Use first-principle thinking to think about how to build a platform to improve productivity

introduction

Now the whole society is engaged in digital transformation, from the government to enterprises, so why is it necessary to carry out digital transformation? In essence, social governance and business operations have become more difficult.

For enterprises, the goal of transformation is to achieve a quality life, such as earning more money or maintaining stable operations. The core of transformation is to build a management system with the help of digital technology to cope with external environment turmoil and corporate competition. Uncertainty brought about by changes and technological updates and developments.

Digital transformation will bring a lot of research and development needs. How to deliver these needs better and faster has become a prominent problem. How to build a platform to solve this problem? Can you use first-principles thinking to deduce the direction of development?

What is first principles

The first principle refers to splitting the problem into the most basic facts or laws, and based on these known information, continuous deduction and calculation, so as to find the optimal path to solve the problem.

Well-known examples of the application of this thinking are

The story of Elon Musk reducing the cost of rocket launches. Subdivide rocket launch costs, develop rocket reusable launch technology, simplify design and improve processes to reduce rocket manufacturing costs
The story of Henry Ford who lowered the cost of making cars. Disassemble the car into the most basic components, and use the assembly line and assembly line workers to mass-produce the car
The story of Elon Musk driving down the cost of car battery packs. Analyze the market price of raw materials for batteries, keep approaching the price limit, and adopt measures such as self-built factories, improved battery design, and improved production processes to reduce costs
The story of Jobs getting rid of the physical keyboard on the phone. Analyze the components of the mobile phone and expand the touch screen area by removing the physical keyboard, resulting in a better experience and lower manufacturing costs

If you think by analogy

Analogical thinking refers to the method of accumulating knowledge with the help of one's own feelings and experiences

Take well-done platforms at home and abroad as examples. Microsoft’s Power Platform and the start-up company Qingliu all aim to help companies quickly develop applications and respond to market development.

Microsoft Power Platform

In addition to the above 5 major products, it also has the following functions

Dataverse: It can manage data, which is equivalent to an easier-to-use database
Connector: can communicate with a large number of APIs, services, and systems
Common Data Model: A standardized and extensible data model that facilitates data transfer and utilization, accumulated over the years by Microsoft and industry partners
Ecosystem: Power Platform can connect to Microsoft's ecosystem, such as Azure (the world's second largest cloud computing platform), Microsoft 365 (including Office office suite and Teams), Dynamics 365 (including ERP and CRM functions)
Power Fx: low-code language, Excel-like formulas, easier to use than professional programming languages

The product concept of Power Platform is to enable people and developers without technical background to solve business problems easily and efficiently. The product architecture takes Dataverse as the core, the built-in data model (Common Data Model) reduces business modeling costs, the visual UI editor (Power Apps, Power Pages) can reduce application development costs, and the report creation tool (Power BI) can mine The value of data, efficiency tools (Power Automate and Power Virtual Agents) can free employees from boring and repetitive tasks. This whole set of combinations is expected to help companies develop their businesses better and save operating costs

Microsoft has done this, should we do the same? Right now

Develop a database that non-professional developers can learn to use
Develop a visual UI editor to support the production of mobile applications, portals, and reports
Develop efficiency tools to automate tasks and support desktop software, Web, and instant chat windows
Develop a programming language that non-professional developers can learn to use

light flow

All-in-One Application Development Platform

Proprietary light stream, multi-role-oriented enterprise-level platform products

Lightflow starts from the high-frequency scenario of form + process management, and is aimed at business personnel rather than developers. With the accumulation of industry cases and the increase of customers,

It has gradually developed into an application development platform for quickly solving business problems and an enterprise-level platform for medium and large customers.

It is worth mentioning that Qingliu, like Sensors, also adopts a subscription-based charging model, which ensures that both public cloud and private deployment can serve customers at a high iteration speed

Qingliu did this, shall we do the same? Right now

Develop a set of tools for data modeling through forms, support process management, and support automated tasks
Develop a visual UI editor to support the production of mobile applications, portals, and reports
Provide plug-in capability to facilitate system integration and function expansion for IT personnel
Continuously integrate external mainstream systems and business capabilities to enrich the platform ecosystem

think

These platforms all provide data modeling capabilities, one is a database-like modeling process, and the other is a form-based modeling process
These platforms all provide UI editors, which can visually build front-end pages and support multiple platforms (mobile and web)
These platforms all provide code development capabilities and can write logic at the code level
These platforms provide extension mechanisms to integrate external systems and business capabilities

It seems that if you want to be a productivity platform, you must provide these four major functions

Think from first principles

Business operation is the process of solving customer problems. From past experience, there are two types of people who need digital solutions when solving problems

It has not been verified that there is a customer demand for a certain problem, and I want to verify it as soon as possible
The existence of customer needs has been verified and needs to be launched as soon as possible

The first type of people, if you use code development to realize the requirements, the cost is very high, you can find a way to verify at low cost through other methods, for example, when Meituan Waimai was first launched, customer service personnel processed takeaway orders manually, and Stripe was manual in the early days Fill in the transaction information and send it to the bank

The second type of people, the customer needs have been verified, and only a programmer is needed. At this time, communication costs, personnel development costs, and time costs account for the majority

How to solve the problem of the second type of people?

