Three Trillion-Dollar Markets: Founders on Fast and Slow | Source Code Capital 2018 Ma Hui

源码资本·April 26, 2018·0·5

Decoding the Rhythm of Entrepreneurship

On April 20, Source Code Capital's 2018 Ma Hui annual meeting, themed "Open Source Iteration · Decoding the Future," was held in Beijing. Yungang Huang, partner at Source Code Capital, Wang Chaocheng, founder of Yijiupi, Zheng Wencheng, founder of Hai Feng Education, and Wang Weihua, founder of Baoxianshi, shared their insights and reflections on the "fast and slow" of entrepreneurship in their respective tracks.

Yungang Huang said: "You can see that the three founders all represent trillion-dollar markets. Yijiupi, which empowers new retail; Hai Feng Education, which empowers teachers; and Baoxianshi, which empowers insurance agents. How to go faster? Good market, good model, good cognition, good strategy, good execution — that leads to good development."

Founder Quotes / CEO QUOTES:

When your business and cash flow are stable, you must run for your life.

You need to know exactly how much money is in the bank. Your growth speed needs to match that. Investors like to talk about 18 months of cash flow — we have more than enough.

100% growth at large scale is fundamentally different from 100% growth from a base of one.

Rest assured, an industry like insurance doesn't need to be fast. Sooner or later, it will become enormous.

50% of my time is spent on organizational upgrading.

Every single one must be personally interviewed by me; every offer is carefully reviewed and confirmed by me.

On talent management, we spend 50-60% of our time; recruiting takes 20%.

Advocating internal gene growth and transformation — you need to spend more than 40% of your time on transformation.

By the end of this year, barring surprises, we'll be a company of ten thousand people. That makes me feel like I'm walking on thin ice.

The biggest failure lessons were bought with a lot of money.

Set goals rather than KPIs in employee project management.

The essence of platform model is making information connection more efficient, but it doesn't truly reduce costs. You must make the physical pipeline of goods flow change — become shorter, thicker, and more efficient.

A CEO's cognition determines the company's boundaries, but it's very hard to grasp. It's easy to become arrogant and stop hearing voices from outside.

From left: Yungang Huang, Wang Chaocheng, Zheng Wencheng, Wang Weihua

Full transcript of the panel discussion below

Yungang Huang, Source Code Capital: Hello everyone, I'm Yungang Huang, moderator of this roundtable. Today's theme is "speed conquers all." Our three guests' companies — Yijiupi, Hai Feng Education, and Baoxianshi — are all top-ranked in their industries and maintaining rapid growth. I'd like to hear from all of you about where exactly your companies are "fast."

Wang Chaocheng, Yijiupi: Yijiupi started in the alcohol business and now covers the full range of consumer goods as an industrial internet company. In FMCG B2B, we're temporarily in a leading position, with annual transaction volume of over 7 billion RMB, including 3 billion in beverages and alcohol, with food just starting to gain momentum and currently in second place. Alibaba does a bit over 3 billion, JD.com is smaller. In terms of merchants, we cover 350,000 small B customers, with monthly active users staying above 100,000.

Zheng Wencheng, Hai Feng Education: Hai Feng Education provides live one-on-one online tutoring for K-12 students. We've been continuously thinking about the problem of personalized teaching in education. For two thousand years, personalized teaching has remained just a slogan. The reason is that teaching demand is huge, but supply is limited — especially quality supply — and matching efficiency has always been constrained by geography and time. Thus, the vast majority of Chinese families cannot access quality teaching resources. Through live one-on-one online teaching and our online teaching platform, we can:

First, solve the time and space problem. Teachers and students break through geographic restrictions, eliminating commute and offline venue costs, with higher matching efficiency;

Second, solve classroom transparency. From the perspective of parents and institutions, every tutoring session is a black box — unable to monitor teaching quality or control standards. Online teaching enables transparency and openness, with full process visibility;

Third, use technology to solve scalable supply. Using AI, big data, and other technologies, enabling more ordinary teachers to quickly become standardized good teachers, achieving teaching empowerment;

Currently, Hai Feng is in a state of having completed steps one and two, while iteratively exploring step three. In terms of the definition of "fast," Hai Feng's business speed is relatively fast, especially in the vast second- and third-tier city markets. From 2015 to now, we've maintained 5-6x growth annually, and will continue to do so for the next 1-2 years.

Wang Weihua, Baoxianshi: Hello everyone, I'm Wang Weihua from Baoxianshi. Before introducing Baoxianshi, let me briefly talk about the insurance industry. Before 2016, insurance was seen as a scam industry in ordinary people's minds. Since late 2017, perceptions have shifted — the wealthier people are, the more they realize the importance of insurance. The insurance industry is a ten-trillion-dollar market, and annual growth is above 10% every year, with original premium growth above 20% annually — all trillion-level figures.

