Rule 3: How can architects maximize their incremental value within a certain period of time
As an architect, it is necessary to create enough commercial value in order to guarantee the long-term career of oneself.
So how do you, as an architect, provide quantifiable incremental value to your company, department or team?
There are mainly two paths of increasing income and reducing costs.
In today's lesson, we will combine several real cases to analyze in detail. How to find opportunities to expand income? Some architects don't pay attention to things other than software, such as seldom caring about the company or department's income. Although this kind of personality allows him to focus on software work, in the long run, if he does not think about how to create business value for the company through technology, it will be difficult to maintain or expand his influence in the team, and his career development may also be frustrated . So even if no one puts pressure on you to increase revenue for the company, you have to actively think about how you or your department can create more revenue for the company. So where do we look for opportunities to expand revenue? In addition to the deep understanding of user psychology and business models mentioned earlier, as well as the different horizontal issues we often encounter, such as stability, security, testability, and maintainability, are there other methods? You may have heard the phrase "see big opportunities in small data, and small opportunities in big data", which actually applies to all of us technical people. The first half of the sentence refers to the opportunity to abstract commonality from individual needs. That is to get inspiration from the process of solving a specific problem, and find an opportunity that can be applied on a large scale. For example, when doing user research, you find that a user always takes screenshots and then sends the product to another person via WeChat. Then you will realize that you can develop a sharing tool to get more social fission through applets and DeepLink. The second half of the sentence refers to relying on data to polish the user experience. That is to find opportunity points through data analysis, and then through product and technology improvement, continuously improve conversion and reduce losses. This is a common approach to algorithms and digital operations. For example, the bounce rate of the homepage of Taobao App is 2%, which is continuously improved through data analysis and polishing.
How to find opportunities to reduce costs?
Some people put too much emphasis on the geek spirit and pursue perfection in everything. I think it's a good starting point.
But in an enterprise, the pursuit of perfection must be based on the premise of cost control.
I invented a method that can accurately measure the expected output of each page after performance optimization. And what we have to do is to divide the expected output by the expected input cost, that is, the expected return ROI to sort, and then optimize in turn. The specific formula is relatively long, so I will not expand it here. The core principle of this algorithm is shown in this picture [2]:
As shown in the figure, based on the histogram of the conversion rate distribution of each page and the expected performance optimization results, we predict the conversion rate of the page that will be lost if performance optimization is not performed. We call this predicted value the page performance loss Lpage. Because we have the conversion funnel and traffic statistics of the whole link, if we optimize a certain page and recover the performance loss, then this optimization can be calculated in advance through big data statistics for the downstream, as well as the prediction of orders and GMV . So we made architectural planning with the goal of optimizing the number of orders, and then coordinated all the optimization actions we could do. As shown in the figure below, some of the optimization actions that are not equivalent at all are at the network layer, some at the front end, and some at the back end. Now we can compare on one indicator, because each of our optimization actions can be attributed to order contribution in the end.
How do I practice the Law of Survival? The key to the success of this architecture project is that I have been following the survival rules I introduced in these two classes. Now let's go through this performance optimization case and see how I use the rules of survival to guide architectural activities. In addition, through the process of case inspection, we will resummarize and emphasize some of the core points we conveyed in these two lessons.
First, as an architect, you must pursue commercial value in architectural design. When we do performance optimization, we do not simply optimize performance indicators, but aim to improve business value from the very beginning. Therefore, our optimization goal is the number of orders, not a technical indicator.
Second, if you want to create business value, you must constantly measure the incremental value you create, so as to ensure that you are at the forefront of value creation. And be able to keep looking for your own value-added space in a relatively unknown environment. From the beginning of this project, we never thought about optimizing the performance of the whole site, but found the single point with the greatest return and made targeted optimization. In order to do this, I invented a formula to accurately measure performance loss and found a single optimization goal (number of orders) at the department level. And attribute all our possible optimization actions to this single optimization goal. With this measurable value, we no longer do blanket performance optimization, but do targeted performance optimization actions with the highest returns. Also, we didn't get wider and wider. When it is found that the return of performance optimization is not large enough, we no longer do performance optimization. Instead, change the track to create value: do the performance monitoring of the existing network. The best proof is that at that time we cooperated with Akamai, which has the strongest research strength and the most perfect monitoring ability in the world. And we found that our ability to monitor the CDN network is far better than Akamai in many countries. Even later, we simply asked Akamai's operation and maintenance personnel to pick up some of our alarms. After that, we expanded the application direction to troubleshoot business transformation and so on.
Third, as an architect, it is necessary to minimize the cost of the entire architecture activity. What you have to do is: ensure the feasibility of the final solution; find the optimal implementation path to ensure that the implementation can be completed in the end; try to maximize the structure of the final solution and amplify your output with the minimum cost.