As of November 12, 2026, China's largest shopping festival has a total and almost nothing else: 1.78 trillion yuan across the major platforms for the October 7 to November 11 window, a rise of 4.8 percent, and no headline gross merchandise value from either Alibaba's Taobao and Tmall or JD.com. For the category teams whose bonuses still hang on a topline, that is an uncomfortable combination. The number that replaced the old scoreboard is not another scoreboard. It is a set of category and engagement counters that each platform chooses, and reading them correctly has become the actual skill of the job.
Why the missing number changes the work
Gross merchandise value was a crude measure, but it was comparable across years and across companies, and its disappearance is not a formatting change. When the topline goes away, every public claim becomes a category claim, and a category claim is only meaningful if the reader knows the base. A statement that AI related product sales more than doubled says nothing about whether the appliance category grew, shrank or traded down, and a brand that mistakes one for the other will plan next year's range against a number that never described it.
The counters that replaced the topline
The replacement counters fall into three families. Volume counters report units or orders, such as JD.com's 50 million AI related products and 300 million service orders. Momentum counters report multiples against a base the platform does not always publish, such as transaction value for AI personal computers, workstations, cameras, glasses and recording devices more than doubling year on year. Attention counters report behaviour, such as searches for the word token rising more than thirteen fold from the preceding period.
| Counter family | Example from the 2026 campaign | What it can support | What it cannot support |
|---|---|---|---|
| Volume | 50 million AI related products sold on JD.com | A claim about units and breadth of range | Revenue, because unit price is missing |
| Momentum | AI glasses and AI cameras more than doubling | A claim about direction inside a niche | Category growth, because the base is unpublished |
| Attention | Token searches up more than thirteen fold | A claim about intent forming before purchase | Sales, because search does not equal an order |
| Aggregate | Syntun total of 1.78 trillion yuan | A claim about the whole market | A brand's own share, unless it has the same cut |
The practical consequence is that a brand now has to build its own denominator. If the platform will not publish a category total, the brand can still assemble one from its own order data plus the third party panel it already buys, and it can hold that series stable across campaigns. A series that changes definition every November is worse than no series at all, because it produces confident conclusions from incompatible numbers.
What the 4.8 percent actually contains
Two readings of the same growth rate are circulating. Syntun's estimate covers a five week window and reports a 4.8 percent rise against a year earlier, while an earlier cut covering October 7 to October 31 was reported at 2.4 percent. Neither is wrong; they measure different spans of a campaign that no longer happens on a single day. The wider lesson is in the comparison base, since last year's festival grew 14.2 percent, and a market that decelerates from that pace while extending its window is telling you that the promotion is carrying more days rather than more demand.
The macro backdrop explains the deceleration without explaining it away. Retail sales of consumer goods rose 1.1 percent in the first nine months of 2026 while the economy grew 4.6 percent, and a long property downturn has left households more cautious about discretionary purchases. Inside that environment, a 4.8 percent festival is a reasonably strong result, and the interesting question is not why it was not 14 percent but which categories paid for the growth that did happen, and whether what looks like category growth is demand moving forward rather than demand being added.
Which categories carried the campaign
Two groups did the work. AI hardware is the obvious one, with transaction value for AI personal computers, workstations, cameras, glasses and recording devices more than doubling, and token packages for large language models rising more than tenfold in sales. The second group is newer consumer brands: Alibaba reported that 333 brands launched within the last three years ranked first in their specialised categories between October 15 and October 27, including an AI companion robot maker and a smart sleep company.
For a category manager the signal is about entry price rather than about technology. AI hardware doubled from a small base, which means the absolute volume is still modest and the shelf space is still being allocated. New brands ranking first in narrow categories is the same story from the supply side: when a category is young, a single well specified product can take the top slot, and when it matures the slot goes to whoever can fund the promotion calendar. Deciding which of the two situations a category is in, before committing a launch budget, is the difference between riding a wave and paying for one.
How to report the next campaign internally
The reporting template is the part a team can fix immediately. Instead of a single number that no platform will confirm, the report should carry three lines that a finance team can tie back to invoices, and one line that records what the platform claimed, kept separate so the two are never averaged together.
FAQ
Is the 1.78 trillion yuan figure an audited total?
No. It is a third party estimate from Syntun covering the major platforms, and it has become the benchmark precisely because Alibaba and JD.com stopped publishing a headline figure. It should be cited with the provider and the window attached.
Why does the same campaign show 2.4 percent and 4.8 percent growth?
Because the two figures cover different spans. The 2.4 percent reading covered October 7 to October 31, while the 4.8 percent reading covers the full window to November 11. A longer window includes the final subsidy peak.
What does category growth mean when a platform reports multiples?
Very little on its own. A multiple needs a base, and the base is usually unpublished. Category growth is only usable when the brand can reconstruct the same denominator from its own data and a stable panel.
Should a brand chase AI hardware listings?
Only with a range decision attached. The category is growing from a small base with modest absolute volume, so the risk is listing many SKUs into shelf space that has not formed yet. Start with two or three specified products and a clear return policy.
