Test plan・Test metrics (Part 1): Schedule and Deliverable Progress Management

27/04/202121/10/2021Test qualityy

In test management, first decide which metrics to use for management

Since the software testing is difficult to see in the eyes, test plan is the way of test management It is important that you be as specific a. The monitoring of the test situation use any metrics what, the status of the test to do is interval to monitor whether the test is when a deviation from the plan has occurred to any correspondence what, write a specific method of management to plan It’s a good idea to keep it. In this article, I will give an overview of what metrics are available, using the test plan that Father Gutara has seen as an example, and then look at the specifics of the metrics that are mainly used to manage the progress of tests.

If you use test goals for metrics , subdivide by test period

The first metric that comes to mind for test management is the test goal . This can be used as a management metric because concrete and quantitative test goals should be set from several viewpoints such as delivery date, cost, product quality, process quality, work quality, and control quality . However, test goals often indicate the final goal reached at the end of the test, which is difficult to manage and needs to be subdivided by test period .

For example, suppose you set a bug detection rate of 0.1 bugs / test items as your test goal . The meaning of this goal is to set the goal of the value calculated by the total number of bug detections / total number of test items to 0.1 when all the tests are completed . Trying to use this target value as it is in test management, calculate the actual value of the bug detection rate at the present time with respect to the target value 0.1 bug / test item value, and whether the actual value meets the target value We will monitor whether or not, and if there is a difference, we will proceed with countermeasures .

However, in reality, simple bugs are often found in the first 30% of the test, so the bug detection rate is about 0.2 bugs / test items , and in the last 30% of the test, only very complicated bugs are found. A common pattern is that the bug detection rate drops to 0.05 bugs / test items because they are not found . Comparing the final goal of 0.1 bugs / test items with actual values ​​for the entire test period is not very good management.

Then, what to do is to divide the test period into several parts , and subdivide the target of the bug detection rate for each period, for example, 0.2 bugs / test items for the first period and 0.05 bugs / test items for the last period. By making it and setting it, you can manage it more correctly. In this way, when using the metrics determined by the test goals for test management, write down the goals subdivided for each test period in the test plan as necessary.

Metrics other than test goals may be used for management

Test goals are not the only metrics used for test management. Even if it is a test goal, if it takes a lot of time to collect achievements, it is difficult to use for test management, so it may be excluded from the test management metrics. On the other hand, if you have a metric that you don’t include in your test goals but need to keep track of your performance on a regular basis , it’s a good idea to use it for test management. So, the first step of the test management is, that whether you want to manage any item item that you want to manage Decide, suitable for the management choose the metrics thing. Now let’s take a closer look at what test management items and metrics are.

Progress management is considered in two ways: schedule progress management and deliverable progress management.

　In the progress management of the test process (also called process management, but the term progress management is used in this article), we will introduce two specific examples of quantitatively developing software and testing progress management in another article. As you can see, it is good to manage the progress from the two perspectives of schedule and deliverables. Both are easy to use for management by calculating metrics (evaluation values) that are easy to compare with the purpose based on the collected actual values. Let me give you a few concrete examples of what the metrics are.

Test schedule progress management metrics

The difficulty in managing the progress of a test schedule is the criteria for activating management, such as how much delay the planned schedule should take before taking countermeasures . It is important to adopt metrics that quantify the progress management of the schedule so that the decision to start management can be made as easily as possible.

(1) Milestone completion rate

In the test scheduling, you will have a day plan that starts on what day of the month and ends on what day of the month for each detailed work that breaks down. You can use the completion rate of milestones, which regards the end date of each detailed work as a milestone, as a metric for progress management.

Completion rate = (Number of milestones actually completed by the monitoring date * 100) / The number　of milestones scheduled to be completed by the monitoring date . If all milestones are completed as planned, the completion rate is 100%, but if some milestones are past the scheduled date, the completion rate will not be 100%. When this completion rate falls below a certain level, it indicates that the process delay is increasing. When the completion rate falls below 85%, it may be time to revise the schedule.

(2) Milestone achievement rate considering man-hours

Milestone completion rates are an easy-to-use metric, but if the granularity of the detailed work that breaks down is different, it will not be possible to express the actual progress well . For example, if the man-hours for detailed work A are 5 man-days and the man-hours for detailed work B are 20 man-days, the effect of process delay is greater for detailed work B, but this is not taken into consideration in the milestone completion rate. Hmm.

In this way, if there is a variation in the particle size of the man-hours required for each detailed work, there is a means to correct it using a coefficient that takes into account the variation .

Corrected completion rate = (correction coefficient * number of milestones actually completed by the monitoring date * 100) / The　correction coefficient will be added to the number of milestones scheduled to be completed by the monitoring date and the calculation formula. In the above example, the detailed process A can be calculated with the correction coefficient = 1 and the detailed process B = correction coefficient 4, so that the particle size of the man-hours for each detailed process can be corrected.

(3) Man-hour digestion rate

You can also manage the progress by focusing on the work man-hours rather than the schedule. In the test scheduling, plan how many man-hours are required along with the end schedule of what month and what day each broken down detailed work will be completed. It is a method to calculate the man-hour digestibility based on how much man-hours are completed for each detailed work and use it for progress management metrics.

Man-hour digestion rate = (total man-hours　for detailed work completed by the monitoring date * 100) / Total man-hours for detailed work scheduled to be completed by the monitoring date . If there is a delay in inputting the planned work man-hours, it means that some kind of work delay has occurred, so you can manage the schedule by monitoring the man-hour digestion rate. Strictly speaking, the man-hour digestion rate is not the progress management of the schedule, but since it is not the progress management of the deliverables, it is classified as the progress management of the schedule in this article.