In the two weeks before and after, Google and Salesforce respectively invested heavily in the acquisition of Looker and Tableau, which brought unprecedented attention to the field of business intelligence (hereinafter used: BI). Compared with the former, Salesforce's announcement of the acquisition of Tableau is more interesting. After all, Tableau is the creator of the world's leading visual analysis software. However, on the day the news was announced, Tableau stock rose nearly 40%, and Salesforce stock fell 5%. As Xu Fei, a geek time columnist, said, the market expressed appreciation for Tableau’s high price and some concerns about Salesforce’s high price purchase. .
That being the case, there are still many issues worth discussing, such as why the two giants chose to acquire these two important companies in the BI field at this time? What is the reason for the acquisition? What is the future development direction after the acquisition? Does this mean that the BI field will enter a new turning point? How do domestic BI vendors view the acquisition event? What is the future trend of BI? This article synthesizes and interviews the opinions of many experts in an attempt to answer the above questions.
Talking about two acquisitions
Regarding the reasons behind the two acquisitions, there have been many interpretations. You might as well select some valuable information from the official statement:
Screenshot from Google statement
It is not difficult to see that Google believes that this acquisition is in line with its multi-cloud strategy. Users will continue to benefit from the way the Looker multi-cloud platform provides services and continue to enjoy the data capabilities provided by different cloud platforms. At the same time, Google's BigQuery data analysis engine will be integrated with Looker to provide users with faster and more operable data insights, which will help promote business development and better serve customers.
Screenshot from Tableau's statement
For Salesforce and Tableau, the integration of the capabilities and user groups of the two parties is of great significance to future development, which will bring new growth points to the businesses of both parties. In this regard, Percent CTO Liu Yijing said in an interview with InfoQ that in the years of development, Salesforce and Tableau have gradually approached the ceiling of user growth, and new value points must be provided to continue growth. After user data has accumulated to a certain extent Later, it is logical to increase data analysis and visualization capabilities. Here you can refer to Microsoft's PowerBI, which is a completely cloud-based BI product and is well integrated with the Microsoft ecosystem. In recent years, it has steadily occupied Gartner's BI and data analysis leader quadrant. If you only rely on a single product, Salesforce and Tableau will have a hard time getting to the end.
As for why the two parties chose to acquire rather than build themselves, Liu Yijing believes that this is not a technical issue, but a commercial issue. From the perspective of technical strength, Google and Salesforce must have the ability to do data analysis, but the development of technology is to solve the needs of users. If there is no market, it is difficult to reflect the advanced nature of technology. Compared to self-built, it is obviously a better choice to acquire a manufacturer with a good technical strength and customer base and incorporate it into its own system.
In the past few years, Salesforce has been called the "acquisition madman" and has successively acquired Buddy Media, MetaMind, Demandware, Kurx, Quip, MuleSoft, etc. According to its insiders, it usually only takes three to six months to complete the acquisition of Salesforce. Products and technologies can be digested (integrated into Salesforce's products), which indicates that Salesforce will begin to integrate Tableau's products and technologies in the next period of time. Once data connectivity is achieved, customers can use both products through a single portal.
In summary, in this business activity, both parties have clearly found a positive part. Although Wall Street believes that Salesforce paid too high a price that caused its stock to fall on the day of its acquisition, there are also views that can refer to Facebook’s acquisition of Instagram. Facebook was also criticized for the same problem, but now Instagram’s value is 100 billion U.S. dollars. The return on investment is 100 times, and the same thing may happen to Salesforce. In any case, the value of BI began to be paid attention to by business and capital.
Behind the attention of BI
Looking back at the development history of BI, these two acquisitions have brought unprecedented attention to this field. However, many users do not have a good understanding of this concept, and it is difficult to make accurate judgments about the current state of development. Liu Yijing said that the usual concepts of data analysis and data mining are still too broad. In fact, BI is specific. This concept was first proposed in the 1990s and mainly includes ETL, data warehouse, OLAP, and data. Visualization and other aspects. The main meaning in the early days is to view and analyze the business from the point of view of data. The main idea is to use a series of indicators and dimensions to describe the business situation. Taking the financial situation of an enterprise as an example, it can be divided into indicators such as revenue, profit, and cost, and dimensions such as monthly, quarterly, and annual. Through the analysis of indicators and dimensions, one can understand the business situation of the enterprise and make strategic decisions and plans accordingly. .
Development History
In the first stage of development, BI mainly includes four layers: ETL, data warehouse, OLAP, and data visualization. Data visualization mainly refers to traditional reports. At that time, from the beginning of ETL data access to the presentation of the final report, the whole process was very cumbersome, time-consuming, and it took months or even years. Therefore, the application was not very extensive, and most of them were large financial institutions or Operators will use it, and the purpose is to understand the development of the business.
