The Snack Chain Has 26,405 Stores. It Owns 9.

The receipt said 0.299 kg

Two Chinese snack chains, both listed companies, run almost all of their stores through franchisees. In August, customers at several of those stores found their receipts showed far more weight than they had bought. Both companies apologised on September 7 and pledged to refund ten times the difference on any mis-weigh. Regulators have so far penalised individual stores, not the listed companies. This is the story of how a RMB 64.58 receipt connects to a HK$95 billion IPO.

On August 10, 2026, in Qing County, Cangzhou, Hebei province, a customer bought four pieces of beef jerky at a Zhao Yiming Snacks store.

The receipt said 0.299 kg, RMB 64.58.

At home, the same four pieces weighed 0.08 kg. At the store’s own price per kilo, that is RMB 17.29. The receipt was RMB 47.29 too high, 3.7 times what the jerky was worth.

The store manager’s reply, as reported in Chinese media: the camera showed nothing wrong and the scale was fine, they did not know how the system had read it, and a mistake was a mistake, they would own it. The store offered to refund three times the whole bill. The customer declined.

The next day, the Qing County market regulator opened an investigation. A day after that, Zhao Yiming’s head office responded that it may have been an employee error and that the company would step up self-inspection.

The next day, at a Haoxianglai store in Zhoukou, Henan, a customer was charged RMB 111.35, asked for a re-weigh and paid RMB 64.80. On August 17, at a Haoxianglai in Wusheng County, Guang’an, Sichuan, a customer paid RMB 111.71 and got a re-weigh of just over RMB 60; nine customers were involved and the county market regulator sealed the scale.

These stores belong to two listed companies. Zhao Yiming is a brand of Hong Kong-listed Mingming Henmang (1768.HK). Haoxianglai belongs to Shenzhen-listed Wanchen Group (300972.SZ). As of June 30, 2026, the first had 26,405 stores and the second 23,802, according to their interim results — 50,207 between them.

By mid-August the market regulator in Funing County, Yancheng, Jiangsu, was already checking snack stores in its district. In early September, regulators in Baotou in Inner Mongolia, Yuncheng in Shanxi and elsewhere published the results of their own checks, made with calibration weights in hand; Baotou’s notice came on September 4.

On September 7, both companies apologised on the same day, with the same promise: if a store weighs something wrong, it pays ten times the difference.

Mingming Henmang’s shares went from HK$444.80 at the close on August 28 to HK$361.80 on September 16. With 217.8 million shares outstanding after the over-allotment (214.1 million on listing day), that is a market value of HK$96.9 billion falling to HK$78.8 billion — HK$18.1 billion less in thirteen trading days.

Not all of that belongs to the scale. August 28 was already a post-results high, HK$7.8 billion of the drop had happened before Baotou’s notice on September 4, and on September 14 the stock bounced 7.2% in a day after a foreign bank reiterated a buy rating. Wanchen Group published interim results on the evening of August 28, with profit more than doubling, and its shares fell 7.9% the next trading day anyway. The market was digesting more than one thing.

But a RMB 64.58 receipt is enough reason to go back and look at how a HK$96.9 billion company was put together. That, on its own, is strange.

The shares opened up 88.08%

Rewind to January 28, 2026.

Mingming Henmang listed in Hong Kong at HK$236.60 a share and opened at HK$445.00, up 88.08%, for an opening market value of HK$95.28 billion. The Hong Kong public tranche was oversubscribed 1,899 times.

Chairman Yan Zhou held 24.06%, which Chinese financial media valued at more than HK$21.7 billion at that day’s price. Vice-chairman Zhao Ding is the second-largest individual holder. Both were born after 1985.

Its largest outside shareholder is HongShan, formerly Sequoia China. A HongShan partner said that day that Yan Zhou had a distinctive “retail aesthetic” and had turned snack shops into containers for emotion. Financial media shortened that into a headline: a victory of retail aesthetics.

The story goes like this. Yan Zhou opened a 40-square-metre snack shop under a bridge in Changsha in 2016. Zhao Ding, from a family of roasted-nut sellers, opened Zhao Yiming in Jiangxi in 2015. The two chains fought street by street across Hunan and Jiangxi; one pedestrian street in Changsha had five of their stores on it. In November 2023, in a private room at a Japanese restaurant in Panyu, Guangzhou, the two men agreed to merge in about two minutes, sixty-forty.

At the merger they had roughly 6,500 stores between them. Two and a half years later, 26,405.

All of that is true.

Third-party research cited in the company’s prospectus put its prices, for products of equivalent quality, about 25% below offline supermarkets on average. It works with more than 2,500 factories and runs 56 warehouses; stores are usually within 300 km of the nearest one, with delivery inside 24 hours. In the prospectus, inventory turned in 11.6 days in 2024 and warehousing and logistics cost 1.7% of sales; in the first half of this year the turn was a little faster, 11.2 days.

Those numbers cannot be faked. The cheap is real and the fast is real.

Here is where the causation got reversed. “Cheap” was credited to the founders’ insight and taste. The store network that makes cheap possible was mostly paid for by franchisees.

99.2% of revenue comes from franchisees

In 2025, Mingming Henmang reported revenue of RMB 66.17 billion. Of that, RMB 65.66 billion was sales of goods to franchisees — 99.2%. Franchise service fees were RMB 510 million, 0.8%. Management has said it has largely stopped charging franchise fees at all.

In other words, almost all of the company’s revenue is wholesale to its own franchisees.

At the end of 2025 it had 21,948 stores, of which 21,927 were franchised and 21 company-owned. By June 30, 2026, of 26,405 stores, 9 were company-owned.

It had 10,327 franchisees, averaging just over two stores each. It opened 7,813 stores in 2025, 21.4 a day.