I think among all industries, such growth rates are relatively rare. And in the process of several trillion RMB in original premium growth each year, Baoxianshi is a platform providing services and help, serving those at the very bottom who create this multi-trillion market — the insurance agents who were once considered "scammers."

Yungang Huang, Partner, Source Code Capital

1 How to Balance Speed and Excellence?

Yungang Huang, Source Code Capital: Every company pursues faster development; every investor wants to back faster-growing companies. How to achieve speed? Sometimes you need to polish your product and perfect your service before aggressively developing the market; sometimes you need to seize market share first, get capital support, and only then have the chance to polish your service. I'd like to ask the three founders: how do you balance and trade off between speed and excellence?

Wang Chaocheng, Yijiupi: Fast and slow differ in every industry. It depends on the balance of three factors: competitive landscape, cash flow, and business quality.

First, competitive landscape. Some industries have very short windows of opportunity. Take bike-sharing — competitors expanded rapidly, and if you couldn't keep up, the war was over. In B2B, however, because offline small shops' traffic is location-based rather than online, giants can't monopolize. If you have exclusive advantages in certain categories among 2,000 SKUs, giants can't kill you. Of course speed is still needed, just not as fast as bike-sharing.

Second, business quality. Category selection matters a lot. B2B companies starting from beverages generally struggle. We started with alcohol, where per-unit margins are high, so our business quality was good from the beginning. With eighty to ninety distribution points nationwide, selling only alcohol was basically break-even. Expanding into non-alcohol categories from this base was relatively relaxed.

Third, cash flow. You need to know exactly how much money is in the bank. Growth speed needs to match that. I'm a relatively conservative person. Investors like to talk about 18 months of cash flow — we have more than enough.

Overall, "fast" differs in every industry. It's related to competition, and also requires balancing business quality with cash flow. If the company goes bankrupt, speed means nothing. When your business and cash flow are stable, you must run for your life. Actually, there's no such thing as regional internet or category internet — internet is platform-based, national.

Zheng Wencheng, Hai Feng Education: Our industries differ greatly, so I'll try to address commonalities. The balance of speed and excellence may differ across the 0-1, 1-10, and 10-100 stages. The 0-1 stage requires more pursuit of excellence, ensuring healthier subsequent development from 1-10 to 10-10,000. Take Hai Feng as an example: can you find good acquisition channels, so that decline during future expansion is more controllable? Can you find better ways to improve human efficiency, so that as scale grows, future organizational expansion can be relatively slower?

At the 1-10 and 10-100 stages, I think the emphasis on scale grows larger. As the company's maturity increases, marginal improvements from excellence become increasingly faint. The company's focus gradually tilts toward speed and scale, increasingly toward market share. 100% growth at large scale is fundamentally different from 100% growth from a base of one.

For Hai Feng specifically, the early 0-1 stage may have been slower, spending a lot of time polishing teaching research, teaching processes and models, testing front-end and back-end, finding a relatively reliable path, and then scaling — this way risks are more controllable.

Wang Weihua, Baoxianshi: Investors prefer entrepreneurs to be "fast." Many investors looked at Baoxianshi before, and ultimately didn't invest. Because although we had some scale, we weren't fast enough. But I've always remembered a senior's words: "Rest assured, an industry like insurance doesn't need to be fast. Sooner or later, it will become enormous." I also want to thank Source Code Capital for believing in and investing in us.

I think fast corresponds to slow — fast is slow, slow is fast. Before you make your strategy, the path must be slow, and you must think it through. As a founder, you shouldn't be quickly swayed by external interference. Once you make a decision, you must be firm and swift in action.

Wang Chaocheng, Founder, Yijiupi

2 Organizational Upgrading and Management

Yungang Huang, Source Code Capital: I'd like to ask the three founders: how do you prepare personnel and organization before business takes off? In building good business teams, especially in assembling core executives and core teams, how do you spend your time?

Wang Weihua, Baoxianshi: My partner Chen Xiao and I spend 50% of our time on organizational upgrading. If the organization doesn't upgrade and iterate, problems with employee cultivation will affect the company's subsequent development.

Zheng Wencheng, Hai Feng Education: Currently at Hai Feng, I interview people two reporting levels down. Every single one must be personally interviewed by me; every offer is carefully reviewed and confirmed by me, and if there are issues I'll raise them. Regarding executive recruitment, Yungang often recommends good talent to me. Chatting with candidates until midnight — there's no problem spending more time on such good talent. This is the most worthwhile investment during a CEO's rapid growth phase.

