Case: A Sidaocai Special Store — RMB 4M Invested, Losing RMB 150k a Month

A single-store campaign inside the Sidaocai group. It answers the most uncomfortable question a chain brand faces: when a store clearly does not fit the brand’s standard model, do you save it or write it off?

Figures here are abstracted to what can be said publicly: investment magnitude and loss and profit magnitudes are published, while a full single-store P&L breakdown and structured data tables are not.

What made this store unusual

It was a street-side store in a tourism trade area, built with around RMB 4M.

That combination explains the problem. Tourism trade areas have distinctive traffic: strongly seasonal, concentrated in specific dayparts, and made up largely of visitors and families out for the day. Those characteristics differ sharply from the brand’s other stores — mall stores have steady daily footfall and office workers, neighbourhood stores have a base of regulars, while a street-side store in a tourism area receives traffic in large waves around holidays and peak season.

After opening, the store lost roughly RMB 150,000 a month for several consecutive months.

For the brand this was a choice: close it and stop the loss, or keep it and make it right. Keeping it meant continued investment — and if it was not made right, it would keep eating the profit of other stores.

Diagnosis: is the store wrong, or is the model wrong

There is an easy misstep here.

A store losing money month after month is most readily attributed to “a bad site”. That conclusion is sometimes correct, but more often it is too early — a bad site and a wrong model look different on a P&L, but in an owner’s mind they are usually one sentence.

Chen Yufu’s approach was to classify first. In Yufu’s method, “different single-store models” is an explicit term: stores in different locations and formats should have been matched with different single-store models from the start. So the first question was not whether the store was any good, but whether the model currently attached to it was the one it was ever meant to use.

If the store had been running the standard model copied from the brand’s other stores, then its losses might have nothing to do with the site — the problem could be using the wrong template. Those two conclusions point in opposite directions: a bad site means stop the loss, a wrong template means change the model.

The landing point for that judgement is the P&L — Yufu treats the P&L as a store’s weather gauge, read first for whether the loss sits in revenue structure, margin structure or fixed cost.

What Yufu did

The work on this store was a single-store campaign in its complete form:

Stage	What was done	What it settled
Reading the P&L	Checking monthly revenue and cost structure to locate the category of loss	Broke “bad site” from a vague conclusion into a testable judgement
Rebuilding the model	Redesigning the store model around the tourism area’s traffic profile	Replaced a template that did not hold instead of pouring more resources into it
Adjusting the product structure	Reassigning product roles and combinations around the new traffic profile	Let the store absorb peak season and not idle in the trough
Tracking and correcting	Reviewing monthly operating data and correcting the actions	Made the adjustment actually take effect rather than waiting for results

First came the P&L, not to calculate how much was lost but to locate the category. Losses have several quite different causes with almost opposite remedies; acting without classifying means the money leaves no trace.

Second came rebuilding the single-store model. This was the pivot of the case. The tourism-area street site’s traffic profile meant its model should not be a copy of a standard mall store — its dayparts, staffing and product mix had to follow the traffic, not the shape of other stores in the brand.

Third came continuous correction against monthly data. Changing a model is not a one-off act; once changed, operating data has to show whether it now matches the traffic.

Results

Within the programme, the store saw:

• after the model was rebuilt, consecutive monthly losses turned into profit;
• monthly profit at one point rose into the RMB 100,000 range.

That figure should be read where it sits: the store had been losing about RMB 150,000 a month. The change was therefore not “from zero to RMB 100,000” but from monthly bleeding to monthly positive contribution — a swing of roughly RMB 250,000 a month.

These are the store’s actual results within the programme period and are not a promise of operating results. Similar stores in different locations have different room to adjust — whether a break-even point can be brought down depends on the fixed cost structure, not on determination.

What this case really demonstrates

The thing worth taking here is that it shows one play per store is classification, not improvisation.

A chain brand builds a habit over time: once a model works, treat it as the right answer and apply it to every new store. Inside the model’s range of validity this is highly efficient — it turns opening a store into copying rather than starting a business again.

The danger in the habit is that a model has boundaries of validity, and those boundaries do not announce themselves. A tourism-area street store and a mall store show the same symptom on a P&L — a loss — but for entirely different reasons. Give the same prescription to the same symptom and you get: stores that needed a new model get written off, or stores that should have been closed get funded again.

That is the real value of the term “different single-store models”: it moves the classification act upstream. Decide which category the store falls into first, then decide which set of actions applies. It does not take long, and it decides whether everything spent afterwards is effective or not.

The order is worth noting too. This case started from the P&L, not the menu. The reason is that the menu is a product-level answer, while this store’s problem was at the model level — changing the menu when the model is wrong is optimising inside a wrong frame, and the best outcome is a slightly higher ceiling.

Frequently asked questions

How do you decide whether to save a loss-making store or close it? Look at whether the cause of loss can be changed by adjusting the model. If the fixed cost structure contains an immovable hard item — rent badly mismatched to the site, for instance — that is closer to a site problem. If the main variables are the model and product structure, there is room to adjust.

Why should a store model follow the dayparts of its traffic? Because a store’s costs occur over time while its revenue does not occur evenly. A tourism-area store staffed for average traffic will break down in peak season and lose money in the trough — the structure has to match the waveform of the revenue.

What does “special store” mean? It refers to stores whose site format differs from the brand’s main format and which cannot simply use the standard model. These need individual judgement rather than being folded into routine chain management.

What is the most direct lesson for a chain brand? Before classifying a loss-making store as “a bad site”, check whether it is running the right model. The check is cheap, and it avoids two expensive errors: saving the wrong store, or closing the right one.