Principles. Improvement and Measurement

12009

Overview

No matter what kind of improvements are made, if the results cannot be measured in some way, it cannot be proven that real improvements have been achieved. Before addressing each issue, we must find appropriate measurement methods and place them on an equal footing with the corresponding functional requirements to complete them together.

For example, in a data annotation volunteer recruitment examination system, although it is divided into multiple iterations, each releasing many features, the lack of data collection and statistical analysis functions that product personnel are concerned about means that the team only knows that people can use this system to complete tasks, but cannot determine whether they can do so efficiently, making it difficult to propose the next steps for product optimization.

Application

For code quality, we can measure it using various tools and metrics, such as code complexity, code coverage, and defect density. Through regular measurement and comparison, we can clearly see changes in code quality, thereby identifying areas that need improvement.

For development efficiency, we can measure it by tracking developers' working hours, task completion times, and bug fix times. Through this data, we can identify inefficient links and then take measures to improve them.

For system performance, we can measure it by monitoring system response times, throughput, and resource utilization rates. If performance metrics do not meet standards, we need to identify performance bottlenecks and optimize them.

For user satisfaction, we can measure it through user surveys, user feedback, and web browsing data. If user satisfaction is low, we need to identify the reasons for user dissatisfaction and make improvements.

Allowing Failure

Although every product manager hopes that all solutions will succeed, we cannot guarantee that every solution will be successful. However, as long as we maintain an open mindset, we can learn a lot of new knowledge from all solutions, and these gains are crucial for both the future success of the product and the improvement of the team's capabilities.

For example, an e-commerce website is redesigning its product detail page. To promote the order conversion rate, the product manager plans to redesign the product detail page. He provided 14 experimental solutions, each with only minor changes. For instance, displaying the price in the currency used by the user's location; placing product images on the right side of the webpage. Other solutions included changes in font or button size and color. The final result of these 14 experiments was that each solution had little impact on the order conversion rate, and some even resulted in a decrease in the conversion rate.

However, during these 14 experiments, the product team gained new insights into users, such as which elements users focus on and which they are not very sensitive to. Ultimately, based on these experiments, the page underwent a comprehensive redesign, and the results of this version significantly improved the order conversion rate.

## Overview

No matter what kind of improvements are made, if the results cannot be measured in some way, it cannot be proven that real improvements have been achieved. Before addressing each issue, we must find appropriate measurement methods and place them on an equal footing with the corresponding functional requirements to complete them together.

For example, in a data annotation volunteer recruitment examination system, although it is divided into multiple iterations, each releasing many features, the lack of data collection and statistical analysis functions that product personnel are concerned about means that the team only knows that people can use this system to complete tasks, but cannot determine whether they can do so efficiently, making it difficult to propose the next steps for product optimization.

## Application

For code quality, we can measure it using various tools and metrics, such as code complexity, code coverage, and defect density. Through regular measurement and comparison, we can clearly see changes in code quality, thereby identifying areas that need improvement.

For development efficiency, we can measure it by tracking developers' working hours, task completion times, and bug fix times. Through this data, we can identify inefficient links and then take measures to improve them.

For system performance, we can measure it by monitoring system response times, throughput, and resource utilization rates. If performance metrics do not meet standards, we need to identify performance bottlenecks and optimize them.

For user satisfaction, we can measure it through user surveys, user feedback, and web browsing data. If user satisfaction is low, we need to identify the reasons for user dissatisfaction and make improvements.

## Allowing Failure

Although every product manager hopes that all solutions will succeed, we cannot guarantee that every solution will be successful. However, as long as we maintain an open mindset, we can learn a lot of new knowledge from all solutions, and these gains are crucial for both the future success of the product and the improvement of the team's capabilities.

For example, an e-commerce website is redesigning its product detail page. To promote the order conversion rate, the product manager plans to redesign the product detail page. He provided 14 experimental solutions, each with only minor changes. For instance, displaying the price in the currency used by the user's location; placing product images on the right side of the webpage. Other solutions included changes in font or button size and color. The final result of these 14 experiments was that each solution had little impact on the order conversion rate, and some even resulted in a decrease in the conversion rate.

However, during these 14 experiments, the product team gained new insights into users, such as which elements users focus on and which they are not very sensitive to. Ultimately, based on these experiments, the page underwent a comprehensive redesign, and the results of this version significantly improved the order conversion rate.