Shulou(Shulou.com)06/03 Report--

This article is based on the keynote speech "Make Data Your Killer App" delivered by Alex Lu, the data director of American travel platform Lime, at the data-driven conference of Shenze 2019. This article will focus on the nine major ideas of big data's success, including the following:

Data culture

Quality first

No fast, no break.

Data citizenization

Decision closed loop product and operation closed loop

Data security

Make good use of artificial intelligence

Use one's strength to fight

"turn data into killer APP."

In the smartphone era, everyone has a lot of APP on their mobile phone, and we use all kinds of APP every day. Without APP, our life will be inconvenient. In big data's era, data to business practitioners is like APP to individuals, we have to use it every day. Without data, the work of practitioners will become difficult. We need to use data to report performance to our boss, use data to help us make daily decisions, and use data to make our product operations smarter. We want to do data like a killer APP, hoping to turn the data into a daily rigid demand APP for practitioners.

"my first time."

I studied automation at Tsinghua University and electronic engineering abroad. I first became acquainted with big data in 1997. At that time, AMGEN, which is now a drug company with a market capitalization of hundreds of billions of dollars in the United States, did not have its own data team. We were invited here to help them build a data warehouse. At the project launch meeting, the other party attached great importance to it. All the project participants were excited by his COO speech. At that time, my English was very poor, but I understood two sentences. The first sentence was "turn data into a competitive advantage of AMGEN", and the second sentence was "there is no shortage of money for bold work."

Through half a year's efforts, a team of more than a dozen people completed the project. The effect of the follow-up project is amazing. Data warehouses make AMGEN smarter to make decisions, increase product sales very quickly and make a lot of money. As a well-known pharmaceutical company, AMGEN in addition to spending a lot of time and energy to develop new drugs, its core business is to sell drugs. They have a strong marketing team and spend a lot of money on advertising, which is now called "branding". At the same time, they have to make efforts to open up pharmacies, insurance companies and other sales channels, according to the current words is to "build channels." More importantly, they need to take care of the doctor network, because only doctors can sell their prescription drugs, which is now called "conversion". The data warehouse we built has indeed become a competitive advantage for this pharmaceutical company.

In fact, I have done other projects before participating in this project. I have done the development of WEB web development, electronic, commerce, CRM and other systems. But I didn't know what I was interested in until after I finished the pharmaceutical project, I knew that the data was what I had to do. Despite the technical challenges, what really attracts me is the difference that data can bring and the value of real money.

"do Internet data"

Later, I was lucky enough to join Yahoo in the early days of the Internet and became the first engineer to do data. Although there was no such thing as big data at that time, the amount of data was indeed very large, and it was growing rapidly. To sum up my seven years of work at Yahoo in one sentence, that is, "how to use data to better understand users, understand thousands of users, understand hundreds of millions of users, and continue to provide them with better services", all for users.

Later, at an industry conference, I met my later colleagues at Google. At that time, I asked him, Yahoo traffic is ten times that of Google, but the revenue is similar to Google, why on earth? He said, if you want to know the answer, join in. In this way, I joined Google. It is also because of this curiosity that I have an in-depth understanding of Google's commercial products and profit model. In one sentence, I have been working at Google for 6 years, that is, "how to use data and technology to bring more value to customers while ensuring the user experience."

After leaving Google, I joined Baidu. In Baidu met Wen Feng and Shenze friends, in Baidu 6 years, my work has always been around big data, cloud computing, artificial intelligence, broadening horizons. After leaving Baidu, I joined Pinterest. Pinterest is the ancestor of Photo Waterfall Stream APP. At Pinterest, we began to practice the idea of making data as APP products for the first time, making data a daily tool for everyone in the company. "do O2O + Internet data"

A year ago, I took a small step across the border and joined Lime. Lime is trying to redefine travel, using green electric scooters to solve the "last kilometer" problem, where data is everyone's daily rigid demand. We continue to use the idea of doing APP products to do data, and take data, cloud computing and AI into consideration, in order to accelerate product innovation closed loop and fine operation closed loop.

Twenty years ago, achieving data value was an art that required harmony between the time, the place and the people, so almost 80% of projects failed. Today, the realization of data value depends on technology. According to a 2018 statistics, more than 80% of enterprises regard data and AI as their strategic priorities. At present, no one has ignored the value of big data. The famous investor Mary Meaker mentioned in the Internet trend report in 2019 that successful enterprises before 1995 gained competitive advantage by grasping data manually. From 1995 to 2000, they gained competitive advantage by relying on electronic data captured by the Internet. In the two decades since 2000, they have made good use of big data / AI technology to achieve more data value and gain competitive advantage.

At present, data / AI technology is dazzling. So many options also make it difficult for big data practitioners. Perhaps in the future big data and AI technology will be integrated into the head of several major platforms. As a practitioner of big data's technology platform, it has become a dream for many people to become a part of the head platform. But at present, the technology has not yet developed to that stage, and the successful big data still has a lot of challenges. Next, I will share with you some thoughts about being a successful big data.

1. Data culture

People always say that data culture is a matter of CEO, yes, without the support of CEO, there is no money to invest. But companies can't expect CEO to understand big data. The person in charge of big data, the person in charge of each business of the company has the responsibility to clearly describe the data value and vision of the company, show it to CEO, also show it to the whole company, and promote the data culture of the company.

