This article was originally published in Zhihu’s column “Mai Suji” with the title “How to Grow Users – My Practical experience in Uber and Mobike”

The author column: https://zhuanlan.zhihu.com/maisuiji

Over the past year I’ve had a lot of friends share their startup ideas and ask me what else I can do in the post-Internet startup era. I’ve always said that 2017 is the year when online traffic prices hit a tipping point, and startups now have to fit into one of these two categories:

A. Either extract High value from each user (High LTV, such as cash loans, big games)

B. Either there is a channel that is so cheap that it costs almost nothing (Low CAC)

If you can’t find an economic and reasonable way to realize the unit, or there is no unique channel to obtain customers, do not easily resign, returnees, the future of the Internet industry belongs to giants (TATBMD ranking points in order), not entrepreneurs.

For example, a friend took Xu Xiaoping’s angel investment to do skills payment platform (Thumbtack), which is very delicate and runs very well. However, it took nearly 100 RMB to acquire an effective user, which was not a natural amount, so we closed our business last year. Similarly, big strategy games now make headlines for a download activation (CPA) of $120, which is incredibly expensive.

That’s why the headlines are filled with ads for kryptonian mobile games (Blue Moon, Chu Liu Xiang), English training and used cars. Without exception, they all meet the condition I said a, a user can make a lot of money, so they can afford to buy ads directly at a high price.

The success of Mobike and Ofo in 2017 mainly meets condition B. In fact, the two companies hardly advertise information flow and app store effect, and most of the effective customers come from the TWO-DIMENSIONAL code of the car body at subway entrances, bus stations and roadside. As a result, both companies are advertising heavily in subway stations and bus stops to enhance the main acquisition channel and get users to choose themselves rather than their friends when exiting the subway.

In the post-mobile Internet era, the most important means of production are traffic and capital, and it is difficult for technologies and products themselves to form barriers (such as AI, VR/AR and blockchain). Referring to Douyin and Muse, the two products look exactly the same. The unit data (unit price and retention time) are several times different, and after a period of time, the number of users is thousands of times different. Tiktok is a copy of Muse (http://Musical.ly), but the apprentice won. Sure enough, in November Toutiao (tiktok’s parent company) bought Musical.ly, which failed to beat musical. ly.

Readers may ask, how can such a pixel copy of the product have such a wide range of data? This is the difference between a grounded team and a grounded team. Having a sound growth team and methodology is critical in this era. In the following article, I will roughly summarize the insights I learned from Ed Baker, Andrew Chen and other experts in the growth team of Uber in the past, as well as the business experience in Mobike, and talk about the ideas in the current industry.

What is “growth”?

Growth refers to Marketing mass Internet products with a programmer’s mindset. The first wave of Internet companies hired traditional AD agencies and marketing directors, and more than half of the super Bowl ads in 2000 came from Silicon Valley, before all but collapsing during the bubble. When Mr. Ma launched Taobao in 2003, he started with television ads, print ads and outdoor ads.

There’s a famous saying in traditional advertising: “I know 50 percent of this AD is wasted, but I’ll never know which 50 percent.” Gradually, we found that Internet products can be done in a new way, let engineers take the lead, product iteration, AB testing such thinking to reduce waste, improve conversion rate, eliminate “self-promotion advertising”. Domestic call this flow dozen method. Gradually, people woke up and fired their marketing directors, replacing them with product managers, product operations and programmers.

What do companies do during outbreaks?

I went from “3 million orders per day” to “25 million orders per day” at Mobike. Before the outbreak, the founding team had found a smooth channel to acquire customers: the qr code registration of the bicycle body scanned at the subway entrance. Until I left Mobike, this channel was still the largest and most efficient source of new users. With a deposit of 299 yuan and a car full of multi-bet, Mobike achieved a magical positive cash flow pattern:

Cost per vehicle < deposit * Number of deposits per vehicle

It also generates positive cash flow, which should make every Internet dreamer laugh. Also therefore, sequoia, Kaoling, Tencent and other investment admission.

Now that unit economy is proven, the second step is to increase efficiency while expanding operations. So we continued to increase the number of vehicles we put in, and we also started to improve the activation conversion rate.

In the new user registration process, we use Growing. IO to look at the churn funnel for each page. After understanding this, I decided to cut down the guide page and replace the previous four steps of mobile phone verification, deposit, ID card and invitation with three. The deposit before the id card verification is a clever design: for users, after hundreds of deposits, what is privacy? So this is a very high conversion rate.

All of these are cases where the App is already installed. In a real usage scenario, the step that has a bigger impact on conversion is download and install. They’re all outdoors, they don’t have a lot of traffic, and when they see the jump to the App Store, a lot of people close the page. Therefore, Mobike made wechat mini program, and copied the revised new user process in the mini program, and the conversion rate further soared. How exaggerated is this channel? For nearly half a year, Tencent told us that more than 50% of all traffic on miniprogram came from Mobike.

Optimized acquisition and activation, time to improve retention and activity. There are many strong competitors and our primary goal is market share. Wang Xiaofeng, a former Uber GM in Shanghai, is playing against Zhang Yanqi, former Uber RGM in the North and West Region. Yang Yujie, a returnee from Uber headquarters, is playing against Chen Wei, a returnee from Uber headquarters. Both Mobike and Ofo are not short of experienced veterans. In 2017, we’ve all seen how veterans fight subsidy wars: they’re well-trained, they recruit, they burn cash. During this outbreak, we summed up the following things:

Flush back — Free day — Red envelope car — Fission free Day — Easter Egg car — Treasure box car — Red envelope car V2 — Member — Monthly card — Monthly card V2 — Red envelope car V3 — monthly card V3

Ofo is basically similar, with more GXGY attempts than we did, but the effect is not so good. Hit up to now, both sides who also did not get the other side, but killed a small player. Bluegogo’s bikes are easy to ride, but its products cannot grow. In this era, the aroma of wine is also afraid of deep alleys.

