虎嗅

The Secret Weapon That Saved Hulan Flowering Club

原文:呼兰拯救开花俱乐部的秘密武器

Summary of Key Points

Comedian Hulan temporarily took on the role of CEO for Shenzhen Kaihua Club from May to September 2026. By reorganizing the product system and leveraging AI tools, he addressed issues such as monthly losses of 120,000 yuan, a low audience turnout, and inefficient operations. By the end of June, the club's losses had been reduced to 20,000 yuan, and it was expected to break even in July. His core approach was to focus on creating a sustainable long-term operation rather than relying on short-term celebrity effects. AI was used to handle the tedious and time-consuming aspects of management, while product improvements were made to enhance the audience experience, ultimately enabling the club to become self-sustainable.

I. The State of Kaihua Club When Hulan Took Over

The situation at Kaihua Club was truly dire:

  • Financial Strains: The club was losing 120,000 yuan per month, and it was struggling to pay the salaries of its six employees.
  • Low Audience Attendance: A theater with a capacity of 300 people had only about ten attendees during open mic sessions. On Hulan's first performance, there were fewer than twenty viewers, leading to awkward silences during his jokes.
  • Inefficient Operations: Ticketing data from multiple platforms had to be manually processed, taking 5–10 minutes to update.
  • Unpredictable Products: The club mainly relied on random open mic sessions where audiences didn't know who would perform, and there were occasional special performances by actors from other cities, resulting in a lack of repeat business.

Hulan initially wanted to find a place to practice his jokes, but upon seeing the club's situation, he realized that if nothing was done, it might disappear. After all, stand-up comedy relies on a lively atmosphere; fewer viewers meant less feedback for the actors, which in turn led to worse performances and even fewer attendees, creating a vicious cycle.

II. AI Isn't for Writing Jokes, but for Handling Menial Tasks

Hulan, with a background as a programmer (Master of Actuarial Science from Columbia University and former CTO), knew that while AI couldn't write great jokes (jokes require personal experience), it could solve operational problems:

  • Stage Design: Using Alibaba's Accio Work, he inputted parameters like a budget of several thousand yuan, black and yellow color scheme, and "Kaihua Club," and the AI generated design drawings, listed required materials, helped find nearby manufacturers, and provided pricing information, saving employees time.
  • Data Dashboard: By integrating past performer registration data and ticketing information from various platforms, a real-time dashboard was created in half an hour, eliminating the need for manual calculations of sales figures.
  • Team Collaboration: The six employees used the same AI system to access financial, ticketing, and promotional data, allowing them to quickly identify which performances were losing money and which actors were popular, avoiding duplicate work.

Although AI didn't make the club immediately successful, it freed the employees from tedious tasks so they could focus on performing.

III. Products Are the Key: Hulan Reformed the Club's Performance Content

Hulan believed that without good products, even the best AI would be useless. He made three major adjustments to the club's offerings:

1. Reducing Open Mic Sessions: From one daily session to two weekly sessions, as audiences didn't want to pay for the uncertainty.

2. **Upgrading the "Golden Joke Contest": The contest was renamed "Divine Golden Joke Contest" and included interactive elements like audience voting (similar to offline variety shows). Hulan purchased 18,000 voting devices, promising a return on investment if they sold 100 more tickets per event. The prize for the winner increased to 1,000 yuan, and the runner-up received 400 yuan. There was also a rule that the same joke could only be performed three times to prevent actors from being lazy.

3. Localizing the Content: They launched "Thursday Cantonese Night," with Jiang Zihao, who spoke Cantonese fluently, in charge of organizing events to attract local Cantonese audiences.

These changes transformed the club's performances from random events into more engaging and predictable experiences, encouraging audience repeat business.

IV. Challenges in Implementing Changes

The biggest obstacles were human-related:

  • Cost Disputes: Employees objected to buying voting devices (60 yuan each), but Hulan convinced them after a long discussion. Later, when audiences complained that the devices lacked backlighting, Hulan solved this issue by using fluorescent stickers on the buttons.
  • Sustainability of AI Usage: Employees were concerned about the cost of continuing to use AI after Hulan left, but Hulan saw this as a sign that they had become dependent on it, which was a positive development. He promised to find ways to reduce costs in the future.

Hulan's principle was to spend money where necessary and avoid waste wherever possible.

V. Hulan's Long-Term Goals

His role as CEO was a temporary experiment. He didn't want to rely on his personal fame to attract audiences (e.g., by organizing special events) because he wanted to establish a system that could function independently. His goals were:

  • Standardized Products: To create a consistent performance lineup recognized by the audience.
  • AI-Fueled Operations: Employees would learn to use AI to solve problems without relying on manual labor.
  • Data-Driven Decision-Making: To make informed decisions based on real-time data.

The club's losses have been reduced from 120,000 yuan to 20,000 yuan, and it is expected to break even in July. Employees are also getting used to using AI tools. Hulan's experiment has been a success; even if he leaves in September, the club will be capable of operating on its own.

In conclusion, Hulan demonstrated that AI isn't just for large companies but can be crucial for small businesses as well. The key is to use it effectively—don't expect AI to solve all problems instantly; focus on getting the basics right first.