On June 5, 2021, the 2021 China Developer Ecosystem Summit hosted by SegmentFault came to a successful conclusion. At the meeting, CMO Kaye (Jiang Hanyu), a netease intelligent enterprise, delivered a speech titled “Explore the Rubik’s Cube for Business Growth with Data Thinking”, and explained some knowledge of other sectors from the perspective of the market.
Guest: Kaye, CMO of netease Intelligent Enterprise
Sort and publish shorthand: SegmentFault Editorial Department
Actually, I happened to be in Beijing on business last week, and I have been in Beijing for four or five days. But I misremembered the time we shared this time. I thought it was June 15th, but later I found it was June 5th, so I flew to Beijing again a few days later. Although the process is a little difficult, but I am very glad to have this opportunity to share and discuss with you. Just now I have listened carefully to the wonderful speeches of several teachers. In contrast, my content is not so closely tied to developers in a strict sense. You can take it as a change of perspective to understand some knowledge of other sectors from the perspective of market.
I am now serving as CMO of netease intelligent enterprise, mainly responsible for the overall market work of netease intelligent enterprise. Although you are not familiar with netease’s TOB business, netease has been expanding its B-end business. Our netease Cloud Service started in 2015 and has now entered its sixth year. Netease Yunxin mainly provides IM and audio and video technology services for developers, so this product is closely related to developers. The other two businesses are netease Eshield and netease Cloud Business, which are respectively the leading digital content risk control service providers and one-stop business growth service platforms in China.
I am outgoing and outgoing. I like traveling and running in my spare time. I am a loyal fan of Elon Musk and I like NASA very much.
Now, let’s officially start today’s sharing.
I’m sure many of you are familiar with the word VUCA, which stands for uncertainty and variability in the world. In such times, I believe that each of us is constantly seeking new knowledge and enriching our own insights. The terms MVP, growth hacker and AARRR mentioned by the guest just now have come into our perspective more and more frequently in the past two years. In fact, each of us has been constantly adding new knowledge. I often tell my team members that in the era of growth, only a restless person can be an excellent marketer. This “restless” growth of the hacker mind gives us more possibilities.
As I have just mentioned, I am in charge of the Marketing team. Our team has a very important job and is responsible for the dual duties of “brand building” and “lead introduction”. From the first day after the establishment of the team, to put it plainly, our KPI was strongly tied to the whole customer acquisition trail, including traffic acquisition, user conversion, etc. What I’m going to share today will be based on the actual work of our marketing team, including exploring some of the lessons learned in the process of acquiring customers for TOB marketing.
When the teacher shared with us just now, I asked if there were any friends involved in marketing. I saw a lot of people raise their hands. In fact, for us marketing people, the most headache or the most concerned point is “where is the traffic”. At present, this chart is also a summary based on the actual work of our team. We have been paying attention to the development of the whole industry, and we can clearly feel that the market of B2B industry has maintained a very good growth trend.
Before my presentation, I ran into a couple of VC friends outside the door. I believe you have noticed that VC is paying more and more attention to enterprises in the TOB field. You may also know that in the past two years, news of some TOB enterprises going public has emerged in an endless flow, and the whole industry is showing a good trend. At present, the overall usage habit of users is still based on the PC terminal search, and the mobile terminal is gradually improving. Many of you are developers, and I believe that you will often search the content you need during your working hours. This overall trend is very obvious. Therefore, the traffic distribution throughout the year is relatively average, mainly concentrated in the working hours of weekdays. These traffic related data are extracted from our background, and are also based on the observation of the overall traffic of Baidu.
In terms of retrieval features, generic words are mainly the core, and brand attention is gradually increasing. Meanwhile, the knowledge requirement of the whole concept is also increasing. In addition, the proportion of after-sale attention also makes us pay more attention to the relevant launch of the post-sale link. We can also extend it here. Although Yunxin is a technology-oriented product, the vocabulary of common words we put in has reached 150,000, which I think is still a very large amount for TOB’s business.