2010 年左右，BI 进入第二阶段。在该阶段，企业的数据量已经变得非常庞大，业务变化也非常快，传统的报表已经不能满足企业的数据分析需求。于是，敏捷式报表的概念逐渐深入人心。这个阶段，技术人员只需要准备好基础数据，剩下的数据分析和可视化交由数据分析师即可。这一阶段，BI 的流程已经大幅缩短，在决策中的作用开始被认可。在这个时代，我们可以粗略的认为，技术人员解决 50% 的问题，数据分析师解决另外 50% 的问题。借助敏捷式 BI 的发展，Tableau、Qlink 等企业在这一阶段得到了快速增长，Tableau 最终于 2013 年成功上市。
但敏捷式报表还不是终点。能否让技术工作变得更少更简单，让数据分析师更快上手承担更多任务？2017 年之后，BI 的发展进入第三阶段，智能化的概念开始对这个领域产生影响，各类 BI 应用的使用门槛得以进一步降低，比如自然语言等交互方式的加入，或者机器通过对数据进行分析给出可供参考的决策建议等。
技术门槛
上文提到，对谷歌和 Salesforce 而言，研发 BI 并不受技术水平所限，市场是更为重要的因素。对大部分 BI 厂商而言，技术反而是最直接的竞争力，可视化的页面设计并不是核心所在。通俗来说，能直观看到的东西都很容易模仿，但底层技术并不简单，尤其是智能化趋势的加入（AI 化）。
刘译璟表示，2017 年前后，百分点 因为在海外项目建设中发现客户存在此类需求，因此决定在大数据技术之上提供 BI 能力。这件事情本身比较顺理成章，因为早前（2010 年前后）大数据概念进入企业落地的第一步往往是 BI，通过对数据分析查看业务运行情况，二者的联系本就十分紧密，在既有大数据能力的基础上研发 BI 对百分点而言是很好的选择。
在实际调研中，百分点 发现商业应用的定制化能力存在不足，且周期较长，难以满足用户需求。于是，百分点决定构建新的解决方案，大量应用自然语言处理和深度学习技术，最终开发出如今的百分点商业智能系统。用户可通过自然语言与系统进行可视化互动，获取和分析数据，大大降低数据分析门槛；在数据分析方面，百分点商业智能系统通过对数据的智能分析，能够给用户推荐可能感兴趣的报告，并可以深度挖掘数据中隐藏的关系和价值，从而为商业决策提供有力支撑。由于该需求最初发现于海外市场，因此整套系统也匹配了多语言设置，目前支持简体中文、英语、法语、葡语四种语言。
在这个过程中，刘译璟表示，基础功能页面的设计和交互体验是可以不断改进和优化的，但智能化技术的演进并不是短时间内可以速成的，这项技术的门槛并不低，其背后的设计逻辑较为复杂，机器学习只是其中很小的一个环节。
发展趋势
百分点目前在“增强型分析“层面进行了大量探索，包含智能数据发现、增强数据准备、增强数据分析这些模块。百分点围绕“增强型分析”已经构建了图表推荐、智能建议、智能问答、图表见解四大能力，自助式分析通过结合图表推荐、智能问答等增强型分析功能，赋能用户分析思维，让普通业务人员和高管快速成为“公民数据科学家”。
刘译璟认为，在 BI 向第三阶段演进的转型期，“增强型分析”使 BI 成为更易用的产品，NLP 和 AI 将会是现代 BI 的主要特征。同时，在“增强型分析”方面的技术突破，使 BI 打开了一个增量市场。随着数据持续发挥价值，数据可视化需求水涨船高，未来的 BI 产品形态可能会发生改变，以 IaaS 的方式让越来越多的人应用起来。
目前，Tableau、Qlik 等 BI 产品已经加入了“增强型分析”这一特性，而在国内这样的产品还不多见，但很明显这会是未来一段时间内的重要趋势。这一名词同样出现在 Gartner 对 2020 年 BI 市场的预判趋势中。根据 Gartner 的报告：
By 2020, enhanced analytics will become the main driving force for new users to purchase BI products, data science and machine learning platforms, and embedded analytics.
By 2020, 50% of analytical queries will be generated through search, natural language processing or speech, or generated automatically.
By 2020, organizations that provide users with access to internal and external data curation catalogs will gain twice the business value from their analytics investments.
By 2020, the growth rate of the number of data and analysis experts in business departments will be three times that of IT department experts, which will force companies to reconsider their organizational models and skills.
By 2021, the two functions of natural language processing and conversation analysis will increase the utilization rate of analysis and business intelligence products from 35% to more than 50% among new users, especially front-line workers.
At the commercial level, many users expressed concern about the prospects of domestic BI vendors. On the one hand, looking back at Tableau’s financial reports, it can be found that since 2015, Tableau has never achieved annual positive profits, and the annual loss in 2018 reached 277 million U.S. dollars. This has to make people think about whether companies that are already in a leading position in the market are not profitable, and whether other manufacturers will be more difficult. Liu Yijing believes that this may be related to the business strategy chosen by Tableau. The company has profitability, but it may be more inclined to invest funds in technology research and development or ecological construction in the early stage. Costs increase, and profits are naturally compressed; on the other hand, two The acquisition may affect domestic cloud vendors and initiate a round of acquisition frenzy. Liu Yijing said that as to whether cloud vendors intend to acquire or not depend on their business development, for some data-based vendors, their own BI capabilities are already fully available. For other cloud vendors, if they don't want to make profits solely through infrastructure, they may consider cooperating or acquiring related companies.
Concluding remarks
In November 2018, the U.S. Department of Commerce's Bureau of Industrial Security (BIS) issued an export control framework for key technologies and related products, including data analysis technologies such as visualization, automatic analysis algorithms, and context-aware calculations. This not only shows the importance of this technology, but also provides a good development opportunity for domestic BI manufacturers. If you seize this opportunity to move forward, you should have a breakthrough in competitiveness and user growth.