Anyone who has read The Lean Startup knows that data is the key to iterating and closing the loop between ideas and practices. Nowadays, data science is very popular, using statistical and artificial intelligence algorithms to find rules from data that can be applied to business and products, in which data is also the key. Lime places great emphasis on data culture. Product innovation closed loop, fine operation closed loop, in many O2O companies are very important ideas. For Lime, the current product is a green travel scooter, and the operator is to put the right scooter in front of the right person at the right time. Because data has become the key to the two closed loops, the use of data has gradually become a daily habit. It is difficult to call it culture if it has not become a habit. Culture is something that you will practice unconsciously. The reason why Lime values data so much is that we know that for startups, no one can predict what the future will look like. But we believe that those who learn quickly will win. The speed of learning depends on data, which can let us know where to succeed and where to fail, find out the reasons, and constantly correct mistakes and try new things. In such a culture, there are signs of rapid growth of the company.

two。 Quality first

When we recruit data talent, we can usually use a simple question to tell whether the candidate really knows how to do the data, which is "how to solve the data quality problem". You should have heard of garbage in / garbage out, wrong data is more terrible than no data, there are many examples, companies rely on data for a large number of decisions, but it turns out that the data used in previous decisions is wrong, and the negative impact of these wrong decisions may be worse than no data to make decisions. High-quality data represents people's confidence in the data. Only when it is credible, people can safely use it in their daily work and become the new normal.

3. No fast, no break.

There is no fast but broken, everyone talks a lot, the Shenze data led by Wen Feng has made a very good product in this respect. So that you can quickly build a data system, at the right time to give you the data to use and generate value. Recently, we have talked a lot about doing the "data lake", and its original intention is to speed up. Usually, after the release of a new product, there are layers of processing from data retrieval to data availability. The data lake eliminates these links, and then a user data query tool can use the data in a timely manner. In this way, whenever a new product is released, you can use the data to verify the iterative effect of the product at the first point in time. Late data will be of less value.

4. Data citizenization

When data becomes a corporate culture, and when data is the daily need of everyone in the company, we must use the idea of APP to do data. Make it possible for everyone in the company to use data in actual work. Data is no longer a dedicated tool for managers. The more people use data for decision-making, the more valuable it will be. It is not easy to popularize data. First, in order to popularize the data, the APP must be extremely easy to use and the threshold must be lowered. Second, everyone has different ways of querying data, and the results may be different. It is necessary to make a clear distinction between standard answers and free answers to reduce the risk of errors caused by the civilian use of data. Third, from the perspective of data regulation, the more civilian the data is, the higher the risk of disclosure is. It is necessary to strengthen data security and use operating standards.

5. Decision closed loop

No company in the world has perfect data. No company has all the data. Many of our decisions are made when the data is incomplete. Therefore, in many cases, it is more appropriate to talk about data support. Data should not replace the dominance of decision makers.

6. Closed loop of product operation

Small-step fast running and fast iteration has become a routine task for doing products, and its core is product experiment automation. This approach can also be extended to operations. Take the product as an example, after having an idea, quickly make a prototype, and then do the Amax B experiment in the experiment to automatically generate comparative statistical results and suggestions for the next step. Many of the data scientific measurement methodologies are automated here. This is a specific big data accelerated product iterative innovation killer APP, which is already a daily standard tool for product managers and engineers in many companies. Before the automation of the experimental platform, my former employer could only do about ten experiments a day, but after automation, he could do 1,000 to 1,500 experiments a day. The experimental platform greatly speeds up the iterative closed loop of the product.

7. Data security

When I was working on my first data project, I dabbled in data security. In the Internet era, the importance of data security / user privacy has reached a new level. Recently, the European Union issued the GDPR Act, and many new data security regulations and regulations have been issued in China. The era of data compliance has come, and data must respect the privacy of God's users. According to EU GDPR requirements, companies may be fined 4 per cent of total revenue if they disclose users' privacy. In January 2019, Google was fined 50 million euros, with Facebook fined $5 billion by FTC as an example. "improper data handling" is no longer an excuse for exemption.

8. Make good use of artificial intelligence

Big data, who does not mention artificial intelligence, is incomplete. It is often said that in the era of artificial intelligence, big data is king because there is an inseparable relationship between artificial intelligence and big data. One specific approach is to use artificial intelligence to improve data quality. How to maximize the value generated by the integration of the two is what we need to think about. Digital transformation is the most frequently discussed topic in traditional enterprises in recent years. Generally speaking, the road of traditional enterprise transformation is to collect more data and to use artificial intelligence.

9. Use one's strength to fight

Which platform should be chosen? Is it a private deployment or a public deployment? It is difficult to recruit people, do you want to set up a data team? To answer these questions, it may be a good strategy to use your strength. Especially for small and medium-sized enterprises and start-ups, seizing the market may be more important than independent research and development. Countless examples have proved that you can be a successful big data by standing on the shoulders of giants. When using the leverage strategy, we also have to deal with the following challenges: first, cross-vendor system docking and optimization, second, integration and simplification of the platform, third, data analysis and data science, there is still little power to borrow, team building is inevitable.

Conclusion

To be a successful big data must be directly linked to value output. Big data, who does not talk about value, is a hooligan. Although data is not omnipotent, it is almost impossible not to make data. Data is the fulcrum of leverage, and enterprises that make good use of data may pry huge commercial value. It was an art to realize data value twenty years ago, but today it depends more on technology. I hope everyone can turn the data into a killer application of the company, so that they can stand in an invincible position in the fierce competition.

The city notice is about to start.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.