Next comes the third part. As we all know, the whole B2B decision-making cycle is generally longer, which will also lead to a longer decision-making chain. Therefore, brand influence and brand value are becoming more and more important. As a marketing person, I can clearly feel the changes of everyone in the past two years, especially when I attended some summits, including technology summits and industry summits. In fact, I can see that everyone hopes to improve the influence of the whole brand as soon as possible. In terms of the media path, the B2B path is longer and deeper. Nevertheless, we still value content marketing very much, because it can affect every link in the decision chain, so content marketing must be done. I believe many of you here today is to do the developers of ecological, and even predecessors in this field, you can feel that, if the company testing requirements for you is, you must be within three months from the community or a period of time into how many business opportunities, such as pressure of KPI for you must be very big, but for us, In fact, all kinds of content output, such as activity salon, are collectively called content marketing. At present, knowledge platforms also carry a large number of search requests, including domestic Zhihu, Toutiao and so on. Many people have regarded them as a second kind of search engine. In fact, the deployment of quality content on these platforms is very valuable to our overall brand impact and transformation.
As we currently provide cloud services, most of our customers are distributed in first-tier and second-tier cities and developed coastal areas, and users and purchasers are not the same group of people. Want to deal with this kind of situation, in fact is that we just mentioned, pay attention to the whole decision-making link, he had at different stages of their customer journey they decided to focus on the contents of his direction is different, even the products for the developers, when he was at the time of selection, comparison, his focus content direction is different, These are better guides to how to deploy different things.
Now that I’ve talked about the industry as a whole, let me share a little bit about what we’re actually doing. In China, you have to admit that you want to acquire users, whether it is short term users or the company requires you to acquire new users, we must be very clear about the types of traffic, which customers are short term customers, which customers need to be cultivated for a long time. We divide the traffic into three types of traffic, which are the bottom flow, the base flow and the incremental flow.
Bottom traffic is what we often do now, similar to channels like Baidu, sogou, 360, Google, and Toutiao. Take netease Yunxin as an example. In fact, after we launched the developer-oriented service in 2015, we spent less than half a year to construct the bottom traffic, mainly to build some common channels, and also to run the whole marketing and transformation link. Cloud letter operation to now, the bottom flow has actually begun to continue to provide some stable high traffic conversion, which is in fact every company should insist on doing. But as the business stabilizes and the marketing links run smoothly, we will pay more attention to the base traffic. Basic traffic will be more similar to brand effect, including some external alliances, such as Baidu Alliance, Sougou Alliance, 360 Alliance and netease youdao Intelligence Selection, etc. We will get through all of these. We currently have for the basic traffic plate to do information flow advertising, such as some brand advertising, we can also see intuitively, with a good conversion rate. The third piece is to do incremental traffic. As the whole B2B business industry continues to flourish, especially after 2020, there will be more people who compete for traffic, and the traffic bottleneck will obviously appear. Is also keen, we found that the so-called incremental flow is actually more high exposure, but the conversion rate is low, such as baidu small program was introduced in 20 years, it is actually more high exposure, but the conversion rate is not so high, but it can achieve high exposure is also very valuable, especially for baidu system. Baidu system will support self-built ecology, including our so-called know, encyclopedia, Q&A, in fact, such as the above self-built ecology of Baidu can even be understood as free flow pool, is worth a good deployment.
Our topic today is “Exploring the Rubik’s Cube for Business Growth with Data Thinking”, which is more about where our traffic comes from. And how we did it. How do we use data thinking to increase traffic? We know that data tracking is not only the eye of promotion, but also the cornerstone of growth. I think you are familiar with some of the work we are doing. As I mentioned, we have 150,000 general words, and we will add some parameters and hyperlinks to each promotion, which I think is very important.
For example, last week we held a conference with a scale of 1,000 people in Beijing. All the registration data, attendance data and WeChat group data were QR codes with parameters added. For us, all traffic tracking is actually a consideration. Where the flow finally falls and how the transformation results are, we have always been paying attention to. Whether it is some activities or promotion and publicity operation, it is necessary to carry out data tracking. As for the specific way to do it, it should be decided according to the project situation. Because some channels do not allow you to put parameters like LinkTag, but the tracking methods are also relatively diverse. Many of our netease data are buried by our own set of self-research system, called netease Hubble, through which we can track users’ habits in a more comprehensive way. After setting up our channels, there is a very perfect evaluation system. At present, after the introduction of all traffic, we can directly track it to the final order. We can synchronize it to the background or the database side of UBA or CRM when the user registers, so that we can know the source of every clue and evaluate the channel effect.
After we introduce all the traffic, will be directly traced to the single. In this process, we will evaluate from three dimensions: basic index, product index and income index. When it comes to building channels, the first thing to do is to set up basic indicators to better evaluate the size of the channel. How is the drainage effect? How much exposure? How many hits? For product metrics, we evaluate the quality of the entire channel, the core page bounce rate? Length of visit per person? At the income index level, we will directly track the traffic I attracted. Will the order be completed? Can it bring results to the company?
