实验操作
使用集合让一切井井有条
根据您的偏好保存内容并对其进行分类。
设置好 Experiment
及其 ExperimentArm
后,并修改实验组以适合您的实验后,您可以执行各种操作来实际运行实验,并在实验完成后对其进行管理。
长时间运行的操作
其中一些操作是异步的,这意味着您的 Google Ads API 调用将开始该流程,但不会立即完成。您必须检查是否已完成。如需详细了解此方面的信息,请参阅有关处理异步错误的指南。
运维
您可以对实验执行四种操作。它们都位于 ExperimentService
中。
- 安排实验。这是异步的。将设计中的广告系列具体化为实际广告系列,在实验开始日期到达后即可投放。
- 结束实验。这是一个正常的同步操作。如果实验已在进行中,您可以立即结束实验,即使尚未达到结束日期也是如此。
- 推广实验。这是异步的。这会将实验组中所做的更改复制到对照组,并停止投放实验组。如果您喜欢这些更改,并希望将其纳入原始广告系列,请使用此选项。
- 结束实验。这是一个正常的同步操作。此操作会永久设置实验组广告系列,使其在实验结束后继续存在。不会对对照组广告系列做出任何更改。如果您喜欢这些更改,但希望它们与原始广告系列分开存在,请使用此选项。
如未另行说明,那么本页面中的内容已根据知识共享署名 4.0 许可获得了许可,并且代码示例已根据 Apache 2.0 许可获得了许可。有关详情,请参阅 Google 开发者网站政策。Java 是 Oracle 和/或其关联公司的注册商标。
最后更新时间 (UTC):2025-08-26。
[null,null,["最后更新时间 (UTC):2025-08-26。"],[[["\u003cp\u003eGoogle Ads experiments allow scheduling, ending, promoting, and graduating experiment variations to manage and implement changes.\u003c/p\u003e\n"],["\u003cp\u003eCertain experiment operations like scheduling and promoting are asynchronous, requiring completion checks, while ending and graduating are synchronous, taking effect immediately.\u003c/p\u003e\n"],["\u003cp\u003ePromoting an experiment applies treatment arm changes to the control arm and stops treatment serving, while graduating creates a separate, permanent treatment campaign alongside the original.\u003c/p\u003e\n"],["\u003cp\u003e\u003ccode\u003eExperimentService\u003c/code\u003e provides the functionalities to manage and control these experiment operations within Google Ads.\u003c/p\u003e\n"]]],[],null,["# Experiment operations\n\nOnce you've set up your `Experiment` and its `ExperimentArm`s, and modified the\ntreatment arm to suit your experiment, there are various operations you can\nperform to actually run the experiment and manage it once it is complete.\n\nLong running operations\n-----------------------\n\nSome of these operations are asynchronous, which means that your Google Ads API call\nwill begin the process, but it won't be complete right away. Instead, you will\nhave to check for completion. For more details on this, see the guide on\nhandling [async errors](/google-ads/api/docs/experiments/async-errors).\n\nOperations\n----------\n\nThere are four kinds of operations you can perform on an experiment. They are\nall housed in the [`ExperimentService`](/google-ads/api/reference/rpc/v21/ExperimentService).\n\n- **Schedule** the experiment. This is asynchronous. Materializes the in design campaigns to actual campaigns, ready to serve once the experiment's start date arrives.\n- **End** the experiment. This is a normal synchronous operation. If the experiment is already in progress, you can end it immediately, even if the end date hasn't been reached.\n- **Promote** the experiment. This is asynchronous. This will copy changes made in the treatment arm to the control arm, and stop the treatment arm from serving. Use this if you liked the changes and want to incorporate them into your original campaign.\n- **Graduate** the experiment. This is a normal synchronous operation. This will permanently set up the treatment campaign to continue to exist, outside of the context of the experiment. No changes will be made to the control campaign. Use this option if you liked the changes, but want them to exist separately from the original campaign."]]