The secret of growth: How did Didi discover the most effective channels and growth methods?

Since I joined Didi in August 2015, I have been responsible for attracting new carpooling owners, increasing shuttle bus users , cold starting the minibus business, and express bus business in Chongqing counties.

Looking back at the 650 days at Didi, I have experienced many growth businesses from 0 to 1. When facing the market and users, being able to explore the most effective channels and the most effective growth methods is what I think is the most meaningful thing.

Before joining Didi, I worked at Uxin Used Cars , where I pushed UV to over 1 million in just half a year and successfully built an Internet used car brand through online channels and media advertising.

I hope to analyze these growth projects in depth, from the essence to the surface, from logic to grasp, from questions to answers, and review and reflect deeply on the work of the past two years.

No matter what job you do, no matter what problem you solve, you will encounter the following problems:

Where do I start? Which direction? How to control the rhythm?

In the face of these problems, we need to first go back to the origin of the problem (the purpose of solving the problem) - demand . Because there must be people behind the demand. The core issue is, if you don’t know who uses your product, when they use your product, and where they use it? There is no answer to this question.

The people behind the demand must have commonalities , and this commonality requires a lot of operational means (research means) to figure out. Simply put, it is to bring out qualitative problems through communication with users, then do quantitative verification, and finally discover the essence of the problem. For example, you can ask certain users what makes them unhappy when using Didi. After getting the conclusion, you can expand the sample analysis, summarize and classify to find the root cause.

What does it mean that the Internet does not create demand ? For example, the need to eat. People must eat from birth, and the emergence of reviews has not made people eat more. Many needs are innate and are not created through the Internet. Therefore, you need to have a clear understanding of your own business and what exactly you are solving. If your product does not have the rigidity and sustainability to improve the efficiency of solving demand, then it is a false demand.

As far as dining is concerned, for example, Meituan provides more choices for dining, delivers food to users faster, and allows users to eat better meals, saving more time and money. In fact, it revolves around the four quadrants of demand: "more, faster, better, and cheaper." The same is true for Didi, with more cars, faster access to cars, better service, and saving time and money.

Looking back, you have to see if you can find the real solution to the needs. You have to consider what the original demand was like. If there was no such demand originally and you created it, it would definitely not be reliable. Too many products that were supposed to create demand have died.

Be sure to remember the above picture:

Let us first interpret this picture from a theoretical perspective. We can call this process of clarifying needs the "seed stage" . The seed stage actually means that I have thought it through clearly, I have also talked to some users, and I am confident. I have made a product model, and the simpler the experience, the better.

Three things to do during the seed stage:

1. Be sure to communicate more with users;

2. There is no difference between 100 users and 1000 users;

3. The purpose is to create a retention funnel.

A large number of businesses die in the seed stage. Although the ways of death vary, the causes are similar: they all did something they shouldn’t have done at a critical time. For example, some people are a little anxious when they have some money in their hands, and they want to spend money to develop the product first, but they don’t pay attention to retention, and don’t analyze the reasons for the low retention rate in time. The result can be imagined. Therefore, the seed stage is very important. There is no essential difference between 100 people and 1,000 people. Only verifying demand is meaningful and can lay a solid foundation for subsequent growth.

In a word: you must pay attention to the retention rate in the seed stage, this is the essence!

What are the signs that the seed stage ends and the outbreak stage begins? The key to the seed stage is retention. If the retention rate reaches a certain threshold, it means that the product has successfully passed this stage. It is said that giving birth is a life tribulation for a woman, and being able to survive the seed stage is a life tribulation for doing business. I still remember that when the ride-sharing service was first launched, we saw that the monthly retention rate of car owners was greater than 30%. When you see these exciting numbers, congratulations, just go for it, the explosive growth period is coming.

As the name suggests, the explosive growth period is about trying every possible means to attract new customers . During this stage, the focus is on these three indicators : LTV, PBP, and CAC , which are all cost-benefit indicators.

CAC is the abbreviation of Customer Acquisition Cost, which means "user acquisition cost".

LTV is the abbreviation of Life Time Value, which means "user's lifetime value".

PBP is the abbreviation of Payback Period, which means "recovery period".

During the explosive growth period, if the customer acquisition revenue is positive, the next step is to focus on user activity. Because we mentioned that we should look at retention in the seed stage, which is actually building a funnel. For example, with this retention funnel, a lot of water can be poured into it during the new customer acquisition stage. Even if a lot of water leaks out, there will always be some water left. However, in the later stage of the explosive phase, you will definitely find that user activity has declined significantly compared to the early stage. On the one hand, this is because a large number of "water users" have entered, and on the other hand, users are tired of the game and it is no longer fresh to them, so you should make some changes.

In a word: the explosive period is to find the ceiling of customer acquisition while paying attention to user activity.

When the explosive period ends, it enters the bottleneck period, during which more attention is paid to indicators of user activity . At the end of the explosive period, we found that the cost of acquiring customers was very high and hit the ceiling. Then it is necessary to increase user activity through better products and better services to increase the ceiling of customer acquisition.