Our landing pages will also be divided into several kinds, functional landing pages, and there are basic flow landing pages. We will also focus on channel quality, such as the flow of some core page looks a lot, but the bounce rate is very high, this is about to consider is out of the question on the content, language is on the content such as lead and keywords with the user to the page after it receives the information does not match, led to jump out the user, We have relevant personnel to pay attention to and follow up on these issues. On the growth level, if the page is hosting 1000 visits, but only ends up converting 10, in some dimensions, we have a lot of room to grow. From the perspective of input and output, we now have a formula for the calculation of market input and final order results: LTV/CAC. LTV mainly refers to the revenue the customer contributes after becoming your paying customer, and CAC refers to all the costs we pay in acquiring traffic. Costing varies a little, but it usually takes into account marketing expenses, marketing staff, and labor costs for sales. At present, we think that if the number is divided by this, it can reach about 3, which indicates that the channel quality is very good. If it is more than 3, it means that the channel is very good, but if it is less than 1, it is like a loss-making business.
Now we have five sub-products under the whole product, and in a better case, we can achieve a figure of 5 to 6.
The point of all this is to tell you that we think that TOB’s data operations, or digital marketing, are essentially about using data analytics to drive growth. I think the concept of data-driven is deeply rooted in our team. We will think that the whole operation is actually a continuous process, which should not be done or not done, nor is it experience-driven, but should be data-driven.
The second important one is the familiar ABTEST. For some websites with large traffic on the C end, ABTEST is relatively easy to be implemented. Sometimes, if you change the guiding language of the registration button, the change of traffic conversion can be seen more obviously. However, for TOB, the traffic is actually much smaller. You’ll also be questioned about his input-output ratio. But I think anyway, anything that is validated and tested needs to be understood in a different way. I think we should firmly know ABTEST because it is also able to really use data to drive growth.
The third piece is to set data indicators. In fact, our company now has a large number of businesses, and in the early stage, each sub-product business is relatively independent development, and there are different sales teams, sales directors and different product teams. The marketing team is the only one I lead. In this process, you will find that people often have different language systems. In this business, he likes to call registered clients, while in that business, he likes to call Leads. He said he preferred to call it an effective customer, and then he called it a transfer customer. In fact, this would not be conducive to the latter-stage overall data pulling. This seems to be a very simple thing, but the same language system, the development of the same set of data index system, I think it is very important. Our marketing team, our data team, our sales team, all talk about conversion rates. Don’t argue or blame the other person, why don’t you follow up when this one works, this one doesn’t work or whatever. We all pull the data out, and as I mentioned above, there’s an SDR team, and I’ll tell you more about that later.
In the first part, I will give you a brief introduction of where TOB traffic is, what we need to do based on this traffic, and how to analyze the value of this traffic.
In the second part I will briefly share with you the promotion, transformation and growth of the marketing data closed loop. In particular, I have mentioned SDR as a team.
This chart is made by our team together with the data team, because we think that the whole growth routine is generally as follows: by sorting out our own business situation, we use data to describe the key indicators of each link of AARRR, and finally determine the North Star index of each growth project. After more than a year of trying. We’ve actually abstracted a growth solution for TOB. All the things in this picture are researched by ourselves, and we will also make some data burying points by ourselves.
As you can see, the top layer on this chart is the acquisition that I introduced to you in the first section, including marketing, sales, and data. This picture you can see traffic is in the top layer, I in the first part mainly introduces or traffic from search engines, but in fact we flow of source is very, very wide, in addition to the search engines also includes is similar to the offline conference today, also including our online activities, include we often do some depth of content marketing, Sometimes for zero dollars you can get a business growth manual or a white paper or something like that, and it works well. Netease is primarily aimed at developers, so we have a set of technical standards for audio and video translation, called the WebRTC standard. Up to now, we are still bringing us a steady stream of potential leads. So far, we have brought nearly 2,000 leads. The content we do is almost free for developers, but our content people spend a lot of time on planning, translation, content arrangement and channel deployment in the early stage. As you can clearly see from this picture, the top layer is where our entire traffic comes from.