To sum up, what Mobike has done is basically in line with the AARRR model, finding a good Acquisition model, using various methods to improve Activation, The product matrix is then used for Retention, Engagement and friend Referral. As the technical threshold is too low and the competition is too fierce, I think it is very reasonable to directly skip the turnover (Revenue) in AARRR. In this kind of tuyere, only step on the accelerator to grab market share to further take financing, alongside the giant, the battle did not finish absolutely can’t rush to do revenue.

As an interlude, speed, execution, is very important in this battle. Both ended 2016 with poor code architectures and ended 2017 with bayonets of technical debt. The two companies maintained a weekly frequency for a while, but the outage was so severe that it was indecent for employees to work overtime. When I talked to the CTO of Ofo over morning tea, I felt miserable.

Later, they could not resist any longer, so both of them agreed to change to two weekdays, and were free to pay off the technical debt. Mobike also had the opportunity to do a reconstruction (Kotlin rewrote the Android terminal), and launched the Manhattan project (a major overhaul of core UX/UI) that Liu Yao and I had been dreaming of. Mobike’s client also won wandoujia’s 2017 Design Award.

What is the system of a mature company?

I wrote code in the growth department at Uber headquarters, starting with Engagement and then leading Loyalty programs. Andrew Chen, the director of our department, is an Internet celebrity who has been writing about user growth for more than a decade. Eric Ries, author of The Lean Startup, has commented that “Andrew Chen’s blog is one of The best Startup content blogs ever written.”

When Kalanick hired Uber to poach him. He bought his failing company outright. He met Marc Andreessen in his early years as a blogger. When Kalanick was out of the running in 2017, everyone was upset. Last month, Andrew left Uber to join A16Z, one of silicon Valley’s upstart venture capital firms, as an investment partner. He had a particularly good point:

All growth tactics become law ofshitty clickthrough rates.

So what works really well, everyone keeps secrets, and whatever you blog about, it’s outdated and it doesn’t work very well. Overall, then, growth departments cannot rely on one or two quirky techniques, but must have a methodological approach to discovering, testing, and promoting new techniques.

When I joined Uber, I just started the re-org process, which was to break up the enterprise structure that had been divided by job types into separate business units with a whole set of products, front end, back end, operation and data. I was going to follow my LinkedIn boss into the mobile group, but instead I went straight into the growth group, working for Ji Yang (who went on to become senior director and head of growth at Guazi Used Cars). It was led by Ed Baker, who came from Facebook. There have been several reorganizations, but in general, our growth department is divided into four groups: recruit, activate, activate, and retain.

When Andrew came, he saw that he did not invite friends, so he added a group, so it became:

  • Acquisition

  • Activation

  • Engagement

  • Retention

  • Referral

We will talk about the KPI of each department in detail, what we have tried and what skills we have.

A customer Acquisition:

Retained the Retention:

Active Engagement:

What if I go from zero to one?

The biggest difference between mature companies and startups is that startups don’t know if they’ve found a Product/Market Fit (PMF) and don’t have a clear pipeline of customers. The key is to build a “hot wheel” of growth.

Growth flying wheel

As anyone who has read The Lean Startup knows, there is one and only way for a Startup to die: it runs out of money. At this stage, if found PMF, growth momentum to do, hot money must be crazy to find you. So happy companies have all kinds of happiness, and unhappy companies are all the same — money running out, no real need to find and endless growth.

No growth is the curse of start-up companies. As long as there is no growth, the user base will shrink and slowly become more and more bleak, which can be referred to renren and Jumei Youpin. Tencent and Facebook, on the other hand, have reached 1 billion days and are still rising, as if there is no end to it.

The direction of finding real needs is so grand that it deserves another article. The cliche 80-20 rule and Andrew Chen’s “use it worse” rule come to mind.

The 80-20 rule says that at any given time, 80 percent of customers for each product come from the same channel. The law of “use it worse” is that if you find a very efficient channel, do not rest easy, good life can only last 3 months to 3 years, and soon lose.

These two principles, must cooperate with the capital chain of start-up companies to practice. Startups are so strapped for cash that they have to split every penny in half. In three steps:

  • Cast a wide net and try all possible ways to get customers at the lowest cost (no more than tens of thousands of dollars) to get the “golden water pipe” that is reliable at this stage;

  • Turn on the “gold pipe” to the maximum, and turn on the surrounding “silver pipe”;

  • Repeat the first two steps before the golden hose runs out.

Creativity is important, and Synergy is also important. Previously, the most effective way was fission Hongbao (PDD, Didi, Mobike) and invitation coupon GXGY (Didi). New ideas come up all the time, but remember, the best ideas are the ones that are used the least.

In addition, Cheng Hao, founder of Xunlei, previously said that all successful startups enjoy a wave of dividends, so the trilogy must be done in accordance with The Times. Zynga could have started off with Facebook user dividends, Uber could have started off with mobile Internet dividends in 2010, and Pinduoduo could have started off with wechat dividends last year, but the dividends of mature platforms have been wiped out, and the golden pipes will surely appear on the new platform.

conclusion

This is how I summarize user growth at three levels: startup, explosion, and maturity. In general, it is the initial period as thrifty as possible to find suitable for their own products, and then gradually transition to the mature period of a set of systems.

, End,

Micro-interaction ∣ Detail design makes for great products

Long press, identify the QR code, add attention

Share the article and let more people know about microinteraction