Moving on to organizational capability — this is a common topic today. Our management's annual meeting set the tone that 2018 is our most important year. Because by the end of this year, barring surprises, we'll be a company of ten thousand people. That makes me feel like I'm walking on thin ice. Among these ten thousand, a large number are full-time teachers, and managing teachers online is completely different from traditional offline training institutions. We have no teaching locations; all teachers are distributed nationwide, teaching entirely from home. There are also part-time teachers — how to manage them well? This is an enormous challenge, one that cannot be solved with traditional offline approaches. We need to learn from management approaches of strongly operation-driven internet companies, and figure out new methods suitable for our industry. How to solve this problem is the hardest thing I think about every day.

Wang Chaocheng, Yijiupi: On talent management, we spend 50-60% of our time; recruiting takes 20%. I believe team upgrading should be organic. I don't really believe that a company doing very well relies on hiring different people every day, constantly bringing in new executives. My view on people is that internal genes can support it. Advocating internal gene growth and transformation — you need to spend more than 40% of your time on transformation. Of course some genes can't be cultivated internally — for example, finance. Food differs greatly from our existing alcohol and beverages, so we tend to look externally for that. But frontline employees, like ground sales staff, are highly standardized and mainly managed through systems and data.

Companies must clearly understand the relationship between their business strategy and people. Understand whether the business strategy's requirements for people's genes can be generated internally or must come externally. A company that spends massive time looking externally every day, without cultivating internally, is taking a huge risk.

Zheng Wencheng, Founder, Hai Feng Education

3 Good Cognition, Good Strategy, Good Execution

Yungang Huang, Source Code Capital: We just talked about rapid execution after decision-making. Indeed, from the results perspective, people look at effects, results, numbers, headcount growth — this is the execution process. How to make decisions faster, how to improve cognition, how to formulate strategy?

Wang Weihua, Baoxianshi: We have some of our own lessons. The biggest failure lessons were bought with a lot of money. Previously we made an important decision, quickly decided it was doable through a series of meetings. In the end, we found it was fundamentally unhelpful to the company, yet the costs were already spent. Now when exploring decisions, we proceed in steps:

First, establish a project team. We generally hold a co-creation session, inviting external institutions, including ourselves and core personnel, for two to three days of closed-door sessions to analyze whether this project is doable or not, and how. Think it through. After careful consideration, it definitely shouldn't be the boss slapping the table saying this must be done, with partners brainstorming in isolation.

Second, execution. After thinking it through, no startup should have problems at the execution level. The main issues are detail control and progress control. Set goals rather than KPIs in employee project management. Set a series of measures for the goals, with the core being organizational allocation. In the process, each person sets their own KPIs, doing things step by step. If no major problems arise, it will succeed.

Zheng Wencheng, Hai Feng Education: Returning to the excellence we discussed earlier — excellence isn't just about doing things excellently, but also thinking things through excellently. You need to clearly know that for the education and training industry, brand matters most. First, make people who know you and have experienced you approve of you; second, make more people aware of you. The first step is very clear — it's essential, it's the prerequisite. In the 0-1 process, why do teaching research first? Why develop the technology platform first? Why endure not scaling until these are done? It immediately becomes very clear.

Why must we use full-time teachers? Even for part-timers, why examine them with the same standards as full-time teachers? Because we know that full-time retention rates and stability are critical to ensuring business quality. The number of students a full-time teacher can serve is vastly different from a part-timer. Full-time teachers allow us to more calmly carry our set annual performance targets. Many companies taking the part-time path become extremely uncomfortable once they reach a certain scale, and must go back to redo the full-time path.

Also, many people ask why we built our call centers in Beijing and Shanghai. We clearly know that the work ethic of sales talent in first-tier cities — second- and third-tier cities can hardly imagine them willing to grind until midnight, grab a sleeping bag in the office, and get up the next day to keep calling. Once you think this through, subsequent execution must be fast and decisive — you must walk this path firmly and quickly.

Wang Chaocheng, Yijiupi: I think a CEO's performance for the enterprise is the result of cashing in their essential cognition of the industry, of their track. A CEO's cognition determines the company's boundaries, but it's very hard to grasp. It's easy to become arrogant and stop hearing voices from outside. My experience has three approaches:

First, prototype thinking. What is the essence of the company? Before the internet emerged, what was its business prototype? What changed after the internet emerged? The prototype business of B2C is Walmart and Carrefour. Before the internet, Walmart and Carrefour had no category restrictions, so you find that today's B2C — Alibaba, JD.com in China, Amazon in the US — are all platform models. But before the internet, B2B prototype agents — you rarely saw beverage agents selling steel, or steel agents selling cars. The world's largest channel companies don't cross categories very much. This inspired us that one-stop B2B companies might be quite dangerous, so from the start we only sold alcohol, then adjacent beverages, now adjacent food. We only try one category at each stage, and it must be adjacent — all must be necessities for tobacco and liquor shops, mom-and-pop stores, and restaurants. B2B and B2C have fundamentally different industry essences. Even with the internet accelerating industrialization speed, ancient industry models aren't so easily changed. Prototype thinking makes your cognition less likely to be distorted.

