This is the first time Xiao Hei Yang has reposted someone else’s article. Having followed business WiFi for a long time, this article resonates deeply with me. I originally wanted to write a critique on business WiFi, but now I feel that this article is enough as a foundation.

Author: He Mingke Link: View Original Source: Zhihu

Although the free Wi-Fi model or business WiFi has passed its peak, and most of the answers to this question occurred in 2014-2015, I still want to angrily respond, dedicated to my youth that will eventually fade away. You might wonder, how does a data analyst programmer get involved in Wi-Fi? However, this story full of bitter tears actually started from big data. Part One: From Big Data to Business WiFi Our team responded to the national call for “mass innovation and entrepreneurship,” hastily starting a business in 2014 with a group of graduates from prestigious universities: Tsinghua, Peking University, Stanford, Oxford, Hong Kong University of Science and Technology, etc. With a solid foundation in big data and programming, we ambitiously began with big data. As we delved deeper into big data, we realized we had no data to analyze, which is like a skilled cook having no rice to cook. The best way to obtain data is to develop an app, but looking around, we found that developing apps was quite challenging. Firstly, tool-type apps have mostly filled the market, and even if there are some, they are resource-intensive and not suitable for small teams. Secondly, creative apps (like social apps) are not suitable for us older and introverted individuals. In our despair, a teammate proposed a brilliant idea: users’ pain point is WiFi in restaurants and shopping malls; we should build WiFi for these commercial venues; while satisfying user needs, we can also collect data and occupy the users’ internet access point. It was only later that we learned this model is called “business WiFi.” What a genius idea: collect vast and complete internet behavior data; it also comes with a business model. For us, software and hardware development and system procurement were not problems, OpenWRT/web/client were not difficult points. We quickly developed a mature business WiFi system. Our model was: commercial venues pay for hardware and bandwidth, we provide installation and operation and maintenance services for free, and we collect data and internet access points. When the system was ready, we started looking for commercial venues, and right from the start, we received orders, indicating the strong demand and the attractiveness of our proposal.

Our first client was a high-end coffee chain.

Construction site

The first few sites were installed by ourselves; the advantage of engineering students is that they can endure hardship.

Part Two: From Business WiFi to Factory Dormitory WiFiAfter securing the coffee shop as our first client, barber shops and 4S stores followed one after another, and user feedback was also fantastic.

WiFi login interface and user praise

We had already begun to fantasize about laying WiFi in one place after another and ultimately succeeding in our entrepreneurial dream. Meanwhile, the data services we provided for the fund had also begun, and we increasingly felt that our model was brilliant: data could accumulate and be immediately monetized while controlling the user’s internet access point. However, good times did not last long; the business WiFi sector quickly became a heavy disaster zone for capital investment:

• In August 2014, Tree Bear received 100 million yuan investment from Alibaba

• In December 2014, Maiwaidi obtained 300 million yuan investment from Tencent

• In early 2015, Ju Wan Network received 120 million yuan financing, led by 360 company, with Haitong Securities and Zhongguancun Xingye Capital participating.

These big players, buoyed by capital, immediately adopted a free model, providing merchants with devices and bandwidth, and even subsidies. Suddenly, our fee-based model was obliterated. Our shortcomings in leveraging capital and storytelling were laid bare, making it impossible to keep pace with competitors. Most critically, we began to realize that 4G was rapidly replacing WiFi. Shenzhen, as a pilot city for China Mobile’s 4G, offered an excellent 4G experience and cheap packages, leading to a significant replacement of WiFi in urban areas.

The double whammy forced us to find new data sources. Shenzhen’s strong global manufacturing base gene allowed us to quickly find the next entry point—factory dormitories. In the Pearl River Delta region, the most abundant resource is the large number of workers, producing over 90% of the world’s electronic products. Factories generally provide dormitories for workers, but to save costs, they do not provide internet. However, today’s workers are mostly born in the 90s, who generally own only smartphones and no PCs, making WiFi access an absolute necessity.

Factory and dormitory conditions around Shenzhen

Thus we entered the factory dormitory to start a new model: find a factory to negotiate, requiring no investment from the factory, just permission for us to construct; we would be responsible for all hardware investment and broadband costs. The following images show our installation of routers and APs in the factory dormitory. The environment was much worse than the previous coffee shop; the ladder in the left image was about to fall apart and was shaky, right next to the balcony, so someone had to hold it while operating. We completed the transformation from urban to rural, with everyone becoming the king of the suburbs; the young man in white in the right image is a technology expert who graduated from Tsinghua.

