why the number is that high
GMV Max is not a bidding tool you point at an audience. It is an optimisation engine pointed at a pool of creative, and its job is to find the videos in that pool that produce revenue at the return you have asked for. If that is new, start with how GMV Max works.
Every video entering the pool goes through the same lifecycle. It sits in testing while the algorithm buys enough impressions to judge it. Most videos fail there and never deliver meaningfully. A small number clear testing and start spending properly. Those eventually fatigue – the audience that responds to them has seen them, performance decays, and spend moves elsewhere.
That lifecycle is the reason the pool has to be large. If ten videos go in and one clears testing, the campaign is running on one creative, and when that one fatigues the campaign stops. If a hundred go in, there is always something entering delivery as something else leaves it.
Given correct affiliate consent, content that meets TikTok’s policies and a budget large enough to test with, every video in the pool will deliver at some point. They will never all deliver at once, and that is the part sellers misread.
Delivering is the metric, not uploaded
The number that matters on the dashboard is how many videos are in delivering status, not how many you have uploaded. A seller with 70 videos and six delivering has a creative problem. A seller with 40 videos and 15 delivering is in a stronger position and should be increasing budget, not creative.
Watch that ratio week to week. If it climbs as you add videos, the pool is healthy and the constraint is supply. If it stays flat while uploads rise, the new videos resemble the ones already failing, and adding more of the same will not fix it.
This is also the honest answer to a campaign that appears to be working but barely spends. A budget that will not clear is usually a symptom of a thin pool rather than a targeting fault. We cover that separately in what to do when GMV Max is refusing to spend its budget.
Where 100 videos actually come from
No brand produces a hundred videos per SKU in-house at sensible cost. The pool is built from two sources.
The first is your own production – a structured shooting schedule producing platform-native short-form video on a repeating cycle. This is where the content standard is set, and it is worth getting the first twenty right because affiliates will copy what they see working. How much you need across the shop overall is a question of content volume rather than of ads.
The second, and the larger share for most shops, is affiliates. Every affiliate video is a piece of creative the campaign can use, produced at no upfront cost. A shop with forty active affiliates each posting two videos has eighty pieces of inventory it did not pay to make.
That reframes affiliate recruitment. Most Malaysian sellers treat affiliates as a sales channel and judge them on orders. They are better understood as a creative supply line, and judged on how much usable inventory they generate. A creator who produces four videos that clear testing is worth more than one who produced a single video that sold well and then went quiet.
One condition governs all of it: affiliate consent must be set up correctly, or the videos cannot be used in ads at all. That is a setup detail, not a strategy question, and it stops more campaigns than it should.
The budget it takes
A hundred videos need enough spend for the algorithm to judge them. Testing costs money whether or not the video works.
RM100 per day is the realistic minimum for a single SKU running properly – roughly RM3,000 a month. Below that, testing stretches out so far that creative fatigues before the campaign has learned anything useful from it, and you are paying for a learning phase that never ends.
Sellers do run on RM50 a day. It is a floor, not a plan, and it suits a shop testing whether the model works at all rather than one trying to scale. At that level, expect a long ramp and do not judge the product on the result.
When only a handful are delivering
If the pool is large and the delivering count is still low, the problem is upstream of the ads.
Check the obvious constraints first. Is the target ROI set so high that the algorithm cannot find anything meeting it? Is the listing itself weak – short description, too few photos, low buyer feedback – so traffic arrives and does not convert? Is the product page carrying enough social proof for a cold buyer to trust it?
A new listing with strong click-through and no sales is almost never an ads problem. The videos are doing their job. The product page is not.
What to take from this
Treat creative as inventory and count it accordingly. A hundred videos per SKU is the working minimum, the delivering ratio is the health metric, affiliates are the supply line, and RM100 a day is what it costs to test properly. A campaign that is short on any of those four will underperform regardless of how well it is optimised.
Getting there is a production and recruitment problem before it is an advertising one, which is why most shops stall at the first ten videos.
Common questions
Can I reuse videos across SKUs?
Only where the video genuinely features that product. Creative that does not match the listing it is selling performs badly and risks a listing issue.
How long does it take to build a pool of 100?
For most shops, three to four months with an active affiliate programme running alongside in-house production. Faster if the shop already has creator relationships.
Do older videos keep working?
They fatigue rather than expire. A video that stopped delivering months ago can return to spend if the campaign’s pool changes around it, which is another reason not to delete anything.
We run social commerce management for Malaysian brands, including the creative supply and affiliate recruitment that GMV Max depends on. Tell us the brand and the goal.