The second level is transformation, where clues are filtered from MQL to SQL. For the corresponding data labels, we have separated the entire lead rating model and the sales force evaluation model, which will also be briefly introduced later. The rating and evaluation model of sales power mainly determines which sales lead is best matched according to the customers the sales person is good at and the characteristics of the lead itself. The last level in this chart is retention, which mainly includes sales and customer success. We can find that the flow of traffic always follows from the top to the bottom. If the customer loses, then our entire data analysis, including subsequent purchase and corresponding customer service, should be adjusted accordingly. In fact, the data burial point has existed since the launch of our Marketing Department. From the interaction of users, such as their registration or browsing, to their complete loss, or even the possibility of becoming my customers for me to further follow up services, they are all included in our entire data growth model of TOB.
We just talked about the flow problem, in fact, B2B clue flow path or have a relatively ideal flow closed loop. I think about flow closed loop can be summed up in simple words, expand the flow funnel, registration timely response, clues efficient allocation, release the sales effort, market cultivation, drainage system optimization, it can improve our overall efficiency by guest, in the whole link, actually see the depth of marketing and sales is very fit.
As for the closed-loop practice of B2B marketing data, I think as I said just now, people must have the same set of thinking and language system for a lot of cognition, where we call it Tao and Art.
The Tao is that we have to be customer-centric and a highly coordinated growth flywheel. At this point, customer-centric talk may sound hollow. But I think a lot of times, especially in the actual work, there are some core ideas, some culture. Still need to continue to pass, we actually under the foundation of take the customer as the center, is a rhetoric art requires a unified system, target and strategy, system and technology, because earlier we have mentioned the whole product business is has different sales team, sometimes in the meeting between different product team will find there are some communication barriers, I’m sure it’s obvious if you have that many different businesses, that many different products, and that many different product teams in your company.
Just gave you a brief introduction, but in B2B marketing is actually more practical, which is the next I want to share with you, here will be divided into three parts to share with you. The first part is just talked about, to break through the marketing data, to track the user status, is to use the data to speak. The second part is the strategy of the thread circulation. We will continue to design the refined thread circulation link according to the actual situation. The third part is the talent management of SDR. I started to introduce the SDR team in 2018, and it was formally established in 2019. From now to 2021, our entire SDR team has taken over almost all the lead traffic of netease smart enterprises. They will release some energy of sales, which can greatly improve the conversion rate.
The first part is to use data to speak, which can be divided into three parts. The first part is that the source of Leads can be traced and the user behavior can be analyzed. As you can see, all the data can be associated with an ID in the end.
In China, due to the characteristics of national conditions, most of us now take mobile phone number as the unique ID. Therefore, the source of Leads is traceable and the corresponding user behavior can be analyzed. Our user behavior analysis tool is developed by us, and our customer relationship management system is also developed by us. We can summarize all the customer data under one ID and all the paths of the user together.
To source for this figure, which was in front of me to everyone when it comes to all of our promotion, whether it’s activity, the content or search engine, is will with parameters, with a channel number, can help us better for data analysis, including page sources and user access time recently, Give us a more complete understanding of the user’s behavior path. On the right, you can see the information users fill in when reading our articles, and the data actively filled in by these users will also be recorded in our system. Based on data, we have also built a marketing management cockpit. In fact, we hope to have a full-link data Kanban from market to sales to service. Now, when we hold weekly meetings, we will look at the data first, not to mention other things. You can see clearly from the data board which part of the problem, which part of the transformation is down.
The second point is to standardize the filling of CRM, do a good job of data precipitation and governance. I still have some feelings about this. I remember that before the SDR team was established, we wanted to analyze why the registration data was invalid. As we all know, invalid data would be thrown into the high seas pool. In fact, the reason is very simple, because everyone is casually to fill in the information, has caused many data can’t do a structured analysis, the export data is in a mess, so after I build the SDR team, have asked us to every stage was to fill out the very standardized, so as to make follow-up data analysis more accurate and clear.
After the implementation of the previous part, it is very important to carry out a fine design of the thread flow link. The one on the left is a picture of foreign countries, and the one on the right is the process of our team’s corresponding lead circulation. In fact, the whole process is the process from the flow of leads to the market recognition and then to the sales recognition, the order making or the order losing, and finally the re-operation. The whole design of a link is very detailed, this requires market classmates or SDR classmate, still say we sell my classmates have to be very clear, I want to do and in different stage should pay attention to what is the point of, for example, I am director of sales to communicate with us, at that time my data already satisfy certain conditions, It can be transferred to the sales person for them to follow up. When the sales person thinks that this is an invalid lead, I will also recheck whether it is because some necessary conditions are not met or because of the sales person’s personal judgment.
