prologue:
However, changes are inseparable, and the core of enterprises' need for data is still to support operations and make scientific decisions quickly. The emergence of self-service better meets the needs of data-based operations for companies to reduce costs and increase efficiency.
I am very happy that there is a manufacturer who has been doing various innovative and leading exploration work, and she is Smartbi. At the just-concluded September 7th Smartbi2018 Big Data Analysis Summit Beijing Station press conference, Smartbi showed us her infinite possibilities, whether it is an innovative application based on Excel, and a thorough practice of "full self-service and true sharing", It is still an active exploration of the future of SaaS cloud reports and AI, which are all at the forefront of the industry and are worth looking forward to.
——Chu Ranran, Chairman of China Electronic Form Application Conference
The following is the speech given by Sematic Software VP Xu Jing at the Beijing press conference just past
host:
Sematic Software Market VP Xu Jing, as the first guest to appear, brought a speech with the theme of "Smartbi V8.5 Magical AI Big Data Analysis". Let us follow Mr. Xu Jing's wonderful speech and take a look at Smartbi's attempts and efforts in leading the era of "everyone is a data analyst".
01 Always change
Xu Jing:
Hello everyone, I am the screenwriter and director of the Smartbi brand. Today, the leading actor and CEO Wu Huafu was not present due to traffic control (affected by the China-Africa Conference). I will serve as an extra to introduce some thinking behind Smartbi V8.5. .
In the field of data analysis that global CIOs are highly concerned about, there are many new concepts and new technologies, from DSS, OLAP, BI, DW when it was born more than 20 years ago to Big Data, DT, Self-BI, Saas BI, I Trying to find the essence of it will make us think more calmly and objectively. After all, everything will never leave its origin.
First, let's look at the "game" that exists in every enterprise-the relationship between business and IT department.
Regardless of the early C/S architecture or the B/S products in recent years, the IT department always hopes to find a data analysis software that can match or even replace Excel, and the business department (even our BI vendors themselves) cannot give up this ease of use. Tool of.
Let's not discuss how convenient Excel's functions are for the time being. Only from a management perspective, can such contradictions be resolved?
I want to use the simplest way of thinking. In the past, the connection point between the two parties was "software", which is like an apple. We can divide it into two apples to distinguish between "data" and "analysis". Is it easy to find ideas? Up?
02 Thinking about the evolution of automotive tools
When Smartbi is classified in software, it belongs to BI_Tools. Since it is a tool, I want to make some categories with cars, which are familiar to everyone, and see the very interesting relationships among them.
Many years ago, a Santana was more than 300,000 yuan, and owning one is enough to prove the financial strength of a company or an individual. Ordinary people can meet each other (Beijing 20 years ago). Later, the auto industry began to compete. Through joint ventures and localization, cars became cheaper and cheaper (especially relative to housing prices). At this time, private cars entered the homes of ordinary people, and everyone could get a driver’s license. The profession of driver was also very successful. The career of money has become a skill for ordinary people.
Well, come back and look at the field of BI tools, is there a very similar situation. In the past, this kind of data analysis software was very expensive, just like Santana, and can afford to use this software, that is, institutions such as finance and telecommunications. Later, there were more and more similar software tools, and the price/performance ratio improved (just like a private car). Now every company has the financial resources to use this BI.
That person is the same, he was a report engineer at first, and now he is called a data analyst. The changes between these two kinds of people are just like when we were full-time drivers before, now everyone can drive, everyone is a data analyst. Let's take a look at the intensity of this change. Report engineers can search for training from the Internet, and they can search about 3.5 million items, and data analysts can find 25 million. This is when everyone is a data analyst and the driver becomes a skill.
We seem to see the trend of “self-service data analysis” from the perspective of management and the evolution of tools. Using Gartner’s prediction is: enterprise organizations are adopting self-service analysis and business intelligence (BI) and bringing these capabilities to Business users at all levels. This trend is very obvious. According to Gartner's prediction, by 2019, the analysis output of business users with self-service capabilities will even exceed that of professional data scientists.