One brokerage model puts the cost of a store at RMB 550,000 to 600,000, with a blended gross margin around 18%, a net margin of 8% to 10%, and payback in eighteen months to two years.

Put the two sets of numbers side by side.

At headquarters, gross margin was 9.8% in 2025 and 11.5% in the first half of this year, earned on the wholesale spread. At the store, on the brokerage model, gross margin is 18%, and rent, wages and utilities all come out of that 18%.

The same goods, from factory to your hand, have their margin cut into two layers, and both are thin. Headquarters multiplies its thin layer by scale into RMB 2.33 billion of net profit. The store has nothing to multiply by.

This year’s numbers point in one direction. In the first half of 2026, store gross merchandise value rose 55.6% on the company’s figures, while one brokerage, dividing head-office revenue by store count, put monthly revenue per store at RMB 310,000, up 3.2% year on year. Management said at its results briefing that payback now runs around two years, according to Chinese media; three or four years earlier, the company’s own franchise-recruitment materials had said 12 to 18 months. One franchisee in Nanchang found another store from the same group 700 metres away.

Put another way: the capital behind the ten thousand stores came mostly from franchisees, each putting up half a million yuan or more.

The franchisees have not walked away, either. In the first half of 2026 the chain closed only 121 stores, its franchisee count grew by 655 to 10,982, and on July 20 it said signed stores had passed 30,000.

Headquarters’ customer is the franchisee. You are the franchisee’s customer.

The scale sits at the thinnest point between those two layers of margin. That does not mean anyone is weighing heavy on purpose. It means this is the spot where errors are easiest to make and hardest for headquarters to see.

The scales are mostly fine

The results of the inspections are coming in. Start with the strictest one.

In late August the Zhejiang Consumer Council sent staff into 22 stores across eight brands in Hangzhou, posing as ordinary shoppers, and had the samples re-weighed by a licensed third-party metrology lab. Twenty-one stores passed, including every Zhao Yiming and Haoxianglai store in the sample. The one store it named belonged to another brand, where beef slices labelled 240 g weighed 234.3 g — 5.7 g short, about RMB 2.05 overcharged. Three items together came to about RMB 3.75.

The regulators’ numbers look similar. Baotou checked 123 stores and 284 scales and found one that had not been certified. Huangshi checked 13 stores and found no short weight. Longhua district in Shenzhen checked 43 stores and found one scale not certified as required. Lingwu in Ningxia found five scales uncertified or expired and one with a forged certification mark.

So far, every penalty has gone to an individual store. Neither listed company has been subject to administrative punishment.

So the scales themselves are mostly fine. That was also Zhao Yiming’s position on August 12: every store scale has to be inspected by the market regulator and carry a certification label, it said, so faking one is not possible.

That leaves the question of where 0.299 kg came from.

Every explanation the companies gave landed on people. In Cangzhou, later reporting relayed the investigation’s conclusion as improper operation by staff. For Zhoukou, Wanchen Group said an employee had made an error: the till had been restarted and not zeroed.

One number explains why people make mistakes here. As of the end of September 2025, about 38% of Mingming Henmang’s products were sold loose by weight, according to Chinese media citing company data. A store carries at least 1,800 products. Reporting on the chains describes no dedicated cashiers, staff rotating across every job, high turnover, and new hires needing time to learn the barcodes and pricing of thousands of items.

A scale that passed its annual check, a new hire who rotates through every job, and 38% of the goods going over that scale. Whether it errs, and in which direction, is not something the person on the customer’s side of the counter can see.

This can only be inference: the errors most likely go both ways. Only the ones that overcharge become complaints.

One researcher quoted in Chinese media put it more bluntly: measurement errors at multiple stores at once cannot simply be written off as random operator mistakes. The reverse also holds: the inspection results so far are not enough to call them deliberate, either.

The same researcher added a line that goes to the heart of this piece: in an asset-light franchise system, the brand owner’s responsibility for measurement compliance cannot be pushed entirely onto franchisees.

What actually tells you something is three becoming ten.

The industry norm was to refund three times the difference on a mis-weigh. On September 7, both companies said ten, and put the scales in every store under central monitoring. As I read it, more than tripling the payout says one thing: headquarters cannot watch the scales in 26,405 stores on its own. What is left is to let customers weigh things themselves, and pay handsomely when they catch something.

A calibration weight next to the scale

One item in the remediation plan: every store gets a standard calibration weight, checks the scale before opening each day, and sends every weighing record back to headquarters.

The regulator is moving further. A draft amendment to the rules on weighing retail goods, published for comment on May 12 by China’s State Administration for Market Regulation, says: when goods are weighed in front of the customer, the unit, the process and the reading must be shown, and a customer who re-checks and finds short weight may claim compensation. On September 7 the same regulator launched a pilot in Hebei, Zhejiang and Fujian for price-computing scales that lock themselves if the casing is opened; from January 8, 2027, scales that do not meet the new standard may no longer be sold. Hebei is the province Cangzhou is in.

So the company’s plan and the regulator’s draft land on the same person: the one standing in front of the scale. You.

Mingming Henmang’s chief financial officer, discussing expansion and operations at the annual results briefing, said that for years the company had put its energy into opening stores fast, and its operations had not kept pace with its expansion, according to Chinese media.

My reading of that: the speed of opening is set by franchisees’ money, 21.4 stores a day. The speed of operations is set by headquarters’ people. The gap between the two is where a 0.299 kg receipt can appear.

Cheap by 25% is real. 26,405 stores is real. The 88.08% opening day is real.

It is just that headquarters supplies the scale and does not employ the person operating it. The scale is in a franchisee’s store, at the elbow of a staff member who rotates through every job, and now there is a calibration weight beside it.

Before you pay, you can ask for it to be weighed again.