Just told everyone to our ratings are a set of system for the clues, and when the user through online registration or other means to our platform, their data is not directly to the sales or Marketing Department, the data will also pass us a clue to the rating system, to automatically calculate his scoring, L1, L2, L3 at three levels. So you’re going to ask, what’s the grading? In fact, we have a feature weighted scoring model, which is mainly led by the Marketing Department and will make regular model modifications together with the sales team. The model is Lead Score = W1F1 + W2F2 + W3F3 + WIFI, where WI is the weight of each feature and FI stands for features, for example, If the user filled in the registration information is to contain a complete enterprise name, a complete personal name, as well as a complete mobile phone number, then he has the corresponding characteristics, his weight will be relatively high. When I was just talking about population characteristics are also said to our current most clues from the first – and second-line cities, that if the registered clue is from Beijing, or is in guangdong, the relatively speaking, its weight will be slightly higher, and companies such as the user is Top100 Top industry is strong, it weights will be higher. In addition, since we can capture the registration data, which pages he browsed or even downloaded our Demo before formal registration, once he has these interactive behaviors, his weight will be increased. As for the specific weight points represented by each W, We adjust our standards for different stages and conversion rates. We have been doing this for nearly six years now, so we have a huge amount of user data, so our SDR team will also stratify user data according to the scoring model when we do data tracking and data fishing.
The characteristics of our Scoring model have been introduced to you before. Here, I would like to show you in detail that the characteristics of our Lead Scoring model include the behavior of official website, interaction information, company information, product use and registration information, all of which actually become the characteristics of our Scoring model.
Because every company has its own characteristics and the actual situation of different, so if you company want to do a similar score model, we would be doing the model test the data out of the past to clinch a deal before, what kind of characteristics to analyze the clues are relatively conversion and sell-through rate will be higher some, and then the corresponding data abstraction, Form the final model. In addition, our data team will also be able to interface with such interfaces as Skycheck, because as the Marketing Department, we are familiar with some famous products. For example, I clearly know that Douyin is a product owned by Bytedance, but it may be unfamiliar to the sales staff. Through docking such systems as Sky Eye, it can help us identify whether the registered enterprises are Top level in the first time.
As mentioned just now, between our Marketing Department and sales department, we will conduct data cleaning through SDR. Only after the registered users meet the lead standards, can they pass our cleaning smoothly. As for the screening criteria of MQL, Bant is A common one in the market. B stands for budget and A stands for influence, which actually involves the skill of our SDR team. When the team members go outside to call, Can through the standard operation to ask the user whether have enough budget and the registrant have decision-making power and influence of his company, the third point is to go to asked him whether the product requirements we offer the service to meet, the fourth is to ask specific project start time, they might want to do a project, but still want to do in the stage, The real project launch will be half a year later, so we will try our best to screen out the projects that are in the model selection stage and are already in progress, so that the sales can quickly follow up. Our SDR plays a very important role in the initial screening of these clues against the criteria.
Just now we have been talking about the term SDR, and some students are wondering what is SDR on earth. In fact, the Chinese name of SDR is sales development representative. We also researched a lot of companies when we started our department. Currently, there are two types of responsibilities and division of labor for SDR in the market, but different companies may have slight differences, but in general, they are divided into these two types. The first type mainly cleans the market through external call, obtains clues through market cleaning and classifies the clues. In fact, this kind of SDR is more to do some passive basic external calling work, and does not need to external expansion by itself. However, the other type is the B type SDR that needs to be actively called outside. They should take the initiative to expand the SDR just like sales. Such SDR can get commission. The first type of SDR he gets is the base salary and assessment salary. For example, our company is currently based on Class A, and my SDR team is mainly operating in A relatively standardized process. The working process of SDR can be shown in the following diagram, including the introduction of market leads, sales follow-up and customer transactions.
The first is to clean the data of the potential customers. If it is the intended SQL users, they will be transferred to the sales for follow-up. If the sales can make a deal after follow-up, we will carry out our final SO service. Intent users who are not SQL will come into our aquarium, and we will breed them again. In the whole process, it will be combined with the data model rating we mentioned above. In fact, before SDR calling, the clues have already been rated. After the completion of the rating, we will follow up all his behavior data in the form of an ID until the final transaction.