For example, if the user activity is 1 hour per day on average, the cost of acquiring a new user is 50 yuan. Now the activity becomes 2 hours, the cost of acquiring a new user is not 50 yuan, because the user lifetime value has doubled, so the cost of acquiring a new user becomes 25 yuan.

Getting through the bottleneck period can help us enter the next round of outbreak . Therefore, the logic of the entire growth is very clear: demand, seed stage, explosive period, bottleneck period , and in the later stage, the "explosive period" and "bottleneck period" will alternately iterate. This is the logic of growth.

Let’s talk about the law of growth. This chart represents the pace of Didi’s development from 2012 to the present. The percentage means market share (of course it is a hypothetical number, and the valuable information is the ratio between the four links).

In the early days , we attracted active users, and only 3% of them were new users. These users are characterized by strong curiosity and willingness to provide feedback.

The next 13% stage is the battle between Didi and Kuaidi, attracting them through subsidies and red envelopes.

The remaining 33% refers to the word-of-mouth stage . At that time, the most effective word-of-mouth communication scenarios for Didi were dinner parties and company groups. Our research found that the "dinner party" scenario was the most prominent scenario for users to use Didi for the first time. For example, when two people are having a meal, they often ask each other how they got here. The other person answers that they got here by Didi, and then their companion will ask what Didi is? At this time, the other party can show off, and then when they say goodbye after dinner, people who have never used Didi will have their first order with Didi, or will be left with a deep impression of Didi.

The last step is the trend stage . Today, if you see your good friend hailing a taxi on the street in the wind and rain, you will definitely feel sympathy and sadness, because it would be difficult to get a taxi without Didi. When the general trend is right, you need to look beyond growth and focus on how to make profits and how to maintain a stable scale.

Looking back at Didi, we did ground promotion in the seed phase, fought with Kuaidi during the explosive phase (issuing coupons crazily), and started private cars and express cars after entering the bottleneck phase. Users found that these products were cheaper than taxis and provided a better experience than taxis, thus entering a new period of explosive growth. Although Didi’s core now is express rides, I believe Didi will still explore areas such as chauffeur services, finance, insurance, and the after-market. We didn't go far, but we followed logic.

This table is a summary of the best and most effective growth methods at each stage from all the growth projects I have experienced.

We look at it from three dimensions: attracting new users, activity and retention .

The most important thing in the entire seed stage is to focus on "retention" , which needs to be achieved through product optimization. The most effective way to attract new users in the seed stage is still ground promotion, and the most effective way to activate is research. Don't give coupons to users, just chat with them, because the seed stage often attracts new users, and these users can also give you the most valuable feedback. This is a very critical thing to promote activation in the seed stage.

Is the purpose of research just to find problems? Of course it is not that simple. On the one hand, you can discover problems to the greatest extent through research, and on the other hand, you will find out what the users at this stage are like. The characteristics of seed-stage users are that they are curious and willing to give advice . The best way to activate them is to chat with them, provide two-way feedback, and make users feel respected.

When conducting a survey, you can tell the user, I am the product manager of xx. In order to make your use more comfortable, I hope to do a small survey. When the user feels that you respect him, he will continue to use the product and provide feedback . When you know about the problems with the product or realize that the product has harmed the users, you offer comfort and certain material compensation. The users will also give you unexpected support. When you adopt this approach when facing a group of users, you will definitely benefit greatly.

The core of the explosive period is to attract new customers . You need to establish more channels and pay attention to the channel forwarding rate and customer acquisition revenue. The best way to promote activity during an explosive growth period is to conduct sales promotions, the best way to retain users is to conduct research, and the best way to recall users is to chat with them. If users churn during an outbreak and you continue to provide them with subsidies, they may feel that this is an insult to them - -! .

For example, when we were operating buses, we did not conduct direct research on lost users, but instead directly issued subsidies to recall them. As a result, the conversion rate was very low. What was the real reason?

I remember that time when I was communicating with a user, I found out that the user bought a ticket excitedly and was scheduled to depart at 5:50, but he ended up waiting until 6:30 and still didn’t see the bus. When he called customer service, no one helped him solve the problem. He was very sad and never used the customer service again. Two days later he had forgotten about this annoying thing. At this time we informed him that I would give him a 2-dollar coupon to continue buying tickets, and his anger was rekindled. What would you think if you were in his shoes? While we have no idea about the user’s pain, we still keep telling him about discounts and promotions... Therefore, the best retention method at this stage is to chat more with the user.

During the bottleneck period, the most important thing to focus on is activity, which should be improved through products and efficiency models. Growth during the bottleneck period does not require the same trend as during the explosive growth period, but requires more stability. Of course, we should also cultivate our internal strength, build more brands, make our voice heard in the market, improve our reputation, and tap into some new demand points. The best way to retain customers during the bottleneck period is to become a member and accumulate points.

