• Hardware Specifications & Operating System:
Windows 11 Home 64-bit (10.0.26200)
CPU: Intel Core i5-10400F @ 2.90 GHz
RAM: 32 GB
GPU: NVIDIA GeForce RTX 3060
NVIDIA Driver: 616.56 (August 26, 2026)
DirectX 12

Additional Restaurant Testing

I have continued testing the restaurant business after my previous report, and I would like to share some additional observations from almost two weeks of gameplay.

Restaurant Business – Additional Testing and Observations

I have continued testing the restaurant business after my previous report, and I would like to share some additional observations from almost two weeks of gameplay.

The most important conclusion from our testing is that the biggest remaining problem does not appear to be the number of employees, but the autonomous behavior of customers and the unreliable progression of the dining process.

Employee testing

At first, we tested the restaurant with regular autonomous employees. In this setup, the restaurant often did not function properly. Employees could switch between roles, take over each other’s tasks, abandon tasks, stand idle or perform unrelated activities while customers were waiting.

We then employed two Zois from the same household, which allowed us to control them directly. We played with this setup for almost two weeks. The restaurant became noticeably easier to manage when we could directly control the employees and intervene when necessary.

This was an important difference.

However, we also tested the restaurant with the owner working almost entirely alone.

In the latest test, six customers arrived gradually during the day. The owner was able to cook and serve them without significant problems. Another employee was present for approximately one hour, but her help did not appear necessary for the actual workload.

The restaurant finished that day with a +M134 profit.

This suggests that simply adding more employees does not necessarily solve the main problem.

Customer behavior

The biggest problem remained the customers.

Customers can arrive at the restaurant and sit down, but they do not always continue normally through the dining process.

During our testing, we observed customers who:

• remained at the table for a very long time without ordering,
• did not proceed to eating even when food and staff were available,
• interrupted or abandoned the expected dining sequence,
• left the restaurant without paying.

In our latest test, one customer remained in the restaurant for approximately three hours without completing the meal process. Eventually, the restaurant closed and the customer left without paying.

At the same time, the owner was available and capable of serving the customers.

This makes us believe that the main issue is not necessarily the employee’s ability to perform the task, but the customer’s autonomous behavior and the transition between the individual stages of the restaurant visit.

The expected sequence is:

arrival → seating → ordering → food preparation → serving → eating → payment → departure

This sequence is still not reliable.

Ordering problem

Ordering itself is another important problem we observed.

When a customer is ready to order, an order request appears above the customer. However, the request can disappear before the server reaches the table. The customer then gets up and starts walking around, and the server has to try to reach the customer again.

We observed this happening repeatedly, sometimes up to eight times with the same customer. In some cases, I had to control the server manually just to make sure they reached the customer and could accept the order.

We suggest that the order request should remain active for a longer period, preferably until a server reaches the customer and accepts the order.

Another possible solution would be for the server to remain at the table until the customer has successfully placed the order.

This would prevent the customer from repeatedly losing the opportunity to order simply because the server did not reach the table quickly enough.

Financial results

The financial results also show how strongly this affects the business.

Earlier test days with two controllable employees still produced significant losses, for example:

• −M459
• −M272
• −M210
• −M61

These results were with only two employees. If all of the employees listed in the restaurant were employed, the wage costs would be higher and the resulting losses could therefore be even greater.

In contrast, the latest test, with the owner operating almost entirely alone, ended with +M134.

The current seven-day statistics show:

• Sales: M3,602
• Expenses: M3,974
• Overall result: −M372
• Visited customers: 8

The latest episode was nevertheless rated positively by the game:

Service speed: +265

“They helped me as soon as I needed it. I’ll come back!”

This is interesting because during the same period we observed a customer who remained in the restaurant for several hours and ultimately left without paying.

Conclusion

Based on our testing, we do not think that simply increasing the number of employees is the main solution.

The restaurant can be operated by one owner when the workload is small enough. Directly controllable employees also make the restaurant considerably easier to manage.

However, the most persistent problem we observed is the autonomous behavior of customers and the unreliable progression from one stage of the dining process to the next.

More employees can increase operating costs without necessarily solving this problem.

We believe that improving the customer flow — especially ordering, eating and completing payment — would have a much greater impact on restaurant performance than simply adding more employees.

We will continue testing the restaurant and observing whether this behavior remains consistent.

For our earlier, more detailed testing and suggestions regarding the restaurant system, please also see our original post here:
[Restaurant Business – Testing, Problems and Suggestions] Restaurant Business – Testing, Problems and Suggestions