Why an SDR team? In fact, we all know that although we all emphasize the importance of data labeling and timely CRM follow-up, in many companies sales are just sales, because they have strong KPI assessment. For sales people, they pay more attention to how to win the order. As for whether the remarks are clear in the follow-up process, whether the follow-up is timely and whether the label is perfect, these are not their primary concern. In such a case, our SDR, as a first-round data cue filter, can better improve the data and do a good job in data annotation, including data feedback. This actually goes back to what we said in the first part. In launching, we need very accurate data feedback, which is mainly accomplished by the cooperation and complement of our SDR and data delivery personnel. At the same time, they also play an important role in data upgrading, transformation and cost saving.
In terms of cost savings, I also have some practical experience, in my first year on the SDR, actually we are against the director of sales, this also is normal, because before all the clues and the data are in sales, sales are used to all the clues to knead in his own hands, they will think what clues as to sales to decide should be, They are reluctant to give clues. Against this background, it is indeed difficult for SDR to push forward in the first year. So how do we break through? How do we settle this matter? Therefore, in the first year, our SDR team only recruited three members. The first requirement was to make the increment, which means to make the data model from the discarded data of sales. As mentioned before, sales will throw useless clues into the high seas pool, so we will make the data model in the high seas pool. We think we can salvage the priority data, such as a batch of data within three months because of a function is not satisfied, or because of the price system and other reasons. There are clues, including our activity because a lot of company’s sales for clues to follow up the activities are not interested in, I believe that our developers conference on ecological? Many people today, if your company do ecological corresponding product developers, that you put these registration data directly to your sales, sales for the most part are not willing to follow up, Because he would think this kind of data is extremely inefficient, so our SDR started from here at that time. In fact, three SDRs are really not expensive. With a salary of five or six thousand yuan plus social security, each person can get less than 100,000 yuan a year, or even 300,000 yuan for three people. However, we made a profit of 2.6 million yuan through this kind of cleaning of high seas data salvage and other clues. You can understand it as that I spent 300,000 yuan on labor cost and gained 2.6 million yuan in revenue. It was really a good deal. After this event, it has also proved that SDR has great value in both data label improvement and cost saving. Gradually, SDR began to undertake the data clues of our whole netease intelligent enterprise. I think a lot of times, we think of data as a big, big thing, but I think data is in the day-to-day work. We think about whether data should be a set of very lofty systems, a set of very advanced theory, need various models. I think these are necessary, but what is equally important is that the early delivery buried points, the improvement of data labels and the daily follow-up of data labels are timely and accurate. Only when these bits and pieces are fused together can we really do a good job in data driving. This is also our real feeling after several years of actual combat.
Today is a developer conference, and as I said before I shared it, a lot of what I’m talking about doesn’t seem to be very developing-related, and I also told the host that I’m going to have a dinner at the end, and I’m going to talk about the Twilio case.
I’m sure those of you who are developers know something about the Twilio product. There are only 3 pages of this material, and this is actually a sharing made by BVP, the investor of Twilio. Twilio is a product for developers, aiming to provide developers with all the convenient tools. It was founded in 2008 and launched in 2017. BVP, who invested in Twilio in its early days and later led Series B, C and D, has grown from $50 million to $57 billion since Twilio’s initial public offering, an increase of 1,140 times. I believe its growth rate or rate of return must be the envy of everyone. As we’ll see, Twilio is a very typical company, very focused on the developer ecosystem, and very focused on the whole developer bottom-up dissemination.
In fact, I think he really emphasized the third point, that for a long time, they really focused on organic growth, bottom-up distribution of developers. Many of the products on the market now, including Slack or Zoom, are based on self-propagation systems. Twilio will organize some activities in the early days, including content marketing, but they are adamant that the only way to truly reach the market is through the developers themselves.
On the last page, BVP mentions that it is not so well known in China, but it is actually a very established investment firm in North America. In the sharing, they offer eight suggestions for the developer platform. In fact, they also believe that a developer platform with fine operation can provide a very good boost and support for sales. At the same time, if the whole developer ecosystem can be well built, it can actually bring into play the inherent natural value-added trigger mechanism of engineer resources and products. Developer sermons can also drive product usage and sales. His eight suggestions, in general, still emphasize the importance of building a developer platform so that our products can be better updated and iterated, and our cognition and engineering resources can be optimized and used to the maximum extent. Actually I think developers ecological netease doing this also is just beginning, even I don’t feel completely introduction, to be here today also, not to share with you, more exchanges, is to learn, as I have just talked about overseas cases, I also in learning some advanced experience abroad, also hope that today’s speech to be of help, thank you.