There are thousands of ways for a project to succeed, but only a few ways for it to fail. Therefore, what I want to advise everyone is to remember the logic of growth, always understand the stage you are in, and don’t do things that are inappropriate.

Now that we have the logic of growth, what are the “factors” of growth? Logic allows us to find patterns, while factors help us grasp the key points.

I have summarized that there are five factors for growth, namely demand, scenarios, brands, channels, and user education. To achieve high growth, these five points are indispensable (growth is not easy, remember that none of them can be missing).

Among them, demand and scenarios are the first layer and are the main reasons for growth. In the short term, we should seize channels through scenarios and then build brands based on demand. There are two key points in user education. The first is to tell users who I am, why they should use me, and why they must use me. The second is to lower user expectations and give users Aha-moments in future experiences.

Below we use Didi Minibus’ cold start from 0 to 1 as an example to restore these five factors to this project.

First, we started with verifying the needs at the earliest stage. At that time, we assumed that the scenario that minibuses solved was the last mile of transportation, so we verified the needs in Huilongguan and Wangjing in Beijing, completely using manual methods. We found that the retention rate was good, with a weekly retention rate of 50%. Facts have proved that it can solve the needs and users are buying into it.

Next, we decided to cold-start in Huilongguan. We decided to choose Huilongguan for two reasons:

1. The demand is stable regardless of peak or off-peak hours, weekdays or weekends;

2. Traffic in the temple is not good and the pain points of public transportation are obvious.

Now that the target area has been selected, the next step is to use the cold start method. The most effective tools we use are scenarios + ground promotion + events .

There were two options at the time. The first was to organize an event . Since Huilongguan Shopping Mall is the largest business district in Huilongguan, we cooperated with Huilongguan Shopping Mall to organize a store anniversary celebration. At that time, we distributed flyers in Huilongguan Shopping Mall. Although there were many people in the mall on weekends, the conversion efficiency was terrible. The activity focused more on brand exposure, and with the help of the cooperation with Huilongguan Hualian store anniversary, many local people knew the name "Didi Xiaoba".

The second way is the field promotion that can truly achieve conversion results. At the beginning, we distributed flyers at the entrance of Huilongguan subway station, using the earliest ticket-style DM flyers. The conversion rate was 2%, which means that only 2 out of 100 people converted. I was very dissatisfied with this result.

At the field sales site, if I saw someone not accepting the DM or throwing away the DM, I would shamelessly go up to them and ask them the reason. Many people gave very clear and common reasons : Sorry, I don’t have any need, I live across the street, or I’m waiting for a bus or someone.

Since there is no demand, it is normal to throw it away. The next step is to find out where the demand is. I found that there was a bus stop 20 meters away from Huilongguan subway station, and many people were waiting for the bus. We shifted our marketing efforts to bus stops, and the conversion rate increased to 7%, which is more than a three-fold increase. Because the scene of "waiting for the bus" itself is a manifestation of demand, minibuses can just solve the pain points of waiting for the bus and getting crowded.

But at the same time, we found that some users still sent DMs. I continued to ask shamelessly, and the users said, "I don't know what you are?" What's the difference between you and a fast car? I don't understand this single page, what does it mean?

The question comes back to user education. Later we made a new big poster and put it on the body, and the conversion rate reached 9%.

Later I found that the user was still throwing it away, so I continued to ask what the reason was. User asked, what is a minibus? How big is the minibus? I know about buses, but what are minibuses? These problems truly reflect that users still don’t understand us well enough.

Finally, we drove a minibus for advertising. The driver held a horn, and the salespeople no longer needed to hand out flyers. We just had to arrange for one person to stand at the bus stop holding a sign, and the other salespeople would act as agents and stand there in line. When the bus arrived at the stop, the agents would get off the bus and say loudly to the driver, "Thank you, master. This bus is really convenient."

After stopping there, the station sales staff shouted that the Didi minibus had arrived at the station and was going to a certain destination. It was two yuan per person and the first ride was free. The helper then got on the bus. At this time, some of the other passengers waiting for the bus nearby just scanned the code and got on the bus.

At that time, we basically didn't need to send out single pages, users would actively ask for them, and finally the conversion rate reached 16%, and the weekly retention rate reached more than 70% . In fact, it was a performance that combined events and local promotion. This car travels between major subway stations and residential areas in Huilongguan.

This incident tells us that you have to find the right scenario, the right time, and make users understand in order to maximize efficiency.

The premise of growth is accurate demand and understanding what problems your product solves for whom and at what time.

We need to explore the essence of demand, not create it .

After clarifying the needs, we must do the right thing at different stages of growth, namely the seed stage, outbreak stage, and bottleneck stage , pay attention to different indicators, allow outbreaks and bottlenecks to appear alternately, and keep growth going. In the three dimensions of attracting new customers, promoting activation and retaining customers , there are different efficient growth methods in different periods.

Under the correct growth logic, we must grasp the five factors of growth, namely demand, scenarios, brands, channels, and user education . These five points are indispensable to achieve high growth.