03 unchanged demand: data-based operations
This shows that what companies need data for so many years is to support their own operations and make scientific decisions. But why there is such a trend of self-service is because of the simplest principle-the essence of operation is to reduce costs and increase efficiency. If the process of data analysis is not "fast" enough, no matter how effective scientific decision-making is, there is no cost Come down. Therefore, self-service has become a trend because it meets the two necessary conditions of data-based operation, fast and effective.
After getting this idea, I borrowed the idea of management consulting and created a "four-quadrant moment of data-based operation" with "fast" and "effective", and understood Excel, Smartbi and similar tools from two necessary dimensions.
Undoubtedly, the convenience of Excel is the best: data analysis, data modification, and data processing are very fast, so it is at the top from the Y axis; then why it is at the leftmost side of the X axis, which is also very obvious. Because of its poor data management capabilities, the original data and reports need to be shared together, there is no security, and it is easily damaged, and sometimes the performance is very slow.
Later, report software appeared, which was based on a database, so it solved various data management requirements, such as security, performance, multi-table association, and solved many problems. But it ignores a very important issue, that is, it sacrifices convenience. The cost of learning these report software is relatively high, so there was only such a profession as "report engineer" before. I remember that the previous report software included Anxun and CR.
Next back to Smartbi, there is a component in our product called a spreadsheet. It has made a breakthrough. The report tool must run on the WEB browser. We use a plug-in on Excel to directly design the report, but the data is stored in the database, such as Oracle, DB2, and Hadoop. At this time, the professionalism of data management is retained, and the convenience of data analysis has been significantly improved (the Y-axis position is higher).
Therefore, Smartbi spreadsheets have been recognized by many users in the past five years. For example, the earliest IBM delivery team used this product to do regulatory reporting projects for banks. As of V8.5 this year, the spreadsheet function has made many improvements in interactive controls, security watermarks, and data visualization. Welcome to log on to the official website to learn about specific functions.
Next, we launched Smartbi self-service analysis in 2014. The core function at that time was perspective analysis and combination analysis. It was very similar to OLAP software like Essbase. It can also be understood as an Excel perspective table, which can be drilled, sliced, and rotated based on the table. And other interactive operations.
Now, self-service analysis is based on data visualization, and the ability of tables is much weaker than before. V8.5 does not abandon the form, but adds the function of self-service dashboard. Our definition is called full self-service, which will be described in detail later.
In the future, many companies will choose to use SaaS to satisfy their own data analysis, which is also an obvious trend. Because this kind of SaaS BI provider will definitely guarantee data management capabilities and provide a convenient and easy-to-use functional interface. This is located in the upper left corner of the first quadrant, so the future SaaS BI will surely become popular.
Based on this judgment, we also tried a fresh product-Smartbi Cloud Report (planned by myself, welcome to try), and sent Excel to the mobile APP at the touch of a touch. At this time, the data is stored in Excel and there is no database, but why does it still reach The first quadrant.
This shows that it has simple data management capabilities, such as separation of data and reports (hiding sheet pages), fast data modification (no SQL required), and so on.
In addition, Smartbi cloud reports inherit the ECharts visualization capabilities of spreadsheets, which means that users can simultaneously render reports with Excel charts (conditional formatting, etc.) and Echarts charts based on Excel static data, and publish them to the APP with one click. It is very simple, Easy to use and fun.
At present, there are already some users who use cloud reports daily to complete progress monitoring, inventory management, and information statistics.
After introducing the four quadrants, everyone can understand the work Smartbi has done from V5 to V8, that is, regardless of reporting requirements, visualization requirements, and interactive analysis requirements, they are evolving to reduce learning costs.
From the perspective of the two pillars of V8.5, the first one is that we rely on Excel plug-ins, word plug-ins and PPT plug-ins (office BI), so that data can be completed in the Office that users are familiar with. Self-service analysis capabilities (self-service BI).
