To talk about the proportion of developers and testers-test

People often still like to dwell on some specific numbers, in particular, westerners like to use data to explain the problem, because it is objective, specific, but also often lead people astray, easy to metaphysics, because each company, the company's products, products, various projects or stages are different, can not use a one-size-fits-all approach. In software companies, the question often asked of test managers is "how much of your company's developers and testers." "If you answer" 2:1 ", the reaction may be different, the other side may say:" Your company attaches great importance to quality, testing good AH. " It may also be said that there are too many testers in your company, and the developers are not very well.

The proportion of developers and testers in a software enterprise is often a matter of concern to managers, or it may be a headache for the test manager below, and no one seems to know what proportion is appropriate. I've written an article about how much of the tester's ratio to the developer is reasonable. To elaborate on this issue from all aspects.

Fortunately, there is a scholar to do this investigation--tester to Developer Ratio Initial study findings, because the idea also entangled him for a full ten years. He gets some data from 4 questions, which are available to those interested in it. 4 Questions are:

1 How many developers are there in your organization?

2 How many testers are there in your organization?
3 if the range of 1 to 6, of which 1 represents low, 6 represents high, you give the current rate of the effectiveness of the number of points.

4 There are other anecdotes about the effectiveness of the current ratio.

The survey showed the poorest testers: 20 Developers for 1 testers (but less efficient, 2 testers are the most abundant: 15 developers have an exception data for 8 testers (high efficiency, 4): The average ratio of 4 developers to 0 testers (efficiency is 3) is The most common scenario for 4.52 developers for 1 testers is: 3 developers to 1 Testers: 2.5 developers to 1 testers most of the developer and tester ratios are 3:1 or lower (that is, 2.5:1 or 2:1)

Effectiveness is also an important consideration, because although the proportion of developers and testers is an objective existence, the proportion is too high or too low may be feasible, but whether to bring high-quality products is an important consideration, or whether it seriously affects the business. As a result, the ratio of developers and testers in successful companies is often more persuasive, and that is how much of the test-developer ratio is reasonable, as exemplified by Microsoft and Google. discussed in.

From another point of view, the quality is constructed, from the requirements, design to coding, only each link is done, quality can go up. And these links are done well, especially if the developers have done enough unit testing, testers can be significantly reduced. If more thorough, developers have strong sense of responsibility and good quality, ability, from the beginning of the project on the overall quality responsibility, the developer not only outstanding design and code, but also fully complete the relevant unit testing, functional testing, performance testing, security testing, then no need to test personnel. In other words, software testing can completely allow a highly responsible developer to complete, although such developers in the country is not much (also rare in foreign countries).

Improve developer: The proportion of testers, there is also a positive side, it will drive product designers, developers have a stronger quality ownership, take more responsibility, do a good job analysis, design, write code and fully complete unit testing, improve the quality of the output of the phased results. To achieve such a goal, you need to start with management, have the same knowledge-such as quality is built, the quality of the SOFTWARE PRODUCT is more dependent on the product designers and developers, to improve the quality and efficiency of the entire team for the purpose of continuous drive product designers and developers to do a good job, lead the continuous improvement process, Ultimately the whole team is benefiting. From this perspective, the "developer: the ratio of testers" problem is not a challenge to the Test manager, but a challenge to the entire team, and more to the challenge to the developer, who should always ask the developer: Design refactoring is often done. Why write so many flaws. The code specification is improved based on the results of the defect analysis. The next version of code defects per thousand lines can reduce 30%~50%. Unit test coverage is over 80%. ......

Another way of thinking, if developers like that to do the work of testers, cost-effective. Testers are less, more developers, the overall development of the cost is not reduced, but the cost is increased, and why to pursue which 3:1 or 4:1 of the data. Testers are more at the customer point of view, the complementary role of developers can not be ignored.

The most fundamental of software development is quality and productivity. If productivity is not measurable, try to reduce costs. Everything that helps improve product quality and productivity is something we have to do. Sometimes, numbers don't matter.

Reference: Discussion: What is the most appropriate development test in the research and development team?