Construction site

Our initial profit model was to distribute apps and games, but we found that workers were too tired during implementation. The manufacturing workers generally work in two shifts: day shift from 7 am to 10 pm with lunch and dinner breaks; night shift from 7 pm to 7 am the next day with a midnight snack. Moreover, the crowd is too fixed, which is not conducive to constantly installing new apps and games. Therefore, the distribution effect and conversion rate were particularly poor, and we couldn’t recover the previous investment costs. So we considered transforming from the free distribution model to charging workers a monthly internet fee, ranging from 10 to 30 yuan per person. Our strong software and hardware development capabilities allowed us to quickly complete the transformation, finishing development in a month. The following images show our charging interface and notification interface, which also provide a great experience; the payment process is smooth, and the payment rate is high. Considering that workers’ usage of Alipay and online banking is low, we also developed a human proxy charging model: increasing the payment rate without increasing labor. We hired a worker in each dormitory building who could use online banking or Alipay as a proxy to collect cash and help people recharge online.

Charging interface and notification interface

However, machines are not completely reliable; as long as there are machines, they will malfunction. However, 90% of the solutions rely on restarting, requiring someone to unplug and restart. To save labor, we developed a remote control system for routers based on Arduino, which can restart routers located far away in factory dormitories by sending SMS to control the hardware switch, ensuring internet experience; moreover, the Arduino module can also send SMS to report the router’s status.

Machine room environment

Part Three: From Factory Dormitory WiFi to Data Analysis Subsequently, it was proven again that idea is cheap; our “brilliant ideas” cannot maintain a long-lasting lead. As the commercial WiFi in the city became chaotic, more and more business WiFi companies also noticed the factory dormitory market and entered the competition. However, relying solely on better technology, we outperformed competitors in investment costs, service experience, payment rates, and other aspects. With technology, we can recover costs within 6-10 months after investment, making us the most profitable project in this niche industry. However, in the face of this market, even though every factory dormitory is profitable, we still reflect: what is our original intention? Are we going to fight to the death in a shrinking market with a group of reckless and moneyless individuals?

• We have reduced the management costs of factory dormitory WiFi to a minimum through technology, but the initial development and installation costs can no longer be reduced. These matters are not our strengths.

• Our original intention was data; we did WiFi to solve data source issues, not to do WiFi for the sake of doing WiFi and using data analysis to solve monetization problems.

• We must go with the flow and avoid taking unnecessary detours, especially not going in circles. The popular saying at the time was: I have occupied the blue-collar dormitories, thus mastering blue-collar workers, and can do blue-collar recruitment or blue-collar social networking. These are vast markets without strong competitors.

In the face of such circumstances, we decisively returned to our original intention—data accumulation, analysis, and visualization, which led to this story (What interesting data mining/analysis projects have you done with Python? – He Mingke’s answer). Thus, we transformed into our current model (What is the current situation of the data analysis and data mining market in China? Are there many opportunities? What are the prospects? – He Mingke’s answer). Nevertheless, our factory dormitory WiFi is still running well.

• Every dormitory we set up provides a steady cash flow every month, and we have long since recovered our investment, enjoying a comfortable life of earning small profits while lying down.

• Our entire WiFi system (software and hardware) has gained a good reputation in performance, and more and more people are coming to purchase the complete source code and solution, many of which are large companies that have received VC investment. (Here’s a little advertisement, our system and source code are definitely for sale, we absolutely want to sell them.)

Finally, a summary of the business WiFi modelAs someone who has earned a little money in this industry, here are a few points to summarize:

• 4G has a significant substitutive effect on WiFi, especially after the coverage of 4G has greatly increased and prices have dropped significantly, the survival space for business WiFi is getting smaller. User habits are increasingly leaning towards 4G.

• Business models should not take detours; they should be simple and direct. Most of the business models that can be thought of for business WiFi (big data, precise advertising, coupon pushing, etc.) are just beating around the bush, and currently, there is no case that can prove it is profitable or can recover costs.

• The only profitable model is the internet fee model that we developed from factory dormitories; however, limited by the shrinking market and high execution costs, it can earn small profits but is difficult to scale. If you want to discuss blue-collar recruitment and blue-collar social networking, please refer to the second summary.

• I once saw a saying: In this world, many people have never experienced making something and then selling it for profit. However, our small team has achieved this. Currently, investment is too hot, and many models are just taking VC money to appear prosperous, but in essence, they are completely untenable.

There are many enthusiastic netizens who have early responses to this question; everyone can click on the original link to view.