Why is it "full self-service"? I still want to use the car as an analogy. Everyone knows that cars, SUVs, MPVs, sports cars, and RVs. Why is there no means of transportation that can cover all the needs of users? If you understand this problem, you can also understand the reason why Smartbi advocates "full self-help".
Full self-service self-service analysis, first is the ability to obtain detailed data, which is a common requirement of large financial institutions, followed by interactive self-service analysis, which is the aforementioned analysis needs such as drilling, and the third is the need for data visualization. It is very convenient to make dashboards and publish them, and finally self-organize data. V8.5 has cross-database query and self-service extraction, and even the ability to process data and model calculations in the future.
Smartbi will continue to strengthen functions and simplify operations after V8.5 to achieve our ideal self-service analysis.
04 Be careful, self-service analysis has pitfalls
When all vendors are calling self-service BI, you better calm down, including hearing the "perfectionism" of full self-service. Let's take the automobile industry as an example to see where the traps are!
Look, when everyone is a data analyst (driver), it will be the same as what we encounter every day: traffic jams, gas queues, no parking spaces, and even traffic accidents. This is the companion product of self-service analysis—— "Performance Challenge"!
Why does the self-service analysis of business personnel cause severe performance challenges? My conclusion is that their usage is no where & no how.
First of all, they do not have the concept of optimization. It is very likely that the entire table will be grouped and aggregated directly, that is, there is no where condition. Secondly, vendors of self-service analysis often say a word called exploration. It is precisely this exploration that shows that business personnel sometimes have no purpose. In analyzing data, many are even learning how to analyze data and how to understand visual graphics.
These two situations have caused a huge resource occupation on the data platform, which drags down the response time of the overall system and causes performance problems!
Can the performance problem be solved? Unfortunately, there is no reasonable solution from a technical point of view, unless the cost is unlimited. Technology and management must be combined:
Technically: The first one is space for time. There is also such a function in V8.5-high-speed caching, which is the same as MOLAP. Before users can access the data, the data is already in a semi-ready state, thus saving the front-end report. Generation time.
For the second and third techniques, I did a laptop-based self-service analysis performance study of billions of data last year. From the initial unoptimized model to the relatively satisfactory model, this exploratory self-service analysis average response The time has been increased several times, and with the help of faster MPP database products, the final performance can be run within 5 seconds.
Outside the venue, we used a 5,000 yuan desktop computer, which increased the speed by 40%. You can experience it directly!
Management: The rational way of thinking of managers is to guide, so how we guide business personnel to use self-service analysis is very important.
Let me return to the analogy of transportation. Beijing has customized public transportation services. This is equivalent to when you can clarify your needs, you can combine them and provide them directly with "data products" to avoid exploring and analyzing data. process.
The second idea, if you still don’t know what common needs are, encourage everyone to share your own practices and ideas, similar to a "shunting ride" sharing scheme, so that new users are guided to see if there is an existing "analysis" first. (It can be retrieved and inquired from the analysis store), which can also avoid the waste of resources. Of course, there are many strategic guidance for management, such as platform cost sharing.
05 Looking to the future, BI will be smarter
Time is almost over. After listening to the full self-service, high-performance, do you think you have seen the end? I personally think that AI will completely subvert the tools of data-based operations. Let's take a look at the future of transportation, the automobile.
The combination of the Internet of Vehicles, autonomous driving, and shared cars gives our urban life a sci-fi future. Similarly, AI will also present the dialogue between people and BI software in a holographic projection like a movie.
Smartbi also has some explorations in AI, allowing users to operate the software through natural language and even perform exploratory data analysis.
Above, I described Smartbi's innovative concept through the four-quadrant diagram of data-based operation, and also introduced the main features of V8.5 through the analogy of transportation.
It should be said that Smartbi's V8.5 is an important milestone, marking our own significant progress. In addition to the front-end data analysis and display, it also adds data processing and data mining capabilities. On behalf of Mr. Wu and the R&D team, I welcome Everyone is here, and I hope everyone likes our V8.5. My speech is over, thank you!
(The above is the content of the speech)