To increase productivity, it is better to let the person with the problem solve the problem as soon as possible

give three examples

For example, if we encounter a certain math problem and cannot solve it, the way to improve productivity is to have a teacher teach us to solve it.
For example, we want to go to a certain scenic spot but find that the road is far away. The way to improve productivity is to have a car that can take us there.
For example, when we are hungry, we don’t have any ingredients and kitchen utensils at home. The way to increase productivity is to go to the nearest restaurant or order takeaway instead of cooking for ourselves.

Starting from this law, we have to think about how to make people with problems solve the problem as soon as possible.

Ideally, the fastest way to solve a problem is to pay a professional to solve the problem. Outsourcing the problem will have the best results and performance.

Is there any other second best way besides this way?

Reduce the difficulty of using tools while ensuring the speed of problem solving

Just like in the past, it was necessary to drill wood to make a fire. Later, flints and matches were used to ignite, and now lighters are more and more convenient to ignite, and the difficulty of use is getting lower and lower.

So for the development of digital solutions, how can we quickly solve the problem and reduce the difficulty of use?

Digital solutions generally need to go through the four processes of data acquisition, processing, storage, and display. Let's look at it in turn

data acquisition

Currently, the most popular data acquisition method is Excel. The appearance of forms makes the data filling experience better, but when operating a large amount of data, Excel is the most efficient

data processing

For data processing logic, the method provided by Excel is formula, and the method provided by developers is code. The least difficult to use is configuration (check the box in the configuration interface) and use case (reuse existing logic). No need To memorize the usage of complex formulas, there is no need to practice how to write code. The difficulty here is how to keep the difficulty of using it from increasing exponentially when the logic is complex

data storage

Formatted data (such as credit card numbers, phone numbers, home addresses, product names) are generally stored in Excel or databases, and unformatted data (such as word, audio, video) are generally stored directly on the hard disk. The least difficult to use is Excel, one by one, just fill in the data

data display

The current display terminal is too rich, such as mobile APP, website, small program, computer software, TV large screen, etc. This is also the most difficult part to reduce the difficulty of use. There are too many personalized needs. If you want to achieve the fastest use The lowest level of difficulty requires a UI editor that is very easy to use. It is best to be as simple as Markdown, focusing only on content and automatic rendering of styles. PPT is too difficult for ordinary people

In summary,

A: For data acquisition and storage, we need a solution similar to Excel, which can store data with low difficulty and high efficiency
B: For data processing, we need a configurable and case-based solution that can handle various logic from simple to complex, and keep it simple to use
C: For the display of data, we need a solution that only focuses on the content and automatically matches the style, supporting multiple screens and terminals

solution

For A, we can refer to foreign Airtable, domestic Feishu's multi-dimensional tables, and Tencent document online tables, which support multi-person collaboration, support large-scale data operations, and support multiple data types

Screenshot of Airtable

Multidimensional table screenshot

For B, we can refer to Apple’s shortcut, foreign Zapier, these tools can program logic, and can expand usage scenarios by introducing more applications and operations, achieving a good balance between difficulty of use and logical maintainability

Apple shortcut screenshot

Screenshot of Zapier

For C, we can refer to foreign Notion and Markdown to provide tools that allow users to focus more on content, automatically match styles through AI or templates, and infinitely expand capabilities through blocks

Notion Screenshot

Screenshot of Markdown

Summarize

If you want to build a platform that can improve productivity, the best way is to help people who encounter problems solve problems as quickly as possible

If we use analogical thinking to think, we may have to do very large and comprehensive things, such as

Building Data Modeling Capabilities
Build UI editor capabilities
Build code development capabilities
Build an extension mechanism

If we adopt first-principles thinking, we can divide productivity into four stages of data acquisition, processing, storage, and display, and use tools that are fast and easy to use to process, that is,

Build a new generation of table capabilities, support online multi-person collaboration, support multiple data types, and support large batches of data
Build logical arrangement capabilities, similar to shortcuts and Zapier, to achieve logical order arrangement and expansion
Build a more content-focused editor capability, users only need to write content, and automatically match the style through methods such as AI and templates

Author: JD Retail Ye Meng

Source: JD Cloud Developer Community