Second, competitive thinking. You need to be different from others, otherwise you'll be very passive. Take Alibaba — compared to us, they moved slightly later, but spent so much money yet have been consistently losing money. What's important is that our approach is different. We're very heavy, extremely heavy. We have our own warehouses and also borrow others' front warehouses; we do delivery ourselves; we also have our own ground sales staff, our ground sales are our own employees. We use our own managed drivers for delivery.

My personal view: in B2B, especially FMCG B2B, platform models don't exist. B2C platforms work because selling directly to consumers makes gross margins higher than B2B selling to terminal stores, and because user base instantly goes national, the scale increases, so gross margin boundaries open up — thus they can pay platform fees like Tmall's, and have more money to pay SF Express, ZTO, YTO, Yunda. But in B2B, even as a platform with dealers入驻, you're still selling to small B — the环节 hasn't changed, gross margins haven't fundamentally improved, total user count hasn't changed, so B2B can't change gross margin boundaries. So the essence of B2B is reducing costs and improving efficiency through pipelines. The essence of platform model is making information connection more efficient, but it doesn't truly reduce costs. You must make the physical pipeline of goods flow change — become shorter, thicker, and more efficient. Our goal is to become China's largest FMCG circulation pipeline — my warehouses are the most efficient, my delivery density is higher, my delivery speed is fastest, my manufacturer-to-merchant环节 is shortest. So you must be driven by internal efficiency to reduce costs, in order to meet terminal demand for cheaper prices and better service.

Third, customer value thinking. Customer value thinking always ranks first, but the two key thinking modes for achieving customer value are prototype thinking and competitive thinking. As for other decisions, daily operational decisions require extreme speed. All management decisions must not exceed 12 hours.

Wang Weihua, Founder, Baoxianshi

4 Technology Content Improving Human Efficiency

Yungang Huang, Source Code Capital: In good industries, there are many ways to be fast — model innovation and technological innovation. Why is online education developing so fast now? What changes will technology bring to this industry?

Zheng Wencheng, Hai Feng Education: We have our own definition of human efficiency concept. One is improvement driven by management content, the other is improvement driven by technology content. Management content is your organizational capability, management efficiency. But after reaching a certain ceiling, no matter how good management is, if technology content improvement doesn't go up, there will be huge impact.

Traditional phone call centers — you learn many common套路 to a certain point and hit a ceiling, and must rely on technology to improve. Your business complexity determines that you must build in-house, must deeply understand sales personnel's needs, liberate their common human efficiency, and use big data, artificial intelligence for higher-efficiency matching. For example, order allocation — what kind of people signing what kind of orders is more efficient? For example, quality inspection — how to allocate more work to machines to complete, making efficiency higher and management penetration deeper.

Traditional teacher population management is distributed management, very dispersed. Objectively speaking, one-on-one teachers at scale are largely mid-level teachers in the industry, as determined by cost and income. How to empower them? You need to固化 many empowerment环节 in the technology process, teaching them how to select topics, prepare lessons, and through log data grasp how good teachers perform. First find common rules, find local optimal solutions; after accumulating certain data, find global optimal solutions. This is our approach.

Yungang Huang, Source Code Capital: You can see that the three founders all represent trillion-dollar markets — Yijiupi empowering new retail, Hai Feng Education empowering teachers, Baoxianshi empowering insurance agents. How to go faster? Good market, good model, good cognition, good strategy, good execution — that leads to good development.

Past Ma Hui Articles

3rd Edition | Yiming Zhang: A CEO Must Avoid Rational Arrogance

3rd Edition | Lianjia's Zuohui Zuo: My Understanding of the Real Estate Industry

3rd Edition | Meituan's Huiwen Wang: Persist in Doing the Right Thing, the Payoff Cycle Is Shorter Than Imagined

3rd Edition | Jingbo Wang: Deep Operations in Finance — Licenses Only Work When You Understand Operations

3rd Edition | Meili United Group CEO Qi Chen: The Difference Between Mules and Unicorns Lies in Focus

3rd Edition | Zhi Qin: The Things Only Entrepreneurs Can Teach Themselves

3rd Edition | Source Code Capital Investment Partner Dr. Hongjiang Zhang: The Essence and Opportunities of AI

3rd Edition | Microsoft Research Asia's Yu Zheng: AI Applications in Urban Management and Business

3rd Edition | What Pits We Stepped on in Three Years of Entrepreneurship

2nd Edition | Xiang Li's First Talk on CHJ Automotive: Transforming Urban Mobility

2nd Edition | Yiming Zhang: Why I Don't Agree with Controlling Labor Costs